development of a particle number and …eprints.qut.edu.au/30297/1/diane_keogh_thesis.pdf · jurgen...

Queensland University of Technology

School of Physical and Chemical Sciences

DEVELOPMENT OF A PARTICLE NUMBER AND

PARTICLE MASS EMISSIONS INVENTORY FOR AN

URBAN FLEET: A STUDY IN SOUTH-EAST

QUEENSLAND

Diane Underwood Keogh

A thesis submitted in partial fulfillment of the requirements of the degree of

Doctor of Philosophy

2009

i

KEYWORDS

Air quality regulation, ambient aerosol, emission factors, modality of particle size

distribution, motor vehicles, motor vehicle inventory, particle mass, particle

number, particle volume, PM1, PM2.5, PM10, South-East Queensland,

submicrometre particles, tailpipe emissions, traffic, transport modelling, ultrafine

particles, urban fleet.

ii

ABSTRACT

Motor vehicles are a major source of gaseous and particulate matter pollution in

urban areas, particularly of ultrafine sized particles (diameters < 0.1 µm).

Exposure to particulate matter has been found to be associated with serious health

effects, including respiratory and cardiovascular disease, and mortality. Particle

emissions generated by motor vehicles span a very broad size range (from around

0.003-10 µm) and are measured as different subsets of particle mass

concentrations or particle number count. However, there exist scientific

challenges in analysing and interpreting the large data sets on motor vehicle

emission factors, and no understanding is available of the application of different

particle metrics as a basis for air quality regulation. To date a comprehensive

inventory covering the broad size range of particles emitted by motor vehicles,

and which includes particle number, does not exist anywhere in the world.

This thesis covers research related to four important and interrelated aspects

pertaining to particulate matter generated by motor vehicle fleets. These include

the derivation of suitable particle emission factors for use in transport modelling

and health impact assessments; quantification of motor vehicle particle emission

inventories; investigation of the particle characteristic modality within particle

size distributions as a potential for developing air quality regulation; and review

and synthesis of current knowledge on ultrafine particles as it relates to motor

vehicles; and the application of these aspects to the quantification, control and

management of motor vehicle particle emissions.

iii

In order to quantify emissions in terms of a comprehensive inventory, which

covers the full size range of particles emitted by motor vehicle fleets, it was

necessary to derive a suitable set of particle emission factors for different vehicle

and road type combinations for particle number, particle volume, PM1, PM2.5 and

PM1 (mass concentration of particles with aerodynamic diameters < 1 µm, < 2.5

µm and < 10 µm respectively). The very large data set of emission factors

analysed in this study were sourced from measurement studies conducted in

developed countries, and hence the derived set of emission factors are suitable for

preparing inventories in other urban regions of the developed world. These

emission factors are particularly useful for regions with a lack of measurement

data to derive emission factors, or where experimental data are available but are

of insufficient scope.

The comprehensive particle emissions inventory presented in this thesis is the first

published inventory of tailpipe particle emissions prepared for a motor vehicle

fleet, and included the quantification of particle emissions covering the full size

range of particles emitted by vehicles, based on measurement data. The inventory

quantified particle emissions measured in terms of particle number and different

particle mass size fractions. It was developed for the urban South-East

Queensland fleet in Australia, and included testing the particle emission

implications of future scenarios for different passenger and freight travel demand.

iv

The thesis also presents evidence of the usefulness of examining modality within

particle size distributions as a basis for developing air quality regulations; and

finds evidence to support the relevance of introducing a new PM1 mass ambient

air quality standard for the majority of environments worldwide. The study found

that a combination of PM1 and PM10 standards are likely to be a more discerning

and suitable set of ambient air quality standards for controlling particles emitted

from combustion and mechanically-generated sources, such as motor vehicles,

than the current mass standards of PM2.5 and PM10.

The study also reviewed and synthesized existing knowledge on ultrafine

particles, with a specific focus on those originating from motor vehicles. It found

that motor vehicles are significant contributors to both air pollution and ultrafine

particles in urban areas, and that a standardized measurement procedure is not

currently available for ultrafine particles. The review found discrepancies exist

between outcomes of instrumentation used to measure ultrafine particles; that few

data is available on ultrafine particle chemistry and composition, long term

monitoring; characterization of their spatial and temporal distribution in urban

areas; and that no inventories for particle number are available for motor vehicle

fleets. This knowledge is critical for epidemiological studies and exposure-

response assessment. Conclusions from this review included the recommendation

that ultrafine particles in populated urban areas be considered a likely target for

future air quality regulation based on particle number, due to their potential

impacts on the environment.

v

The research in this PhD thesis successfully integrated the elements needed to

quantify and manage motor vehicle fleet emissions, and its novelty relates to the

combining of expertise from two distinctly separate disciplines - from aerosol

science and transport modelling. The new knowledge and concepts developed in

this PhD research provide never before available data and methods which can be

used to develop comprehensive, size-resolved inventories of motor vehicle

particle emissions, and air quality regulations to control particle emissions to

protect the health and well-being of current and future generations.

vi

ACKNOWLEDGEMENTS

My sincere appreciation and thanks go to my Supervisors Professor Lidia

Morawska, Professor Luis Ferreira and Dr. Zoran Ristovski for their endless

patience, commitment, encouragement, guidance, academic training and help. In

particular, I would like to especially thank Lidia for her support and encouragement

throughout this very interesting and exciting course.

I am also grateful to Professor Kerrie Mengersen for her professional guidance and

encouragement, and to Joe Kelly, Sean Moynihan and Jaime Mejia for their help

with modelling and maths. The financial support arranged by Ray Donato,

Jurgen Pasieczny and Randall Fletcher is gratefully appreciated. Special thanks go

to Rachael Robinson for her support and friendship. Many thanks to Hussein

Kanaani and Afkar Al Farsi for helping me with physics instrumentation, and for

your kind friendship.

Thanks very much to Dr. Nick Holmes for his kind help, patience and support, to

Dr. Graham Johnson for answering my endless questions; to Dr. Rohan Jayaratne,

Dr. Congrong He and Dr. Milan Jamriska for providing very helpful advice, and to

Xuan Ling for our discussions. The assistance of Scott Cormack, Andrew Joycey

and Dan Harney with traffic modelling and data, Jeff Eaton, Andrew Copland and

Vernon Alcantra for providing transport data, and helpful discussions with Bill

Duncan, John Woodland and Dr. Sama Low Choy are also greatly appreciated.

vii

A very special thank you goes to Kate McKee from the Research Centre for being

such an encouraging and professional ambassador for the Queensland University

of Technology.

My sincere appreciation is extended to the Queensland University of Technology

for enabling me to take up this wonderful opportunity; and to the staff and

students at the International Laboratory for Air Quality and Health for providing a

very positive, happy and encouraging environment in which to work.

Last, but not least, thank you to Dr. David Freebairn and Dr. Sunil Dutta who

were my first ‘unofficial’ Professors in the field of applied climate research; and

to Dr. Alfio Parisi and Dr. Michael Kimlin for introducing me to the exciting

world of Physics.

viii

LIST OF PUBLICATIONS

Morawska, L., Keogh, D.U., Thomas, S.B., Mengersen, K., 2008. Modality in

ambient particle size distributions and its potential as a basis for developing air

quality regulation. Atmospheric Environment 42 (7), 1617-1628.

Morawska, L., Ristovski, Z., Jayaratne, E.R., Keogh, D.U., Ling, Z. 2008.

Ambient nano and ultrafine particles from motor vehicle emissions:

characteristics, ambient processing and implications on human exposure.

Atmospheric Environment 42 (35), 8113-8138.

Keogh, D.U., Kelly, J., Mengersen, K., Jayaratne, R., Ferreira, L., Morawska, L.,

2009. Derivation of motor vehicle tailpipe particle emission factors suitable for

modelling urban fleet emissions and air quality assessments. Environmental

Science and Pollution Research – International. Published online, doi

0.1007/s11356-009-0210-9.

Keogh, D.U., Ferreira, L., Morawska, L., 2009. Development of a particle number

and particle mass vehicle emissions inventory for an urban fleet. Environmental

Modelling & Software 24(11), 1323-1331.

ix

ORAL PRESENTATIONS & COURSE UNIT PRESENTATION

18th Coordinating Research Council On-Road Vehicle Emissions Workshop,

“Development of a particle number and particle mass emissions inventory for an

urban fleet”, San Diego, USA, 31 March – 2 April, 2008.

14th IUAPPA World Congress (International Union of Air Pollution

Prevention & Environmental Associations), “Emission factors for estimating

motor vehicle particle emissions in urban areas”, Brisbane, 9-13 September, 2007.

11th International Health Summer School, International Symposium on

Environmental Health, Climate Change & Sustainability, “Modality in


quality regulation”, Queensland University of Technology, 21-22 November,

2006.

Air Pollution & Transport Short Course, Transport Futures Institute,

delivered jointly by the University of Queensland and Queensland University of

Technology, “Emission Factors: The Evidence”, Brisbane, 1-3 August, 2007.

x

TABLE OF CONTENTS

Keywords ………………………………………………...…….….. i

Abstract …………………………………………………………….. ii

Acknowledgements ………………………………………………… vi

List of Publications … ……………………………………………... viii

Oral Presentations and Course Unit Presentation ……………….…. ix

List of Tables …..…………………………………………………… xxii

List of Figures …………………………………………………….… xxv

Statement of Original Authorship ………………………………….. xxvi

xi

TABLE OF CONTENTS (cont’d)

CHAPTER 1. INTRODUCTION …………………....................... 1

1.1. Description of the scientific problem investigated …………. 1

1.1.1. Developing comprehensive inventories of motor

vehicle particle emissions …………………….…….. 3

1.1.2. Combining knowledge from two different disciplines to

develop inventories …………………………….…... 5

1.1.3. Why particle number emission inventories are

important …………………………………….…….. 6

1.1.4. Designing environmentally-sustainable

transport systems …………………………………… 7

1.1.5. Examination of modality within particle size

distributions and its potential as a basis for

developing air quality regulation …………….……….. 8

1.2. The major components of this PhD research ……….……..….. 10

1.3. The objectives of the study ………………………………….... 12

1.4. Account of scientific progress linking the scientific papers…… . 18

1.5. The important and novel contribution of this PhD research…… 29

1.6. References ……………………………………………………. 32

xii


CHAPTER 2. LITERATURE REVIEW ……………….................. 35

2.1. Introduction ……………………………………………………..... 35

2.2. Characteristics of motor vehicle particle emissions ………........... 36

2.2.1. The nature of particle emissions …………………….…... 37

2.2.2. Motor vehicle particle emissions ………………………... 39

2.2.3. Diesel particle emissions ………………………………... 40

2.2.4. Health effects associated with exposure to particles…….. 41

2.2.5. Current air quality standards to control particulate matter .. 43

2.2.6. Modality within particle size distributions …… ..……… 45

2.3. Vehicle emission inventories and local models …….…..…..……. 47

2.3.1. Developing emission inventories ………….……..……… 47

2.3.2. Local inventories and models for South-East Queensland

and Queensland ………………………………………….. 51

2.4. Transport models …………………………………….………… 55

2.5. Estimate of road transport emissions prepared for the UK ..…. 56

2.6. Identifying suitable particle emission factors .…………………. 58

2.7. Summary ………….…………………………………..……… 59

2.8. Knowledge gaps and conclusions from this review …………... 63

2.9. References …………………………………………………… 65

2.10. Bibliography ……………..……………………………… ...... 80

xiii


CHAPTER 3. STATISTICAL TECHNIQUES USED IN THIS

THESIS …………………………………………………………………… 127

3.1. Introduction …………………………………………….………….... 127

3.2. Statistical techniques used in this thesis …………………………….. 129

3.2.1. Kolmogorov-Smirnov (K-S) test ……………………………… 129

3.2.2. Construction of 95% confidence intervals ………………...…. 130

3.2.3. Trapezoidal rule for integration of the area under a curve ……..131

3.2.4. Development of statistical models using linear regression

and ANOVA …………………………………………………. 132

3.2.5. Linear regression for continuous variables ………………….. 132

3.2.6. Multifactor Analysis of Variance (ANOVA) for categorical

variables ……………………………………..…..………..…..133

3.2.7. The stepwise selection technique for statistical model

selection ……………………………………………………. 134

3.2.8. Scheffe’s multiple comparison tests ……………..…….… . 135

3.3. Alternative approaches considered but not used in this thesis …….…136

3.3.1. Principal Component Analysis …………………………... 136

3.3.2. Non-parametric methods of statistical comparison ……..…. 138

3.3.3. Techniques for integrating the area under a curve …….……138

3.3.4. Techniques for selecting variables in statistical model

development …………………………….…………………. 139

3.3.5. Multiple comparison methods ……………………………. 140

3.4. References ……………………………..…………………………. 143

xiv


CHAPTER 4. MODALITY IN AMBIENT PARTICLE SIZE

DISTRIBUTIONS AND ITS POTENTIAL AS A BASIS FOR

DEVELOPING AIR QUALITY REGULATION ............................ 146

4.1. Introduction ……………………………….…………………….... 149

4.2. Methods and techniques ………………………….……................ 153

4.3. Results and discussion ………………………………………….... 157

4.3.1. Contribution of the modes in South-East Queensland to

PM1, PM2.5, PM10 ..……………………………………….. 157

4.3.2. Modal locations in the published literature ……………….. 161

4.3.3. Separation between modal location values in mass and

volume particle size distributions at around 1 µm…………... 166

4.4. Conclusions ………………………………………………………….. 168

4.5. References …………………………………………..………………. 171

xv


CHAPTER 5. DERIVATION OF MOTOR VEHICLE PARTICLE

EMISSION FACTORS ………….....……………………….....……......... 179

5. Overview of Chapters 5.1. and 5.2. .…………………………..……..… 180

CHAPTER 5.1. DERIVATION OF MOTOR VEHICLE TAILPIPE

PARTICLE EMISSION FACTORS SUITABLE FOR MODELLING

URBAN FLEET EMISSIONS AND AIR QUALITY ASSESSMENTS 183

5.1. Background, aim and scope ………………………………………….... 187

5.2. Materials and methods .…………………………………….…………. 189

5.2.1. Model variables examined ……………………………..……... 195

5.2.2. Statistical analysis of variables ………………………………. 200

5.2.3. Basis for selection of the most suitable emission factors …… 202

5.3. Results ……………….…………………………………….………… 203

5.3.1. Sample size of emission factors examined in the statistical

models…………………………………………………………. 203

5.3.2. Statistical models developed to derive average emission factors. 203

xvi


5.4. Discussion ………………………………………………………… 207

5.4.1. Statistical models used to derive average emission

factors ………………………….……………………..….. 207

Particle number model ……………………….………..…. 207

Particle volume model ……………………………………... 208

PM1 model ………………………………………………… 208

PM2.5 model ……………………………………….…….…. 209

PM10 model …………………………………….…………... 209

Total particle mass model …………………….…………… 211

5.4.2. Statistical differences between published emission

factors ……………………………..…..……………..…... 211

5.4.3. Relevance and application of the average particle

emission factors presented in this study ..………………… 213

5.5. Conclusions .…………………………………………………..... 214

5.6. Recommendations and perspectives …….………………….….. 216

5.7. References ………………………………………………….…. 218

xvii


CHAPTER 5.2. DERIVATION OF MOTOR VEHICLE PARTICLE

EMISSION FACTORS – STATISTICAL MODEL OUTPUTS …. 232

5. Introduction …………………………………………….…….……. 232

5.1. Statistical model outputs ………………………..…........... 233

5.2. Statistical relationships between categorical variables …... 245

5.3. Additional comments on particle volume and PM10 emission

factors ……............................................... .......................... 254

5.4. Additional comments on PM10 emission factors used in the

urban SEQ inventory……………………………………….. 256

5.5. References ……………………………………….................. 259

xviii


CHAPTER 6. DEVELOPMENT OF A PARTICLE NUMBER AND

PARTICLE MASS EMISSIONS INVENTORY FOR AN URBAN

FLEET ………………………………..………………………………….. 260

6.1. Introduction …………………………………………………………... 264

6.2. Method ………………………………………………….………….… 268

6.2.1. Study region. ……………… …………….………….……….… 270

6.2.2. Transport model .……………………..………………..…………. 271

6.2.3. Emission factors …………………………….………………….. 274

6.2.4. Variables used in the scenario analyses …………….………….. 276

6.3. Results and discussion …………………..…………………….………. 279

6.3.1. Particle inventory for urban SEQ for 2004 …………….…….….. 279

6.3.2. Comparing the urban SEQ particle inventory with other

inventories and models ……………………….………………… 284

6.3.3. Results of scenario analyses ……………………………….......... 288

6.4. Conclusions ……………………………………..……………………. 297

6.5. References …………………………………………….…..…………... 301

xix


CHAPTER 7. AMBIENT NANO AND ULTRAFINE PARTICLES

FROM MOTOR VEHICLE EMISSIONS: CHARACTERISTICS,

AMBIENT PROCESSING AND IMPLICATIONS ON HUMAN

EXPOSURE ……………………………………………………………… 308

7.1. Introduction ………………………………………………..…..…….. 312

7.2. Capabilities and limitations of particle number measurement

methods ………………………………………………………………… 313

7.3. Sources of particles in natural environment …..……………….…….. 319

7.4. Vehicle emissions as a source of ultrafine particles ………………… 323

7.4.1. Introduction …………………………………………………… 323

7.4.2. Primary Particles …………………………………………… 325

7.4.3. Secondary Particles ………………………………………… 326

7.5. Role of fuels …………………………………………………………. 328

7.6. Role of after-treatment devices …………………………………….. 332

7.7. Role of ions ……………………………………………………….... 335

7.8. Road-tyre interface ………………………………………………… 338

7.9. Emission factors and emission inventories ………………………… 339

7.10. Transport of particles within urban scale and ambient processing ... 342

7.10.1. Role of meteorological factors on particle concentration … 343

7.10.2. Relative role of various processes ……………………….. 346

7.11. Particle size distributions and modal location in urban

environments ………………………………………………………. 351

7.12. Chemical composition of ultrafine particles in different

environments ……………………………………………………… 352

xx


7.13. Temporal variation of particle characteristics …………………….. 358

7.13.1. Diurnal variation ……………………………………….. 358

7.13.2. Seasonal variation ………………………………………. 359

7.13.3. Long term variation …………………………………….. 361

7.14. Spatial distribution of particle concentrations within urban

environment ……………………………………….…….………… 363

7.14.1. Particle concentration as a function of the distance

from the road ……………………………………………… 364

7.14.2. Relationship between on-road and urban background

particle concentration ……………………………………… 368

7.15. Nucleation mode and its impact on urban particle concentrations ….. 369

7.16. Comparison of particle concentration levels between different

environments ……………………………………………………….. 375

7.17. Exposure to ultrafine particles ………… ……………………………377

7.18. Relationship between different particle metrics and with gaseous

pollutants …………………………………………………………… 378

7.19. Conclusions and implications for the exposure and epidemiological

studies ………………………..…………………………………. 381

7.20. References ……………………..………………………………… 387

xxi


CHAPTER 8. CONCLUSIONS ……………………………………… 422

8.1. Introduction ………………………………………….…………….. 422

8.2. Principal significance of the findings ………..………………….. 424

8.3. The principal findings and significance of this study ……………. 430

8.4. General conclusions from this study .…………………………. 470

8.4.1. Modality in ambient particle size distributions …..…… 470

8.4.2. A new mass ambient air quality standard for PM1,

and its combination with PM10 ………………………. 470

8.4.3. A comprehensive set of particle emission factors for

motor vehicles ………………………………….……. 472

8.4.4. The first published comprehensive particle emissions

inventory for a motor vehicle fleet ………..………… 474

8.4.5. Synthesis of current knowledge on ultrafine particles in

relation to motor vehicles ……………………………….… 477

8.5. Scientific challenges and the novel contribution of this PhD

study …………………………………………..…………… 478

8.6 Comparison of emission factors derived in this PhD study

with a selection of Canadian, European, UK and USA

emission factors …………………………………………………… 480

8.7. Future research focus ……………………………………………. 491

8.8. References ……………………………………………................ 497

xxii

LIST OF TABLES

Table 2.1. Estimates of total annual PM10 for South-East Queensland 53 (SEQ) and urban South-East Queensland related to 2004 Table 4.1. Percent contribution of N+A and C modes by mass to PM1, 160 PM2.5 and PM10 in South-East Queensland, Australia Table 4.2. International literature reviewed to identify the location of 162 the modes in a number of different environments worldwide for particle number size distributions Table 4.3. International literature reviewed to identify the location of 163 the modes in a number of different environments worldwide for particle surface area size distributions Table 4.4. International literature reviewed to identify the location of 163 the modes in a number of different environments worldwide for particle volume size distributions Table 4.5. International literature reviewed to identify the location of 164 the modes in a number of different environments worldwide for particle mass size distributions Table 5.1. Source of tailpipe particle emission factors examined in the 190 statistical analysis to derive average emission factors for different vehicle and road type combinations Table 5.1.2. Model variables examined in the statistical analysis to derive 196 average emission factors to use in transport modelling and health impact assessments, to quantify tailpipe particle emissions generated by motor vehicle fleets Table 5.1.3. Sample size of emission factors for different model variables 198 examined in the statistical analysis, listed by particle metric Table 5.1.4. Tailpipe particle emission factors for motor vehicles 205 considered the most suitable to use in transport modelling and health impact assessments, derived based on advanced statistical analysis in this study of 667 emission factors in the international published literature Table 5.2.1. Particle number model explanatory variables and average 235 particle number emission factors

xxiii

LIST OF TABLES

Table 5.2.2. Particle volume model explanatory variables and average 237 particle volume emission factors

Table 5.2.3. PM1 model explanatory variables and average PM1 240 emission factors

Table 5.2.4. PM2.5 model explanatory variables and average PM2.5 241 emission factors

Table 5.2.5. PM10 model explanatory variables and average PM10 243 emission factors

Table 6.1. Tailpipe particle emission factors for motor vehicles used to 275 develop particle number, PM1, PM2.5 and PM10 inventories presented in this study

Table 6.2. Particle emission inventories for the urban South-East 283

Queensland motor vehicle fleet for particle number, PM1, PM2.5 and PM10 on urban and urban-major roads

Table 6.3. Comparison of estimates of total annual PM10 for SEQ and 285 urban SEQ

Table 6.4. Modelled reductions in total particle emissions in urban 290 SEQ in the 24 hour average period Table 6.5. Modelled reductions in total particle emissions in urban 291

SEQ in the peak travel times and in the 24 hour average period

Table 6.6. Scenario 3: Average particle emission factors per 293

passenger per km for LDVs and buses in urban SEQ in the 24 hour average period

Table 6.7. Scenario 4A: Model variables and assumptions used to 295

predict particle number and particle mass emissions in urban SEQ in 2026

Table 6.8. Scenario 4B: Estimated total annual particle emissions in 296

urban SEQ in 2026, compared to the 2004 inventory, this study

xxiv

LIST OF TABLES

Table 7.1. The range of particle number emission factors reported for 341 nano and ultrafine size ranges Table 8.1. Précis of the principal findings of this PhD research and 461 their significance in terms of application Table 8.2. Comparison of Australian National Pollutant Inventory 481 (NPI), Australian Diesel NEPM Preparatory Work, and a selection of Canadian, European, UK and USA particle emission factors, with emission factors derived in this PhD study Table 8.3 Future studies recommended which use the data, 492 knowledge and methods developed in this PhD study

xxv

LIST OF FIGURES

Figure 1.1. The foci of the four major research components 11 of this PhD project Figure 4.1. Normalised number and volume size distributions in 157 South-East Queensland, Australia

Figure 4.2. Published modal location values relating to particle 165 size distributions for South-East Queensland, Australia and for a range of environments worldwide and metrics (n=600) Figure 5.2.1. Multiple comparison plot showing the nature of the 246 statistical relationship between the categorical model variables for different metrics Figure 7.1. Comparison of reported particle number 316 concentrations measured by CPC or DMPS/SMPS Figure 7.2. Mean and median particle number concentrations for different environments 376 Figure 8.1. Diagram of Research Activities 426

xxvi

THE STATEMENT OF ORIGINAL AUTHORSHIP

The work contained in this thesis has not been previously submitted for a degree

or diploma at any other educational institution. To the best of my knowledge and

belief, the thesis contains no material previously published or written by another

person except where due reference is made.

Signed: ……………………..

Date: ……………………..

1

CHAPTER 1. INTRODUCTION

1.1 DESCRIPTION OF THE SCIENTIFIC PROBLEM

INVESTIGATED

Particles emitted from motor vehicle fleets impact on our environment on a

number of scales, from the local scale, polluting areas on or near roads and in

busways and tunnels, to emissions dispersed across regions and by long range

transport across continents. They can also reach into the upper atmosphere, into

the troposphere and stratosphere and contribute to climate change effects and

dimming of the earth’s atmosphere.

At the global scale, aerosols produced from fossil fuel and biomass burning can

reflect solar radiation, leading to a cooling of our climate system (IPCC 2001);

and the effects of aerosols can cause a weakening of the hydrological cycle, which

impacts on the quantity and availability of fresh water (Ramanathan et al. 2001).

A reliable global inventory of aerosol emission rates, concentrations and lifetimes

is needed, as well as breakthroughs in our understanding about how very small

particles in aerosols modify our environment (Ramanathan et al. 2001).

The mechanisms associated with particle formation, and the resultant levels of

particle concentrations in different particle size ranges, and how these relate to

different particle sources, such as motor vehicles, are the subject of ongoing

research. Examining the location of modes in particle size distributions provides

an opportunity to identify the particle size/s associated with the maximum particle

2

concentrations in different environments and for different particle metrics.

Results of such examinations have the potential to aid source apportionment and

inform development of relevant air quality regulations and guidelines. To date, an

investigation of the location of modes in particle size distributions in different

environmental aerosols and for different particle metrics on the broader global

scale has not been comprehensively attempted.

Investigation of the levels of particulate matter emitted from individual vehicles

continue to be the subject of extensive research. However our knowledge about

the quantities of total particulate matter emitted from urban motor vehicle fleets,

including ultrafine particle emissions (diameters < 0.1 µm), are still the subject of

considerable uncertainty. This is because a comprehensive inventory of particles

emitted from a motor vehicle fleet does not currently exist anywhere in the world.

The current state of knowledge is that very little is known about the extent of total

particulate matter emitted by motor vehicle fleets. This means that in urban areas

where vehicles are a major source of particulate matter pollution, urban

populations around the world are being exposed to levels of particulate matter

pollution about which we have insufficient knowledge, and which have not been

comprehensively quantified.

The health effects associated with exposure to particulate matter, however, are

well-documented. There are known serious health effects associated with

exposure to particulate matter, and a number of epidemiological studies have

3

linked exposure with increases in hospital admissions, various respiratory and

cardiovascular diseases and mortality (Pope and Dockery 2006).

This gap in current knowledge about the extent of total particulate matter emitted

by motor vehicle fleets severely limits our ability to develop effective and

relevant ambient air quality standards, and strategies, such as land use and

transport planning, which can protect human health, the ecosystem, and our

earth’s atmosphere.

Development of comprehensive motor vehicle particle emission inventories,

including inventories for particle number, as well as investigations of particle

mechanisms and resultant modality within particle size distributions and how

these relate to different emission sources have the potential to inform

development of effective air quality and vehicle standards, and strategies to

monitor and control this major pollution source.

1.1.1. Developing comprehensive inventories of motor vehicle particle

emissions

Motor vehicles are major emitters of gaseous and particulate matter pollution, and

a dominant source of particulate matter pollution in urban areas. Developing

inventories of particle emissions provide a means for gaining a detailed

understanding of the extent of this major pollution source.

4

To date a comprehensive, size-resolved inventory of motor vehicle particle

emissions for a motor vehicle fleet, covering the full size range of particles

emitted and including particle number and different mass size fractions, is not

available in the literature. Quantifying particle number emissions, in particular,

pose a major challenge as these can be very difficult and costly to measure, hence

an extensive database of information is not currently available.

To develop particle emission inventories, emission factors are used which

quantify particle emissions originating from different vehicle types under varying

driving and road conditions. However, identifying the most suitable emission

factors to use in developing inventories is an extremely complex process because

of the very diverse range of techniques used to derive emission factors. Many

different measurement methods have been used, that have measured different

particle size ranges and been conducted in different parts of the world. A

multiplicity of factors need to be resolved in order to identify the most suitable

emission factors to use in transport modelling and air quality assessments.

Two very important characteristics of effective particle emission inventories for

motor vehicle fleets include:-

• Particle emission inventories needed to be size-resolved. This is because

particle size is a key characteristic of ambient particulate matter which

determines the likelihood of particles depositing in the human

respiratory tract and how deeply they are likely to lodge in the tract

(Morawska et al. 2008). Therefore, from a health effects perspective it

5

is very important that inventories of motor vehicle particle emissions are

size-resolved.

• Particle emission inventories need to quantify both particle number and

also emissions for different particle mass size fractions. This relates to

the fact that particles with diameters < 1 µm are prolific in terms of

their numbers but have little mass and are therefore measured in terms of

particle number; whereas larger-sized particles with diameters > 1 µm

have greater mass and are most effectively measured in terms of

different particle mass size fractions.

1.1.2. Combining knowledge from two different disciplines to develop

inventories

One of a number of complexities associated with developing inventories of fleet

particle emissions is the very significant amount of data required. This includes

transport modelling data, where traffic volumes are assigned to different road

links in a study area; and derivation of suitable emission factors for different

vehicle and road type combinations that are relevant to the vehicle mix in the fleet

being modelled.

Developing inventories of emissions provide data that enable comparisons to be

made between quantified emissions and current air quality standards. This data

also aids identification of emission hotspots, informs development of air quality

guidelines and regulations, and health impact assessments. It also provides

invaluable information for land use and for transport planners to guide their

planning and decision-making.

6

1.1.3. Why particle number emission inventories are important

Most particles emitted from motor vehicles are ultrafine size (Morawska 2003)

and current ambient air quality standards are ineffective for controlling ultrafine

particles as they are mass-based and prescribe measurement of PM2.5 and PM10

(particles with aerodynamic diameters < 2.5 µm and < 10 µm respectively).

Ultrafine particles are more appropriately measured in terms of particle number

emissions, because they have little mass, and are prolific in terms of their

numbers.

Currently ambient air quality standards in terms of the concentration of particle

number emissions do not exist anywhere in the world, which means that the

majority of motor vehicle particle emissions are not regulated. In addition, a

detailed emission inventory for particle number concentration is not available in

the literature (Jones and Harrison 2006). Hence, to address this major gap in our

knowledge it is extremely important to develop inventories which include particle

number, in addition to different particle mass size fractions.

Another important reason for developing particle number inventories is that at

present, in terms of health effects due to exposure to particle emissions, the focus

of current scientific debate is centred on the premise that particle number is more

directly related to health effects than particle mass (ECJRC 2002). Development

of particle number inventories and further epidemiological research will inform

this very important debate.

7

Although considerable toxicological evidence exists on the harmful effects to

human health of exposure to ultrafine particles, current epidemiological evidence

is insufficient to enable a conclusion to be reached on an exposure-response

relationship (WHO 2005). However, two organizations in Europe are taking a

proactive stance to control ultrafine particles by the introduction of particle

number standards, which measure particle emission rates. Based on the

recommendation of the UNECE-GRPE Particulate Measurement Program, the

European Commission is adding a particle number limit and new measurement

procedure to its EURO V/VI emissions standards for light duty diesel vehicles

and to its EURO VI emissions standard for heavy duty diesel vehicles relating to

solid particles (European Union 2007; Commission of the European Communities

2007 a,b; http://ec.europa.eu/index_en.htm). The Swiss Agency for the

Environment, Forests and Landscape has proposed the introduction of a particle

number standard for diesel-fuelled passenger vehicles for solid particles in the

0.02-0.30 µm size range (AQEG 2005).

1.1.4. Designing environmentally-sustainable transport systems

To design and manage transport systems which are environmentally-sustainable

more detailed information is needed about total particulate matter emitted from

fleets to guide land use and transport planning.

One of the problems of major concern in our current transport systems are the

relatively high emission rates of heavy duty diesel vehicles. For example, in

terms of particle number emissions diesel-fuelled vehicles have been found to

emit an order of magnitude higher particle emissions than petrol-fuelled vehicles

8

(Morawska et al. 2004); and although they generally constitute a small proportion

of the fleet in terms of number, their high emission rates make them a major

pollution source. In addition to introducing technologically-based solutions such

as after-treatment devices, eg., particle filters, it is very important to limit

population exposure to these high emitters by monitoring and controlling their

route choices situated close to populations, and to devise and implement lower

polluting options for freight movement, particularly given that diesel exhaust has

been declared a carcinogen in Switzerland (Swiss Clean Air Act 2000;

www.dieselnet.com/standards/ch/).

1.1.5. Examination of modality within particle size distributions and its

potential as a basis for developing air quality regulation

A mode may be defined as a peak in the lognormal function of the number or

mass distribution of an atmospheric aerosol (John 1993). The location of the

mode in a particle size distribution relates to the particle size/s associated with the

maximum particle concentration/s in the aerosol being studied, and are influenced

by particle mechanisms and the dominant sources of pollution.

Most aerosol particle size distributions are not characterised by bell-shaped

distributions of particle size, they are generally not normally distributed and tail

off with increasing particle size, hence log-normal distributions are considered

more appropriate for characterising airborne particle distributions (Ruzer and

Harley 2004). For example, a substantial number of ambient aerosols measured

in terms of number size distributions can be described as the sum of different

9

modes, with each mode being log-normally distributed (Hinds 1999; Seinfeld and

Pandis 1998).

Examination of two specific aspects related to modality within particle size

distributions have the potential to provide important knowledge for understanding

particle formation mechanisms, and atmospheric processes as they relate to

particles. Investigating these aspects could aid source apportionment, inform

exposure and health risk assessments and guide the development of relevant and

effective air quality regulations and guidelines.

These aspects relate to examining the location of modes in aerosol particle size

distributions, and analysing the relative contributions of particle concentrations to

different modal particle size ranges (known as nucleation, accumulation and

coarse modes, which are generally considered to relate to particles with diameters

of < 0.1 µm, 0.1-1 µm and > 1 µm respectively).

It has been shown that a clear separation exists at around 1 µm, or somewhat

above, between the accumulation and coarse modes in ambient air particle size

distributions, where the mass of particles belonging to these two modal particle

size ranges is at a minimum (Lundgren and Burton 1995). The major proportion

of anthropogenic pollution sources are combustion-related and generate particles

with diameters < 1 µm (Jamriska and Morawska 2003); hence a study of modal

location values and particle concentrations related to the 1 µm size range (or

thereabouts) is considered of global significance.

10

1.2. THE MAJOR COMPONENTS OF THIS PHD RESEARCH

The four major research components of this PhD project are depicted in

Figure 1.1. These relate to:-

1. Examination of modality within particle size distributions as a basis

for developing air quality regulation.

2. Derivation of a comprehensive set of particle emission factors for motor

vehicles.

3. Development of a particle number and particle mass emissions inventory

for the urban fleet in South-East Queensland.

4. A review and synthesis of existing knowledge on ultrafine particles

in ambient air, with a specific focus on particles originating from

motor vehicles.

The examination of modality within particle size distributions provided important

contextual information to inform derivation of suitable particle emission factors for

use in development of the motor vehicle emissions inventory. Outputs from the

inventory development, in turn, complemented modality examination findings. The

review and synthesis of current knowledge on ultrafine particles in ambient air

contributed knowledge to both the examination of modality within particle size

distributions and to the derivation of suitable particle emission factors.

11

Is modality useful for developing air quality regulations?

Which emission factors are the mostsuitable to use in transport modelling?

Emission Factors: derivation of a comprehensive set of particle emission factors to estimate fleet emissions

What temporal and spatial characteristics are important in assessing human exposure to ultrafine particles?

Do discrepancies exist between the outcomes of different particle number measurement techniques?

Modality within particle size distributions

Development of a comprehensive inventory of motor vehicle fleet particle emissions for urban South-East Queensland (SEQ)

How much particulate matter does the urban SEQ fleet contribute in terms of particle number and particle mass?

How might this quantification be validated?

Scenario modelling: What are emission levels likely to be in the future? What changes in travel demand might effect reasonable reductions in regional particle emission levels?

Figure 1.1 The foci of the four major research components of this PhD project

Review & synthesis of current knowledge on ultrafine particles, with a specific focus on vehicle emissions

Would a PM1 mass standard suit most environments?

How confident can we be about the values of the derived emission factors?

Are vehicles a significant source of ultrafine particles in populated urban areas, and are ultrafine particle inventories available for motor vehicles?

Are ultrafine particles in urban areas the most likely target for future air quality regulations in relation to particle number?

Do contributions to PM1, PM2.5 and PM10 mass vary for different environments?

To what extent do the derived emission factors explain the variation in published emission factors?

12

1.3. THE OBJECTIVES OF THE STUDY

The overall objectives and specific aims of this study are to:-

1. Examine a wide range of different environments in urban South-East

Queensland, Australia to identify the relationship between fractional

contribution of mass from different sources and modes in particle size

distribution to PM1, PM2.5 and PM10.

1.1 Investigate the characteristics of modality within particle size

distributions in marine-influenced, modified background, suburban

background, traffic-influenced, urban-influenced and vegetation

burning environments in South-East Queensland; and examine the

relationship between the fractional contribution of mass from

different sources and modes in particle size distributions to PM1,

PM2.5 and PM10.

13

2. Examine the relevance of using modality within particle size

distributions as a basis for developing air quality regulations;

and ascertain whether PM1 and PM10 would be a more effective

combination of mass standards than the current standards of

PM2.5 and PM10 for controlling ambient particles generated from

mechanical and combustion-related processes.

2.1 Examine the location of the mode in a wide range of

worldwide environments, including in traffic-influenced

environments, for different particle metrics to assess its

relevance as a basis for developing air quality regulations.

2.2 Determine whether a clear and distinct separation occurs

between the modes at around 1 µm in different environments

throughout the world, and assess the suitability of a PM1 mass

ambient air quality standard for these different worldwide

environments, including for traffic-influenced environments.

14

3. Derive a comprehensive set of particle emission factors for motor

vehicles that are suitable to use in transport modelling to quantify

tailpipe particle emissions from motor vehicle fleets for different

particle sizes and particle metrics.

3.1 Derive the most suitable particle emission factors to use in transport

modelling and health impact assessments for different vehicle and

road type combinations, and for different particle sizes and metrics,

to enable development of size-resolved inventories for motor vehicle

fleets that cover the full size range of particles emitted and which

include quantification of both particle number and different mass

size fractions.

15

4. Develop a comprehensive road-link based inventory of tailpipe

particle emissions generated by the urban South-East Queensland

motor vehicle fleet covering the full size range of particles emitted.

4.1 Develop an inventory of tailpipe particle emissions emitted from

the urban South-East Queensland fleet which quantifies the total

contribution to particulate matter pollution for different particle

size ranges and vehicle and road type combinations.

4.2 Validate the urban South-East Queensland inventory (4.1 above)

with other relevant models and inventories derived for the region.

4.3 Conduct scenario analysis modelling using the urban South-East

Queensland inventory data (4.1 above) to test the air quality

implications of likely future scenarios related to passenger and

freight vehicle travel demand in terms of particle emission levels;

and develop an estimate of fleet emissions in 2026.

16

5. Review and synthesize existing knowledge on ultrafine particles in ambient air, specifically related to emissions generated by motor vehicles.

5.1 Examine and synthesize current knowledge on ultrafine particles in

ambient air, with a specific focus on emissions that originate from

motor vehicles.

5.2 Review and analyse instrumental techniques used for ultrafine

particle measurement, and identify and examine any differences in

outcomes produced by this instrumentation.

5.3 Examine ultrafine particle emission levels and their characteristics

as a function of vehicle technology, fuel used, and after-treatment

devices applied, with a specific focus on secondary particle

formation in urban environments resulting from semi-volatile

precursors emitted by motor vehicles.

5.4 Review existing knowledge on the spatial and temporal variation in

ultrafine particle concentrations, long term monitoring and the

existence of any inventories available for particle number and

ultrafine particles for motor vehicle fleets.

17

5.5. Review current knowledge on ultrafine particle chemical

composition, and the relation between ultrafine particles

and gaseous pollutants.

5.6. Examine the extent of contributions of ultrafine particle

concentrations to different environments.

5.7. Assess the implications of existing knowledge related to the

characteristics of ultrafine particles and dynamics in the air, in

the context of human exposure and epidemiological studies, and

in relation to management and control of particles in vehicle-

affected environments.

18

1.4 ACCOUNT OF SCIENTIFIC PROGRESS LINKING THE SCIENTIFIC PAPERS

This thesis contains a collection of papers that have been published in refereed

journals. The focus and results presented in these papers are précised below.

Paper One:

Examination of modality within particle size distributions as a basis for

developing air quality regulation

The study reported in the first paper (presented in Chapter 4) focused on

examining the suitability of using modality within particle size distributions for

developing air quality regulations; and investigated whether PM1 and PM10 mass

standards may be a more effective combination of standards than PM2.5 and PM10

for controlling mechanical and combustion-generated particles, such as emitted

from motor vehicles.

It emphasized developing an understanding of the differences between locations

of the modes in different worldwide environments for different particle metrics,

and the relative contribution of particle mass to nucleation, accumulation and

coarse particle modes in urban South-East Queensland.

19

It aimed to produce a comprehensive review of the location of modes for different

particle metrics in a wide range of different environments around the world. A

second aim of the study was to examine whether a clear separation exists between

the modes at around 1 µm and determine whether a PM1 mass standard is relevant

for a range of different environments worldwide.

The study presented evidence that modal locations in particle size distributions

have the potential to be used as a basis for developing air quality regulations, and

provide useful information about contributions from different pollution sources

and particle mechanisms. It also presented evidence that a combination of PM1

and PM10 mass standards may provide a more suitable and discerning

combination of particle mass standards than the current mass standards of PM2.5

and PM10 for combustion and mechanically-generated sources, such as motor

vehicles.

The main conclusions of the study were:-

(i) PM10 measurements provided information mainly about the coarse

mode generated from mechanical processes (eg., particles emitted

from tyre wear or resuspended by motor vehicle traffic) but not about

motor vehicle tailpipe emissions.

20

(ii) PM2.5 measurement data can relate to a mix of nucleation and

accumulation mode particles (combustion-generated) and coarse

mode particles (mechanically-generated), making source

apportionment very complex and suggesting that PM2.5 is an

inadequate basis for standards.

(iii) PM1 measurement data related to nucleation and accumulation mode

particles and enabled a much clearer distinction to be made between

combustion and mechanically-generated aerosol contributions. This

finding provides evidence to support the view that PM1 and PM10

mass standards would be more desirable from the legislation point

of view than the current mass standards of PM2.5 and PM10.

(iv) The study suggested that more discussion is needed to consider the

best combination of particle mass and number concentration

standards for a major source of particulate matter pollution such as

motor vehicle fleets, for example, the introduction of particle

number standards for submicrometre and smaller particle size

ranges (eg., ultrafine particles).

These results make an important contribution to developing an understanding

of the value of examining modes within particle size distributions as a basis for

development of air quality regulations and for source apportionment.

21

Paper Two:

Derivation of a comprehensive set of particle emission factors for motor

vehicles.

The second paper (Chapter 5.1) describes the derivation of a comprehensive set

of particle emission factors for motor vehicle tailpipe emissions which can be

used in transport modelling and health impact assessments to quantify particle

emissions from different vehicle and road type combinations, covering the full

size range of particles emitted and including particle number and different particle

mass size fractions.

The specific objective of the study was to derive the most suitable particle

emission factors, based on statistical analysis of more than 600 particle emission

factors published in the international literature. In order to achieve this goal, five

statistical models were developed that estimated average emission factors for

particle number, particle volume, PM1, PM2.5 and PM10.

From the outputs of these five statistical models, the final set of particle emission

factors were selected which are recommended as the most suitable to use in

transport modelling and health impact assessments. These average particle

emission factors, and their 95% confidence intervals, relate to different vehicle

and road type combinations for particle number, particle volume, PM1, PM2.5 and

PM10 for light duty vehicles, heavy duty vehicles and buses. The outputs of these

statistical models are presented in Chapter 5.2.

22

The main study activities included:-

(i) Conducting an extensive review of all published particle emission

factors for different motor vehicle types in the international

literature.

(ii) Identifying suitable model variables to use in developing

statistical models to predict average emission factor values for

different particle metrics.

(iii) Developing statistical models for different particle metrics to produce

average particle emission factors.

(iv) Identifying the most suitable emission factors to use in transport

modelling and health impact assessments, based on examination of the

statistical characteristics of average particle emission factors produced

by the statistical models.

23

Information concerning which motor vehicle particle emission factors are suitable

to use in transport modelling and health impact assessments is currently patchy

and ill-defined. The results of this study advance scientific knowledge of the most

suitable emission factors to use to develop comprehensive, size-resolved

inventories of tailpipe particle emissions from motor vehicle fleets for both

particle mass and particle number, covering the full size range of particles

emitted.

Most importantly, the emission factors derived in this study have application for

urban regions in developed countries, and have particular application for regions

which lack measurement data, or funding to undertake measurements, or where

experimental data is of insufficient scope.

The study also identified gaps in our knowledge and found that very limited data

exists relating to emission factors for particle volume, particle surface area, PM1,

brake and tyre wear, road grade, engine power, on-road bus measurements, and

vehicles travelling at speeds < 50 km/hr. Information and methods that can be

used to discriminate resuspended road dust from tailpipe emissions, particularly

for PM2.5 and PM10 road emission studies, were also found to be limited.

24

Paper Three:

Development of a particle number and particle mass emissions inventory for

the urban fleet in South-East Queensland.

The third paper (presented in Chapter 6) presents the first published

comprehensive inventory for a motor vehicle fleet of particle emissions generated,

covering the full size range of particles emitted, and includes quantification of

emissions in terms of particle number and different particle mass size fractions.

The specific objectives of the study were to:-

(i) Develop a motor vehicle particle emissions inventory for urban South-

East Queensland that included quantification of particle number, PM1,

PM2.5 and PM10 emissions for light and heavy duty vehicles and buses.

(ii) Model the particle emission implications of different proportions of

passengers travelling in light duty vehicles and buses, and to derive an

estimate of vehicle fleet particle emissions in the year 2026.

In order to achieve the goals of this study, particle emission factors for

different vehicle and road type combinations were combined with transport

modelling data to quantify emissions relating to model links in the study

region classed as urban and urban-major roads. Different scenarios were

modelled which involved shifting proportions of light duty vehicle passengers

to new buses added to the network to assess the impact on particle emission

levels.

25

An estimate of fleet emissions in 2026 was modelled which considered the

anticipated freight task in 2026, likely increases in vehicle kilometres travelled by

different vehicle types, possible improvements in vehicle technologies (leading to

reductions in particle mass and particle number emissions), likely changes in

transport mode choice and fleet composition, and the introduction of new, lower

emitting vehicle types.

The results reported in the paper were based on government prototype data from

the Brisbane Strategic Transport Model for 2004 and used vehicle kilometres

travelled data, and excluded consideration of specific origin and destination trip

data.

The study results advance scientific knowledge by presenting the first

comprehensive inventory of motor vehicle particle emissions that has been

published, and which includes particle number. The research work also

demonstrated how small changes in transport mode and passenger occupancy

rates can lead to reductions in particle emission levels, and provided an estimate

of expected emissions in the region in 2026.

26

Paper Four:

A review and synthesis of existing knowledge on ultrafine particles in

ambient air, with a specific focus on particles generated by motor vehicles.

The fourth paper (shown in Chapter 7) presents a synthesis of existing

knowledge on ultrafine particles in air, focusing on particles originating from

motor vehicles.

The main study activities were to:-

(i) Review current knowledge on ultrafine particles as they relate to motor

vehicles, including the extent of their contribution to urban

environments; and analyse instrumentation techniques used to measure

ultrafine particles to examine any differences in outcomes.

(ii) Examine ultrafine particle emission levels and their characteristics in

terms of different vehicle technologies, fuels, and after-treatment

devices used, with a focus on secondary particle formation in urban

environments.

27

(iii) Review knowledge on the characterization of the temporal and spatial

variation in concentration of ultrafine particles and long term

monitoring; examine existing knowledge on particle chemical

composition, and the relationship between ultrafine particles and

gaseous pollutants.

(iv) Examine any differences in concentrations of ultrafine particles in a

range of different environments, and investigate whether any

inventories are available for particle number and ultrafine particles for

motor vehicle fleets.

(v) Review existing knowledge on the characteristics of ultrafine particles,

particle mechanisms and dynamics in the air affecting these

concentrations, and their relationship in terms of human exposure

assessment and epidemiological studies, and the control and

management of particles in environments affected by vehicle

emissions.

The study found that motor vehicles in populated urban areas are a significant

source of air pollution and of ultrafine particles, and that ultrafine particles are a

likely target for future air quality regulation in terms of particle number in urban

areas.

28

It found that no standardized methods exist for measuring particle number, and

that discrepancies exist between the outcomes of different instrumentation used to

measure particle number. These discrepancies need to be borne in mind when

examining the quantification of particle number in different studies. The lack of a

standard approach for measuring particle number has particular significance for

epidemiological studies and human exposure assessment.

The review found that ultrafine particle concentrations can differ between clean

and vehicle-influenced environments by as much as over two orders of

magnitude, which has implications for exposure assessment. Large uncertainties

were found in relation to vehicle emission factors for particle number and other

particle size ranges, and no emission inventories were found for ultrafine particles

or particle number for motor vehicle fleets. In addition, it was found that limited

data is available on long term monitoring of ultrafine particle concentrations in

urban environments and related to ultrafine particle composition and chemistry,

which can be influenced by many vehicle-related factors and post-formation

processes. Hence, a better knowledge of ultrafine particle chemistry in different

environments is needed.

The paper also discusses the need to include consideration of secondary particle

formation in vehicle exhaust plumes and particle formation by nucleation, and

their possible relevance if particle number regulation is proposed, as well as

consideration of location-specific meteorological factors which can influence

these formations.

29

1.5. THE IMPORTANT AND NOVEL CONTRIBUTION OF THIS

PHD RESEARCH

This PhD study contributes new knowledge and understanding to the field in

four different knowledge domains:-

(i) Firstly, a new method has been developed for deriving comprehensive

inventories of motor vehicle particle emissions. This used a novel

approach that involved combining knowledge from two distinctly

different disciplines – from aerosol science and transport modelling.

(ii) Secondly, the work developed new concepts for identifying suitable

particle emission factors to use in developing inventories for different

particle sizes and particle metrics related to different vehicle and road

type combinations, that included rigorous statistical analysis of a very

large set of measurement data sourced from the international published

literature.

(iii) Thirdly, a new approach was developed for examining modality within

particle size distributions, which provided valuable information on

particle mechanisms and contributions from different environmental

sources to different mass size fractions. This approach also identified

that a new particle mass standard, PM1, would be suitable for the

majority of worldwide environments, and found that a combination of

PM1 and PM10 standards have the potential to provide a more

discerning set of ambient particle mass emission standards than the

30

present standards of PM2.5 and PM10 for discriminating between

combustion and mechanically-generated particles, such as emitted

from motor vehicles.

(iv) Fourthly, this work is the first inventory of motor vehicle particle

emissions that has been published, and which includes quantification

of particle number emissions. The work also developed new

approaches for modelling future scenarios of travel demand and their

particle emission implications.

(v) Fifthly, the work made an important contribution by presenting a

review and synthesis of existing knowledge on ultrafine particles in

ambient air as they relate to vehicle emissions, and found that motor

vehicles make a significant contribution to both air pollution and

ultrafine particles in populated urban areas. The work identified

discrepancies between the outcomes of instrumentation that measure

ultrafine particle concentrations; and the absence of a standard process

for measuring particle number; as well as gaps in our knowledge in

relation to vehicle emission factors for different particle size ranges

and for particle number. Few studies were found related to the

composition and chemistry of ultrafine particles and long term

monitoring of ultrafine particles; and no inventories were found for

particle number of ultrafine particles for motor vehicle fleets in the

published literature. The work identified key areas which require

31

further research to enable control of ultrafine particles in populated

urban areas and formulation of appropriate air quality regulation.

This PhD research has produced a complementary toolkit of data, knowledge and

methods which can be used to quantify, monitor and control vehicle fleet particle

emissions, and has identified important, future areas of research needed for the

control of ultrafine particles in populated urban areas, and for development of

possible future particle number regulation. These include methods for source

apportionment, developing air quality regulation and for quantifying fleet particle

emissions. The work presents the first inventory of particle emissions for a fleet

that has been published, and a method and comprehensive set of particle emission

factors which can be used to quantify urban fleet emissions in the developed

world. Vehicle emissions were found to be the most common and significant

source of air pollution in populated urban areas, and a significant source of

ultrafine particles, emphasizing the importance of considering ultrafine particles

as a target for future air quality regulation in relation to particle number in

populated urban areas.

32

1.6. REFERENCES

AQEG., 2005. Particulate Matter in the UK. London, Department for

Environment, Food and Rural Affairs.

DieselNet Emissions Standards, Switzerland. www.dieselnet.com/standards/ch/.

Date verified 20 February 2008.

ECJRC., European Commission Joint Research Centre) 2002. Guidelines for

concentration and exposure-response measurement of fine and ultrafine

particulate matter for use in epidemiological studies. EUR 20238 EN 2002.

Editors D Schwela, L. Morawska, D. Kotzias, European Commission, Italy.

European Commission. http://ec.europa.eu/index_en.htm. Date verified 28 July

2008.

Hinds, W.C., 1999. Aerosol Technology, 2nd edn., Wiley, New York,

IPCC., (Intergovernmental Panel on Climate Change) 2001. Climate Change

2001: The Scientific Basis, Contribution of Working Group I to the Third

Assessment Report of the Intergovernmental Panel on Climate Change

Cambridge. United Kingdom and New York, Cambridge University Press.

Jamriska, M., Morawska, L., Thomas, S., Congrong, H., 2004. Diesel Bus

Emissions Measured in a Tunnel Study. Environmental Science & Technology

38(24), 6701-6709.

John, W., 1993. The characteristics of environmental and laboratory generated-

aerosols, in: Willeke and Baron (Eds.), Aerosol measurement: Principles,

techniques and applications,Van Nostrand Reinhold, New York, 55.

Jones, A.M., Harrison, R.M., 2006. Estimation of the emission factors of particle

number and mass fractions from traffic at a site where mean vehicle speeds vary

over short distances. Atmospheric Environment 40(37), 7125-7137.

33

Lundgren, D.A., Burton, R.M., 1995. Effect of particle size distribution on the

cut point between fine and coarse ambient mass fractions, Inhalation Toxicology

7 (1), 131-148.

Morawska, L., 2003. Chapter 3: Motor Vehicle Emissions as a Source of Indoor

Particles in, Morawska-Salthammer (eds). Indoor Environment, Wiley-VCH, 297-

318.

Morawska, L., Keogh, D.U., Thomas, S.B., Mengersen, K., 2008. Modality in


quality regulation. Atmospheric Environment 42(7), 1617-1628.

Morawska, L., Moore, M.R., Ristovski, Z.D., 2004. Health Impacts of Ultrafine

Particles - Desktop Literature Review and Analysis. Department of the

Environment and Heritage, September, Canberra.

Parrish, D.D., 2006. Critical evaluation of US on-road vehicle emission

inventories. Atmospheric Environment 40(13), 2288-2300.

Pope, C.A., Dockery, D.W., 2006. Health Effects of Fine Particulate Air

Pollution: Lines that Connect. Journal of the Air & Waste Management

Association 56(6), 709-732.

Ramanathan, V., Crutzen, P.J., Kiehl, J.T., Rosenfeld, D., 2001. Aerosols,

Climate, and the Hydrological Cycle. Science's Compass 294, 2119-2124.

Ruzer, L.S., Harley, N.H. 2004, Aerosols Handbook: Management, Dosimetry

and Health Effects, CRC Press, Florida, USA.

Seinfeld, J.H., Pandis, S.N. 1998. Atmospheric Chemistry & Physics, Wiley-

Interscience, New York.

34

Swiss Clean Air Act 2000, LRV 00, Appl. 1, 83, Schweizerische

Luftreinhalteverordnung, 16 December, 1985, amended 28 March 2000.

WHO (2005). "Guidelines for Air Quality." World Health Organization, Geneva.

35

CHAPTER 2. LITERATURE REVIEW

2.1. INTRODUCTION

This chapter summarises a review of the literature on particles emitted from motor

vehicle tailpipes and the inventories that have been developed to quantify these

emissions in urban areas.

The most relevant topics related to the research work were reviewed in depth and

focused on the current state of knowledge of:-

(i) the characteristics and nature of particulate matter, and modality

within particle size distributions;

(ii) particulate matter emitted from motor vehicle tailpipes;

(iii) methods for developing motor vehicle emission inventories;

(iv) current local and international motor vehicle inventories;

(v) present ambient air quality standards used for control of particle

emissions.

Review of literature on the health effects of particulate matter exposure and

transport models was more general.

36

The characteristics of motor vehicle particle emissions, including modality within

particle size distributions, the health effects of exposure and current air quality

standards are introduced in Chapter 2.2. Vehicle emission inventories and local

models are discussed in Chapter 2.3. Transport models are generally discussed in

Chapter 2.4. An estimate of motor vehicle fleet particle emissions prepared for the

UK is reviewed in Chapter 2.5. Chapter 2.6 discusses the difficulties associated

with identifying suitable particle emission factors to use in developing motor

vehicle particle emission inventories. Chapter 2.7 provides a summary, and the

final chapter, Chapter 2.8, discusses current knowledge gaps and conclusions

from the review.

2.2. CHARACTERISTICS OF MOTOR VEHICLE PARTICLE

EMISSIONS

Motor vehicle tailpipe emissions comprise pollutants in particle and gaseous

forms which are made up of many compounds that are complex in terms of their

chemical composition. Many of these compounds have been found to affect

human health (Morawska 2003b). These emissions have an impact on a range of

scales, from micro to macro environments, ranging from areas in close proximity

to roads, to regional airsheds. They can also be a major source of pollution in

busways, tunnels and in public transport interchanges.

Globally, their effect on earth’s climate and upper atmospheres, including the

troposphere and stratosphere, has neither been quantified, nor is it well

understood.

37

While motor vehicles are also major emitters of gaseous pollution; the current

focus of scientific air quality debate is centred on particulate matter, in particular

on ultrafine particles (particles with diameters < 0.1 µm measured in terms of

number concentration) (ECJRC 2002). The hazard level associated with inhaled

particles is dependent upon the chemical composition of particles and where they

deposit within the respiratory system (Hinds 1982). The aerodynamic size of

particles determines where in the airways they are likely to deposit (Ferin et al.

1990). For these reasons it is vital that quantification of particulate matter

pollution be derived in terms of different particle size ranges.

This literature review is restricted to discussion on motor vehicle tailpipe particle

emissions, because very little information is available in the international

literature on particles produced from brake and tyre wear, nor on methods that

enable discrimination of road dust particles from particles emitted by motor

vehicle tailpipes. Measurement data of ambient particle size distributions in terms

of particle surface area are also rare.

2.2.1. The nature of particle emissions

Most anthropogenic pollution sources are combustion-related and generate

particles with diameters < 1 µm (Jamriska and Morawska 2003). Combustion

source particles, such as those emitted by motor vehicles, are found mainly in the

ultrafine size range (diameters < 0.1 µm) (Morawska 2003). Inhalation and

deposition of particles deep in the alveoli of a human lung can be very detrimental

to human health (Seaton et al. 1995), and cause serious health effects.

38

Particle diameter expresses a particle’s settling velocity, which can be used to

predict where in the respiratory tract a particle may deposit (Morawska et al.

2004). Diffusional deposition is the most important mechanism for deposition in

the lung of the smallest particles (ultrafine size).

Particles can be measured by a number of particle metrics, including particle

mass, number count, volume and surface area. Particles larger than 10 µm tend to

have atmospheric lifetimes that are relatively short (Harrison et al. 2000) and are

of lesser significance from the health point of view since they are mostly removed

by the upper respiratory tract.

Depending on the instrumentation used and their ranges of measurement, particles

in the larger size range that have diameters of 1-10 µm are generally measured in

terms of particle mass and classified by their aerodynamic diameter. Whereas,

particles with diameters < 1 µm are generally measured in terms of particle count

(also termed particle number) and classified according to their equivalent

diameter (electrical mobility diameter, diffusion diameter etc). Ultrafine

particles are very small and numerous in terms of relative number, they have

little weight (mass) and therefore particle number is the most relevant particle

metric to use to measure these sized particles.

39

2.2.2. Motor vehicle particle emissions

Particle size distributions can be presented in terms of particle mass or particle

number distributions, and in terms of particle number size distribution the

majority of airborne particles are ultrafine size (Morawska and Salthammer

2003). In urban environments, where the major source of pollution is motor

vehicle emissions, more than 80% of particle emissions in terms of particle

number are ultrafine size (Morawska et al. 1998a).

Many studies have shown conclusively that the major source of ultrafine particle

pollution in urban environments is from motor vehicle emissions (Harrison et al.

1999; Shi and Harrison 1999; Shi et al. 1999; Shi et al. 2001; Wahlin et al. 2001).

In environments affected by motor vehicle emissions, ultrafine particles can

account for levels of up to an order of magnitude higher than those in natural

environments, and knowledge about the chemical composition of these particles is

very limited as few studies have investigated these concentrations in ambient air

in different environments (Morawska et al. 2008b).

The diameter of particles emitted from motor vehicles can range from around

0.003 to 10 µm; where 0.003 µm is the lowest size range currently able to be

measured. Most motor vehicle particle emissions are in the ultrafine size range.

By comparison, the diameter of a human hair can range from 50-100 µm (Willeke

and Baron 1993). Particles emitted from diesel engines tend to be in the size range

0.02-0.130 µm (Kittelson 1998; Morawska et al. 1998b; Harris and Maricq 2001;

Ristovski et al. 2006) and those emitted from petrol engines in the 0.02-0.06 µm

size range (Harris and Maricq 2001; Ristovski et al. 2006). At the time of this

40

study the literature review revealed that the majority of heavy duty vehicles

(HDVs) were diesel-fuelled and light duty vehicles (LDVs) petrol-fuelled.

As motor vehicle particle emissions span such a wide size range, in order for

inventories to be comprehensive they need to quantify both particle number and

particle mass emissions for different size fractions. As ultrafine particles emitted

from motor vehicles are a significant source of anthropogenic pollution in urban

areas, it is extremely important that they be included in inventories and be

controlled by air quality standards.

2.2.3. Diesel particle emissions

Diesel particulate matter is made up of many small particles that have very little

mass, and the relatively few particles of a larger size account for most of its total

particle mass. Its chemical and physical properties, how they form in the cylinder

of an engine and their effect on human health are not fully understood and diesel

emission regulations exist worldwide to regulate diesel particles (Morawska et al.

2004). In the US, diesel exhaust has been declared a probable human carcinogen

(Zhu 2003); and in Switzerland diesel exhaust is classified as a carcinogen

(www.dieselnet.com/standards/ch). Diesel particle emissions are an important

emission source that requires dedicated effort in terms of its control and

management.

Diesel vehicles release over an order of magnitude more particles, in terms of

particle number, than petrol-fuelled vehicles (Morawska et al. 2004) and a

significant number of these particles are in the ultrafine size range (Morawska et

41

al. 1998b; Ristovski and Morawska 1998). It was found that particle emissions

from light and heavy duty vehicles measured in a tunnel by Kirchstetter et al.

(1999) revealed that heavy duty diesel trucks emitted considerably more particle

number, fine particles, black carbon and sulphate mass per unit of fuel mass

burned as compared to light duty vehicles, in the order of 15-20, 24, 37 and 21

times more respectively. One study found that when petrol vehicles are driven

with high loads (ie., during acceleration) or at very fast speeds (eg., at around 120

km/hr), their particle number emissions can be similar to those emitted by diesel

vehicles (Graskow et al. 1998).

2.2.4. Health effects associated with exposure to particles

A number of epidemiological studies have linked particle exposure with

increases in hospital admissions, mortality, and various cardiovascular and

respiratory diseases (Pope and Dockery 2006). An association has also been

found with effects such as lung cancer (Pope et al. 2002) and heart attacks

(Brook et al. 2000). Research has shown that particles can penetrate the cell

membranes, enter the bloodstream, and even reach the brain (Oberdoerster et

al. 2004); and there are some indications that they can induce inheritable

mutations (Somers et al. 2004).

When considering exposure in terms of dose-response, it has been suggested that

ultrafine number concentration and surface area may be more appropriate metrics

than particulate mass (Young and Keeler 2004). Ultrafine studies have shown

that particle number and surface area, not particle mass, are the most suitable

particle characteristics to evaluate the potential biological effect of ultrafine

42

particles (Morawska et al. 2004). Studies have also suggested that particles

measured in terms of their number concentration may be more suitable than

particles measured in terms of mass concentrations to assess health effects, which

has raised concerns that large particle number concentrations near freeways may

be adversely affecting the health of people situated in close proximity (Zhang and

Wexler 2004). However, it remains unclear whether particle mass, number,

surface area concentration, chemical composition or a combination of these

properties pose the greatest health risk (Zhang 2004).

Present scientific debate is focused on the notion that particle number is more

directly related to health effects than particle mass. Based on exposure-health

response relationships derived in epidemiological studies of exposure to airborne

particles measured in terms of mass concentrations, the World Health

Organization (WHO) has set new particulate matter guidelines with annual mean

values for PM2.5 and PM10 of 10 and 20 µg m-3 respectively (WHO 2005). PM2.5

and PM10 fractions are mass concentration of particles with aerodynamic

diameters smaller than 2.5µm and 10µm respectively. These guidelines were

based on an American Cancer Society study (Pope et al. 2002) and represent the

lowest end of the range across which have been observed significant effects on

survival (WHO 2005). At present the available body of epidemiological evidence

is not sufficient to enable a conclusion to be reached on the exposure-response

relationship of ultrafine particles, hence WHO has stated that “therefore no

recommendations can be provided as to guideline concentrations of ultrafine

particles at this point in time” (WHO 2005).

43

2.2.5. Current air quality standards to control particulate matter

Current national air quality standards in countries across the world are based on

particle mass concentration, not particle count, volume or surface area, and are

restricted to PM2.5 and PM10 fractions. These standards were based, in part, on a

scientific basis and also, in part, on the data and size range limitations of

measuring equipment used at the time the standards were set (Morawska et al.

2008a).

Prior to setting the PM2.5 standard, the USEPA conducted an extensive

examination of the available data on particle size distributions. A decision was

made to introduce 2.5 µm as the upper boundary range for fine particles and as a

basis for a standard; and this decision was strongly influenced by the fact that

epidemiological data available at that time were obtained using PM2.5

measurements (Dockery et al. 1993).

The concentration values of mass-based particle standards can exhibit bias toward

larger particles, which is a major limitation, and the presence of a few larger

particles can mask concentrations of the finer particles that contribute very little to

mass (Morawska et al. 1999).

Very little information can be obtained about particle number from particle mass

measurements (ECJRC 2002), hence it is important to measure ultrafine particles

in terms of particle number, and momentum has been gaining in both scientific

and regulatory circles in this respect. For example, in terms of particle emission

44

rates, the Swiss Agency for the Environment, Forests and Landscape has

proposed the introduction of a particle number standard for diesel passenger

vehicles for solid particles in the 0.02-0.30 µm size range (AQEG 2005); and the

European Commission is adding a particle number limit and new measurement

procedure to its EURO V/VI emissions standards for light duty diesel vehicles

and to its EURO VI emissions standard for heavy duty diesel vehicles relating to

solid particles, based on the recommendation of the UNECE-GRPE Particulate

Measurement Program (European Union 2007; Commission of the European

Communities 2007a,b; Morawska et al. 2008b).

This Particulate Measurement Programme was formed in 2001 and focuses on

providing recommendations for new or additional particle measurement systems to

be used for EU type approval, testing and development approval, testing and

development of future emission standards for both light- and heavy-duty vehicles

(http://www.dieselnet.com/news/2002/10ricardo.php). Its objectives include

identifying the best metrics for future particle measurements; to determine the

methods and instruments utilizing those metrics; and to investigate a test procedure

for measuring particles during type approval tests and recommend a suitable test

system or systems (http://www.dieselnet.com/news/2002/10ricardo.php;

http://www.empa.ch/plugin/template/empa/*/20988/---/I=1).

In Australia air quality standards for a variety of pollutants are set at a

national level through National Environmental Protection Measures (NEPM)

and at State and Territory levels through Environmental Protection Authorities,

and in 1998 the National Environment Protection Council (NEPC), a statutory

45

body, developed Australia's first national ambient air quality standards as part of the

NEPM for Ambient Air Quality (the 'Air NEPM')

(http://www.environment.gov.au/atmosphere/airquality/standards.html).

The Australian Air NEPM sets national standards for six key air pollutants -

carbon monoxide, ozone, sulfur dioxide, nitrogen dioxide, lead and particles

(PM10), and formal reporting against these standards commenced in 2002;

and for PM2.5 a NEPM advisory reporting standard and goal has been set to

collect national data to enable a review of the standard

(http://www.environment.gov.au/atmosphere/airquality/standards.html).

To inform development of the Australian Diesel NEPM, the NEPC has

commissioned preparatory projects, which include the testing of particulate matter

and toxic emissions generated by a range of different diesel-fuelled vehicles

(www.ephc.gov.au/taxonomy/term/70; DOEH 2003; NEPC 2000).

2.2.6. Modality within particle size distributions

A mode may be defined as a peak in the lognormal function of the number or

mass distribution of an atmospheric aerosol (John 1993). Its location can depend

on the particle metric being examined, for example particle number, volume,

mass or surface area, and can change depending on the mathematical

transformation method used (Morawska et al. 2008a).

Three classifications commonly used to categorise modal diameters in

atmospheric aerosol size distributions are based on particle size and production

mechanisms. These are the nucleation mode (< 0.1 µm), accumulation mode

46

(0.1-1 µm) and coarse particle mode (> 1 µm) (Jaenicke 1993). It has also been

shown that a clear separation usually exists between the accumulation and coarse

modes around 1 µm or somewhat above, where the mass of particles belonging to

these two modes is at a minimum (Lundgren and Burton 1995). However, the

delineation of these modal diameters, as discussed above, can vary. For example,

in Whitby’s model of particle volume size distribution (1978), which was based

primarily on atmospheric aerosol number distributions in the size range

0.01-6 µm, when these were transformed to volume distributions, showed modal

size ranges for the nuclei mode (< 0.1 µm), accumulation mode (0.1-2 µm) and

coarse particle mode (> 2 µm) (Baron and Willeke 2001). More recent studies

using instruments measuring down to the smaller size limit of 0.003 µm, have

shown that the nuclei mode needs to be separated into a nucleation mode (< 0.01

µm) and an Aitken nuclei mode (0.01-0.1 µm) (USEPA 2004).

No studies are currently available which have comprehensively investigated the

location of modes in a wide range of different environments and for different

particle metrics in terms of the global picture. The most comprehensive study to

date which has examined the location of modes in a broad range of different

environments is limited to an examination by Morawska et al. (1999b), who

identified modes in particle volume and particle number size distributions in six

different environmental aerosols in South-East Queensland. Other studies to date

have examined modal location values for a smaller number of environmental

aerosols.

47

Gaining knowledge and understanding about the presence and location of modes

in particle size distributions is critical to our understanding about the mechanisms

of atmospheric processes and, most importantly, for assessing risk and exposure,

as modes in particle size distributions have the potential to be used for developing

air quality guidelines and standards (Morawska et al. 2008a).

2.3. VEHICLE EMISSION INVENTORIES AND LOCAL MODELS

This section discusses methods used to develop emission inventories; including

global inventories and models; and local inventories and models developed for

South-East Queensland.

2.3.1. Developing emission inventories

Emission inventories provide information that is critical to guide development of

control strategies, risk assessments, air quality forecasting and transport and

economic incentive programs (Mobley and Cadle 2004). They have supported

major regulatory programs and withstood legal challenges; conversely, their

effectiveness can be limited by cost factors, timeliness, comprehensiveness, data

quality and representativeness (Mobley and Cadle 2004).

Inventory data can inform health impact assessments, aid identification of hot-

spots, be used to model the effects of future land use and transport planning, and

provide guidelines as to the direction needed for future research and development

activities. They can also inform our understanding of air quality and climate

change issues on global, regional and local scales (Parrish 2006).

48

There are many approaches available for estimating motor vehicle inventories.

These can range from estimations developed using a combination of performance-

related emission factors and road traffic data, to estimations based on total fuel

consumption data and fuel properties (Goodwin et al. 1999) or remotely sensed

data (Shifter et al. 2005). The choice of approach is very often dependent upon

available data, and the scale of the inventory required for decision-making (eg.,

road link-based, local, regional, state, country-specific).

Collecting data for inventories can require the use of very expensive

instrumentation, and therefore some countries with limited access to funding to

conduct measurement campaigns may use indirect methods for estimating motor

vehicle emissions, such as basing estimates on total fuel consumption data of a

vehicle fleet. An example is the Rapid Assessment Method, which was developed

based on a study of six cities in developing countries, that links health damages

and other environmental costs to a particular fuel use or pollution source, for cost

benefit analyses of pollution abatement measures (Lvovsky et al. 2000).

In Mexico, for example, limited data exists that is suitable to use in estimating on-

road emissions from petrol-powered vehicles, hence fuel sales have been used to

estimate vehicle activities and remotely sensed data to estimate exhaust emission

factors (Shifter et al. 2005). On the other hand, in the US micro scale emissions

data for quantifying high emission hotspots along roads have been collected using

real-time measurements of on-road vehicle emissions, and this empirical data can

be used to develop emission inventories (Unal et al. 2004).

49

Techniques for estimating emission rates and typical emission patterns can range

from simply deducing these based on measurement of roadside particle

concentration and vehicle mix (Shi et al. 1999; Querol et al. 2002) to considering

the effects of wind direction and speed (Giechaskiel et al. 2005); basing estimates

on road tunnel measurements which may have reduced wind condition effects

(Sturm et al. 2003; Jamriska et al. 2004) and single vehicle chasing experiments

(Kittelson et al. 2000; Kittelson et al. 2002; Vogt et al. 2003; Pirjola et al. 2004).

Developing accurate emission inventories for mobile sources, such as motor

vehicles, require cost-effective methods for determining activity indicators and

emission factors for a representative sample of vehicles using a broad range of

fuels, under different driving conditions and in ambient conditions that have high

spatial and temporal resolution (Mobley and Cadle 2004).

Examples of models used to estimate on-road vehicle emission inventories

include those used in the US, such as MOBILE (USEPA 1993), EMFAC (CARB

2002) in California; and COPERT used in Europe (Ahlvik et al. 1997;

Ntziachristos et al. 2000; Bellasio et al. 2007), as well as a more recent model

VERSIT+ LD (Smit et al. 2007), to name just a few. There are numerous models

and inventories of motor vehicle emissions around the world but these are mainly

restricted to inventories for PM2.5 and PM10.

Many international studies in urban environments that have estimated emission

inventories and the contribution of motor vehicles to total levels of ambient

particle concentrations have related to Total Suspended Particles or PM10 and, to a

50

lesser degree, to PM2.5. To date very little data and insufficient measurements are

available to use in compiling inventories of vehicle emissions to quantify particle

number emissions or additional ranges of sizes in terms of particle mass

(Morawska et al. 2004). No detailed emission inventories currently exist for

particle number concentration (Jones and Harrison 2006); nor are there any

comprehensive inventories available anywhere in the world of motor vehicle

particle emissions that cover the full size range of particles emitted and which

include both particle mass and particle number.

The general approach taken to quantify emissions is based on the concept of

combining a numerical value for an Emission Factor with a numerical factor for

an Activity, ie., Emissions = Emission Factor * Activity

(http://www.naei.org.uk/index.php). For example, to quantify the total number of

particles emitted from one passenger car travelling on a road link, an Emission

Factor (representing the number of particles emitted by a single passenger car

when it drives one kilometre) is multiplied by the Activity (the number of

kilometres the single passenger car travelled on the road link).

However, developing road-link based inventories of motor vehicle emissions can

be extremely complex, involve consideration of a multiplicity of factors and

require a very large amount of data. These complexities include identification of

suitable emission factors for different vehicle and road type combinations, and

assigning traffic data to different road type links in transport network models.

Selection of suitable emission factors to use in transport modelling require

consideration of numerous factors including, but not limited to, vehicle-related

51

factors such as vehicle type, age, fuel type, engine size, engine power, load,

driving mode (idling, accelerating, decelerating, cruising), speed; road type

(urban, highway, rural, motorway, freeway), speed limit on the road and average

vehicle speed (http://www.naei.org.uk/index.php). Average vehicle speed can

often be reduced by congested traffic conditions.

The complexity involved in developing inventories further intensifies when the

inventory is being developed for a number of different particle metrics and

particle size ranges, and a vital component of developing vehicle fleets

inventories is the derivation of suitable emission factors.

2.3.2. Local inventories and models for South-East Queensland and

Queensland

Local model estimates for motor vehicle emissions for South-East Queensland for

total annual PM10 have been prepared by the Queensland Environmental

Protection Agency (QEPA) (EPA 2004) and for urban South-East Queensland by

Apelbaum Consulting (Apelbaum 2006) and the Bureau of Transport and

Regional Economics (BTRE 2003).

The QEPA’s most recent inventory of total annual PM10 was modelled for the

year 2000 and was 2249 tonne for the South-East Queensland region (EPA 2004)

(Table 2.1). They developed a fleet emissions model using estimates of vehicle

52

kilometres travelled (VKT), emission factors and operating conditions; and

estimated emissions for six vehicle classes, four fuel types, operating conditions

(including average travel speed, road grade, engine hot and cold starts), time of

day, day of the week and the summer/winter season (EPA 2004). In this study,

QEPA’s future projection for SEQ for PM10 for 2005 was, in a low population

scenario, 2188 tonne per annum and, in a high population scenario, 2259 tonne

per annum (EPA 2004).

The Apelbaum inventory used speed dependent emission factors for different road

types based on a combination of Australian and European data and estimated

1549 tonne per annum of PM10 for urban South-East Queensland in the period

2003-2004, shown in Table 2.1 (Apelbaum 2006).

The Bureau of Regional and Transport Economics (BTRE) carried out a study of

metropolitan and non-metropolitan areas in Australia and modelled trends in

future levels of noxious pollutant emissions from motor vehicles, including for

particulate matter emissions (BTRE 2003). They considered growth in the

economy, population, travel demand and urban congestion, as well as future

vehicle design and fuel standards, deterioration of vehicle performance due to

vehicle age and rises in fuel consumption. Their estimate of annual PM10 for

2004 emitted from the urban South-East Queensland fleet was 1840 tonne (BTRE

2003), listed in Table 2.1.

53

The BTRE study reported that the uncertainty in their particulate matter estimates

were high, and the part of their analysis with the greatest levels of uncertainty

(BTRE 2003). Although the BTRE model provided some estimates for particle

sizes smaller than PM10, these are not considered to be relevant due to the large

uncertainties they have reported in their particulate matter estimates. They also

reported that their modelled results for total metropolitan particulate matter

emissions lacked data on the details of average fleet particulate matter production,

including for petrol vehicles (BTRE 2003). Therefore, for the purpose of

comparison, only estimates for PM10 are considered relevant.

Table 2.1 Estimates of total annual PM10 for South-East Queensland

(SEQ) and urban South-East Queensland related to 2004 a

Modellers Region modelled Year of

inventory Estimate of total PM10 emissions, Tonne per

annum Queensland Environmental Protection Agency (EPA 2004)

South-East Queensland

2000

2249

Bureau of Transport and Regional Economics (BTRE 2003)

Urban SEQ a 2004 1840

Apelbaum Consulting (Apelbaum 2006)

Urban SEQ a 2003-2004 1549

a Urban South-East Queensland covers around 26% of the South-East

Queensland region, but its fleet accounted for more than 70% of private

passenger trips in SEQ in 2004 (SEQHTS 2004).

54

Australia’s National Pollutant Inventory (NPI) includes State and Territory

estimates for motor vehicle emissions. The NPI is an internet database available

free to Australians which provides information on pollution emitted to the air, land

and water. It reports data on the source and location of 93 toxic substances relating

to manufacturing sites, households and transport emissions to assess their potential

impact on health and the environment (http://www.npi.gov.au/index.html).

Estimates in the NPI for diffuse emissions, such as motor vehicles, are estimated

by State and Territory governments, and relate to the contribution of non-

industrial sources to Australia’s pollutant emissions. Diffuse emissions are not

estimated annually but every 3-5 years, and motor vehicles are considered the

most significant diffuse source nationally (http://www.npi.gov.au/index.html).

The latest NPI estimates for motor vehicle emissions for Queensland are for the

2006-2007 NPI reporting year, and estimated 13 pollutant substances, including

Benzene, 1,3-Butadiene, Carbon monoxide, Cyclohexane, Ethylbenzene,

n-Hexane, Oxides of Nitrogen, Styrene, Sulphur dioxide, Toluene, Total Volatile

Organic Compounds, Xylenes, and PM10 (DEWHA 2008a). The NPI’s estimate

for the Queensland motor vehicle fleet in 2006-2007 is 2200 tonne per annum

(DEWHA 2008b); and no earlier estimate for Queensland is available prior to this

projection for 2006.

55

2.4 TRANSPORT MODELS

Transportation planning is a critical component in the growth and evolution of

metropolitan areas (Murray et al. 1998). It needs to consider trip purpose, spatial

and temporal distributions of trips, as well as modal splits of travel and cost.

Traffic count data is often not available for significant sections of road networks

and in cases where traffic volume data is required, but unavailable, travel demand

models are used to estimate this data (Zhong and Hanson 2008). Travel demand

model data can provide greater detail for transport planning on the spatial

distribution of vehicle activity, different road types and speeds travelled on these

roads, to enable more accurate estimates to be made of emission rates at the local

scale (Cook et al. 2006).

The most common approach to travel demand modelling is the 4-step demand

model, which incorporates trip generation, distribution, modal split and

assignment (Ortuzar and Willumsen 2001). Transport models require very large

amounts of data and apply a number of assumptions in order to assign traffic data

to road links in a road network. A recent literature review found very limited

quantitative data was available on uncertainties in transport model systems as a

whole (Nielsen and Knudsen 2006).

In developed countries transport modellers often have access to large databases,

such as land use and transport network data, but in developing countries these are

often rare (Walker et al. 2008). One recent example is a study by Walker et al.

(2008) who used data from a 1,000 household travel and activity survey for

56

Chengdu in China (which has an urban population of around 3 million) to develop

travel mode choice models for policy analysis and planning.

As more detailed data, such as data available in geographic information systems

related to land use and population distributions becomes available, more detailed

analyses and modelling related to sustainability are possible, eg., developing

ecological and carbon footprints of the transport system.

2.5. ESTIMATE OF ROAD TRANSPORT EMISSIONS PREPARED FOR

THE UK

An extensive review of inventories developed for motor vehicles revealed only

one study which had attempted to estimate particle emissions emitted by a motor

vehicle fleet. This estimate was prepared for the UK for 1996, 1998 and 2001

(Group 1999; Goodwin et al. 2000; AQEG 2005). Their PM10 estimates were

derived by multiplying emission factors for different vehicle and road types by

annual VKT data. In order to derive estimates for PM0.1, PM1 and PM2.5, they

applied distribution profiles for these size ranges to PM10 estimate data, which

means that the emission factors for size ranges below PM10 were based simply on

mass fractions multiplied by PM10 estimate data values, and were not based on

individual measurements of different particle sizes. The method depended on

PM10 emission rates, which in themselves had substantial uncertainties, and

therefore their estimates for smaller sized particles contain even more uncertainty,

due to additional uncertainties in the size fractions (Group 1999).

57

Their estimates of motor vehicle particle emissions for 1996 and 1998 were

derived by applying the same distribution profiles for petrol and diesel exhaust,

viz., mass fractions of PM10 of 85% for PM1 and 90% for PM2.5 based on 33

different particle size distributions (Group 1999; Goodwin et al. 2000). The PM0.1

distribution profile was based on size fractions taken from a European inventory

(TNO 1997). Mass fractions of PM10 used in the 2001 estimate for PM0.1, PM1

and PM2.5 were derived from distribution profiles taken mostly from the USEPA

compilation of emission factors (USEPA 1995) known as AP-42 (AQEG 2005).

The UK emission factors were based on vehicle classes (or EUROs).

The estimate of motor vehicle fleet emissions for 2001 for the UK estimated that

PM10 tailpipe particle emissions emitted from motor vehicles were made up of

about 90% diesel vehicle emissions and 10% petrol vehicle emissions (AQEG

2005), and these estimates are likely to be heavily influenced by the use of petrol

vehicle emission factors that were several orders of magnitude lower than the

diesel values used. In terms of this UK mass fraction, the estimate for the

difference between petrol and diesel vehicle emissions would not normally be as

large. For example, differences between emission rates for these two fuel types

can vary considerably, however as the UK estimates of motor vehicle particle

emissions were based on such few data, these estimates cannot be considered to

be an inventory, comprising of robust and comprehensive estimates.

58

2.6. IDENTIFYING SUITABLE PARTICLE EMISSION FACTORS

Inventories can provide vital information to inform air quality monitoring and

regulation, and transport planning; however the selection of the most suitable

emission factors to use in developing these inventories can be a very complex

process. This is because emission factors can be derived from a range of

different measurement techniques, such as from direct measurements taken on

or near roads, in tunnels or on dynamometers; to indirect methods such as

deriving values based on estimates of fuel consumption or remotely sensed

data.

Many issues need to be considered and resolved in selecting emission factors,

such as vehicle type, fuel type, instrumentation used, size range measured, study

location, road type, vehicle speed or drive cycle tested, to name a few. The

literature review revealed that more than 900 particle emission factors are

published in the international literature, which have been derived from a wide

variety of different instrumentation methods, for different particle size ranges, and

conducted in different parts of the world. These are discussed in more detail in

Chapter 5. However, it remains unclear which of these emission factors are the

most suitable to use in transport modelling.

With the exception of dynamometer studies, the majority of studies reviewed paid

little attention to identifying the average vehicle speed of different vehicle types at

their study sites. The posted speed limit of a road may not necessarily represent

the actual speed of vehicles travelling on the road link, which could substantially

59

influence particle emission levels. Congestion is a major issue in most urban

areas, and there is little information available on speed-related emission factors to

model vehicles travelling at lower speeds in congested conditions, such as at

< 50 and < 30 km/hr.

2.7. SUMMARY

Characteristics of motor vehicle particle emissions

To fully understand the effects of motor vehicle particle emissions on human

health it is important to derive size-resolved inventories of particle emissions that

include both particle number and particle mass. These inventories need to

quantify emissions at a range of scales from the micro scale, in terms of direct

emissions - emission rates related to on-road emissions, to broader scales such as

quantification of the spatial distribution of particle concentrations over a region.

Data from epidemiological studies which investigate particulate matter exposure

and its health impacts are extremely important due to the serious health effects

associated with exposure to and inhalation of particles. Diesel vehicle emissions,

in particular, require special attention due to their relatively high emission rates,

and the fact that diesel exhaust is a declared carcinogen.

Together, inventory and epidemiological data can guide the development of air

quality guidelines and standards, and provide data for monitoring, planning and

health impact assessments. It is essential that these studies investigate particles in

the submicrometre (diameters < 1 µm) and smaller size ranges, such as ultrafine

60

particles, where motor vehicle particle emissions tend to dominate; as well as

particle mass emissions in the size ranges between 1-10 µm. The importance of

controlling ultrafine particles emitted from motor vehicles is receiving greater

attention by regulators, evidenced by efforts in Switzerland and Europe to

introduce regulations to control solid particle number emissions.

Another aspect about which the global picture is lacking and for which there is

limited research, is the nature of modality within particle size distributions.

Modes represent the particle size associated with the highest concentrations in an

environment, and examining this particle characteristic has the potential to

provide a deeper understanding about the nature and mechanisms of particle

emissions in different environments, and may also provide an important basis for

developing air quality regulations.

Motor vehicles are the dominant source of ultrafine particles in urban areas and as

current air quality standards are mass and not particle number-based, this means

that this major global pollution source is not adequately regulated or controlled.

Limited information is available upon which to base the development of particle

number standards for motor vehicle emissions, in order to protect human health

and the environment.

61

Vehicle emission inventories and local models

Current inventories and models of particles emitted from motor vehicle tailpipes

around the world are restricted to particle mass, PM10, and to a limited extent

PM2.5, and no detailed inventories are available that have quantified particle

number or PM1.

There are a number of approaches available for quantifying motor vehicle particle

emission inventories, which are dependent upon availability of data, or resources

to access this data, and the scale or level of detailed required for the inventory.

Further, to develop comprehensive inventories for motor vehicles these

approaches need to include the full size range of particles emitted for both particle

number and particle mass emissions.

Inventories provide key and vital information for developing targeted control

strategies and regulatory programs, to identify hotspots and problem routes, and

provide data that can be used to model scenarios to test the air quality

implications of future land use and transport planning, and events such as the

introduction of new vehicle standards (eg., EUROs) and fuels. Hence, inventories

need to be made an integral part of transport-related projects. In this way they can

be used to evaluate the impact of development projects pre- and post-construction

(eg., building of new busways and tunnels) and the effect of major shifts in

transport mode choice or vehicle occupancy rates, to assess their potential impact

on particle emission levels.

62

The review found that no comprehensive inventory currently exists either for

South-East Queensland or for any region in the world that quantifies particles

emitted from motor vehicle tailpipes covering the full size range of particles

emitted and which includes particle number and particle mass emissions.

Transport models

Transportation planning is a very important task for metropolitan areas. Often

data on traffic volumes for significant sections of road networks are not available

and therefore travel demand models are developed so that more detailed

information and accurate estimates of local scale emissions can be made. These

models require a very large quantity of data which can be obtained from

household travel and activity surveys.

Estimate of road transport emissions prepared for the UK

The only study which has attempted to develop a motor vehicle emissions

inventory for tailpipe particle emissions was prepared for the UK, and this

estimate contained a large degree of uncertainty (Group 1999; Goodwin et al.

2000; AQEG 2005). Their PM10 emission data contained substantial uncertainty,

and their data for the smaller particle size ranges were derived by applying

distribution profiles this PM10 estimate data (Group 1999; Goodwin et al. 2000;

AQEG 2005), and was not based on estimates from individual measurements of

different particle sizes.

63

Identifying suitable particle emission factors

A substantial amount of data is available in the international literature for

motor vehicle particle emission factors which have been derived using a broad

range of different techniques and measuring different size ranges, however it

remains unclear which particle emission factors are the most suitable to use in

transport modelling and air quality assessments. In addition, few emission

factor data is available for motor vehicles travelling at low speeds, such as at

< 50 km/hr and < 30 km/hr, to enable modelling of congested traffic

conditions.

2.8. KNOWLEDGE GAPS AND CONCLUSIONS FROM THIS REVIEW

This literature review has demonstrated that:-

• Very little knowledge currently exists about total particulate

matter pollution emitted by motor vehicle fleets in urban areas, in

terms of both particle number and for different size fractions of

particle mass.

• A large body of data is available of emission factors for different

motor vehicle types under different driving conditions, however

it is unclear which are the most suitable to use in transport

modelling and air quality assessments; and that data relating to

emission factors for motor vehicles travelling at < 50 km/hr is

very limited.

64

• A comprehensive inventory has not been published which has

quantified emissions in terms of the full size range of tailpipe

particle emissions emitted by motor vehicle fleets.

• Little data exists to enable estimation of motor vehicle tyre or

brake wear emissions, or to discriminate road dust from motor

vehicle particle emissions; and data for particle surface area

measurements are also rare.

There are three very important reasons for quantifying particles

emitted from motor vehicle tailpipes:-

o Firstly, motor vehicle fleets are a major source of particulate pollution

in urban areas, particularly of ultrafine particles.

o Secondly, there are known adverse health effects associated with

exposure to particulate matter.

o Thirdly, if we cannot quantify the contribution of particle emissions

from motor vehicle fleets on or near roads, we have little chance of

controlling them or gaining an understanding about their contributions

on a global scale, and their likely effect on our earth’s climate and

upper atmospheres.

It is important to develop size-resolved motor vehicle particle inventories that

cover the full size range of particles emitted from motor vehicle tailpipes, and

include particle number and particle mass emissions; and air quality standards

are needed for particle number and PM1 to control submicrometre and

ultrafine particle emissions.

65

2.9. REFERENCES:

Ahlvik, P., Eggleston, S., Goriben, N., Hassel, D., Hickman, A. J., Joumard, R.,

Ntziachristos, L., Rijkeboer, R., Samaras, Z., Zierock, K. H., 1997. COPERT II

Computer programme to calculate emissions from road transport: methodology

and emission factors. Technical report prepared by the European Environment

Agency, Copenhagen. Report No. 6.

Apelbaum, 2006. Queensland Transport Facts, Apelbaum Consulting Group Pty

Ltd, Mulgrave, Victoria, Australia.

AQEG., 2005. Particulate Matter in the UK. London, Department for

Environment, Food and Rural Affairs.

Baron, P. A., Willeke, K., 2001. Aerosol Measurement, Principles, Techniques

and Applications, 2nd edn. New York, John Wiley & Sons, Inc.

Bellasio, R., Bianconi, R., Corda, G., Cucca, P., 2007. Emission inventory for the

road transport sector in Sardinia (Italy). Atmospheric Environment 41, 677-691.

Brook, R. D., Brooke, J. R., Urch, B. R., Vincent, R., Rajagopalan, S., Silverman,

F., 2000. Inhalation of fine particulate air pollution and ozone causes acute

arterial vasoconstriction in healthy adults. Circulation 105, 1534-1536.

66

BTRE., 2003. Urban pollutant emissions from motor vehicles: Australian trends

to 2020. Final Draft Report for Environment Australia. Canberra, Bureau of

Transport and Regional Economics.

CARB., 2002. EMFAC2001/EMFAC200. Calculating emissions inventories for

vehicles in California, User’s Guide, California California Air Resources Board.

Cook, R., Touma, J. S., Beidler, A., Strum, M., 2006. Preparing highway

emissions inventories for urban scale modeling: A case study in Philadelphia.

Transportation Research Part D: Transport and Environment 11(6), 396-407.

Commission of the European Communities, 2007a. Proposal for a Regulation of

the European Parliament and of the Council on type-approval of motor vehicles

and engines with respect to emissions from heavy duty vehicles (Euro VI) and on

access to vehicle repair and maintenance information, Brussels.

Commission of the European Communities, 2007b. Annex to the Proposal for a

Regulation of the European Parliament and of the Council on the approximation

of the laws of the Member States with respect to emissions from on-road heavy

duty vehicles and on access to vehicle repair information, Impact Statement,

Brussels.

DEWHA (Department of the Environment, Water, Heritage and the Arts) 2008a,

Substance Emissions - Motor Vehicles, Queensland, 20 September.

67

DEWHA (Department of the Environment, Water, Heritage and the Arts) 2008b,

Particulate Matter 10.0 µm Summary - All sources: Queensland, 20 September.

DEWHA (Department of the Environment, Water, Heritage and the Arts),

Ambient Air Quality Standards,

ttp://www.environment.gov.au/atmosphere/airquality/standards.html),

verified 9 July 2009.

DOEH., 2003. Technical Report No. 1: Toxic Emissions from Diesel Vehicles in

Australia, Department of the Environment and Heritage, Canberra.

Dockery, D. W., Pope, C. A., Xu, X., Spengler, J. D., Ware, J. H., Fay, M.E.,

Ferris, B.G., 1993. An Association between Air Pollution and Mortality in Six

U.S. Cities. The New England Journal of Medicine 329(24), 1753-1759.

ECJRC., (European Commission Joint Research Centre) 2002. Guidelines for


particulate matter for use in epidemiological studies. Italy European Commission.

Environment Protection and Heritage Council, National Environment Protection

Council, Preparatory Work, www.ephc.gov.au/taxonomy/term/70, verified 9 July

2009.

EPA., 2004. Air Emissions Inventory South-East Queensland Region,

Environmental Protection Agency, Brisbane.

68

European Union 2007, Official Journal of the European Union, Regulation (EC)

No 715/2007 of the European Parliament and of the Council of 20 June 2007 on

type approval of motor vehicles with respect to emissions from light passenger

and commercial vehicles (Euro 5 and Euro 6) and on access to vehicle repair and

maintenance information, Strasbourg.

European Commission. http://ec.europa.eu/index_en.htm. Date verified 28 July

2008.

Ferin, J., Oberdoerster, G., Penney, D. P., Soderholm, S. C., Gelein, R., Piper,

H.C., 1990. Increased pulmonary toxicity of ultrafine particles I. Particle

clearance, translocation, morphology. Journal of Aerosol Science. 21(3), 381-384.

Giechaskiel, B., Ntziachristos, L., Samaras, Z., Scheer, V., Casati, R., Vogt, R.,

2005. Formation potential of vehicle exhaust nucleation mode particles on-road

and in the laboratory. Atmospheric Environment 39(18), 3191-3198.

Goodwin, J. W. L., Salway, A. G., Eggleston, H. S., Murrells, T. P., Berry, J.E.,

1999. National Atmospheric Emissions Inventory, UK Emissions of Air

Pollutants 1970 to 1996, National Environmental Technology Centre on behalf of

the Department of the Environment, Transport and the Regions..

Goodwin, J. W. L., Salway, A. G., Murrells, T. P., Dore, C. J., Passant, N.R.,

Eggleston, H.S., 2000. UK emissions of air pollutants 1970-1998. A Report of the

National Atmospheric Emissions Inventory. London, Department of the

Environment, Transport and the Regions.

69

Graskow, B. R., Kittelson, D. B., Abdul-Khalek, I. S., Ahmadi, M., Morris, J.E.,

1998. Characterisation of Exhaust Particulate Emissions from a Spark Ignition

Engine. SAE Paper 980528, 155-165.

Group, 1999, Source Apportionment of Airborne Particulate Matter in the United

Kingdom. Report for the Department of the Environment, Transport and the

Regions, the Welsh Office, the Scottish Office and the Department of the

Environment (Northern Ireland).

Harris, S. J., Maricq, M. M., 2001. Signature size distributions for diesel and

gasoline engine exhaust particulate matter. Journal of Aerosol Science 32, 749-

764.

Harrison, R., Jones, M., Collins, G., 1999. Measurements of the Physical

Properties of Particles in the Urban Atmosphere. Atmospheric Environment 33,

309-321.

Harrison, R. M., Shi, J. P., Zi, S., Khan, A., Mark, D., Kinnersley, R., Yin, J.,

2000, Measurement of number, mass and size distribution of particles in the

atmosphere. Philosophical Transactions of the Royal Society A: Mathematical,

Physical and Engineering Sciences 358(1775), 2567-2580.

70

Hinds, W. C., 1982. Aerosol Technology Properties, Behaviour, and

Measurement of Airborne Particles. New York, John Wiley & Sons.

Jaenicke, R., 1993. Tropospheric Aerosols. San Diego, USA, Academic Press.

Jamriska, M., Morawska, L., 2003. Quantitative Assessment of the Effect of

Surface Deposition and Coagulation on the Dynamics of Submicrometer Particles

Indoors. Aerosol Science and Technology 37(5), 425-436.



38(24), 6701-6709.



techniques and applications,Van Nostrand Reinhold, New York. 55.

Jones, A. M., Harrison, R. M., 2006. Estimation of the emission factors of particle



Kirchstetter, T. W., Harley, R. A., Kreisberg, N. M., Stolzenberg, M.R., Hering,

S.V., 1999, On-road measurement of fine particle and nitrogen oxide emissions

from light- and heavy-duty motor vehicles. Atmospheric Environment 33, 2955-

2968.

71

Kittelson, D., 1998, Engines and Nanoparticles: a Review. Journal of Aerosol

Science 29(5), 575-588.

Kittelson, D., Johnson, J., Watts, W. F., Wei, Q., Drayton, M., Paulsen, D.,

Bukowiecki, N., 2000, Diesel aerosol sampling in the atmosphere. SAE

Technology Paper 2000-01-2212.

Kittelson, D. B., Watts, W. F., Johnson, J., 2002. Diesel aerosol sampling

methodology. CRC E-43 Final report.

Lundgren, D. A., Burton, R.M., 1995. Effect of particle size distribution on the

cut point between fine and coarse ambient mass fractions. Inhalation Toxicology

7(1), 131-148.

Lvovsky, L., Hughes, G., Maddisoin, D., Ostro, B., Pearce, D., 2000.

Environmental Costs of Fossil Fuels: A Rapid Assessment Method with

Application to Six Cities, The World Bank Environment Department, The World

Bank.

Mobley, J.D., Cadle, S. H., 2004. Innovative Methods for Emission Inventory

Development and Evaluation: Workshop Summary. Journal of the Air & Waste

Management Association 54, 1422-1439.

72

Morawska, L., Salthammer, T., 2003. Chapter 3: Motor Vehicle Emissions as a

Source of Indoor Particles in, Morawska-Salthammer (eds). Indoor Environment,

Wiley-VCH.

Morawska, L., Thomas, S., Bofinger, N., Wainwright, D., Neale, D., 1998a.

Comprehensive characterization of aerosols in a subtropical urban atmosphere:

particle size distribution and correlation with gaseous pollutants. Atmospheric

Environment 32(14/15), 2467-2478

Morawska, L., Bofinger, N. D., Kocis, L., Nwankwoala, A., 1998b.

Submicrometer and supermicrometer particles from diesel vehicle emissions.

Environmental Science & Technology 32(14), 2033-2042.

Morawska, L., Johnson, G., Ristovski, Z.D., Agranovski, V., 1999a. Relation

between particle mass and number for submicrometer airborne particles.

Atmospheric Environment 33(13), 1983-1990.

Morawska, L., Thomas, S., Jamriska, M., Johnson, G., 1999b. The modality of

particle size distributions of environmental aerosols. Atmospheric Environment

33(27), 4401-4411.

Morawska, L., Ristovski, Z., Jayaratne, E. R., Keogh, D.U., Ling, X., 2008a.


characteristics, ambient processing and implications on human exposure

Submitted to Atmospheric Environment.

73

Morawska, L., Keogh, D. U., Thomas, S. B., Mengersen, K., 2008b. Modality in



Morawska, L., Moore, M. R., Ristovski, Z.D., 2004. Health Impacts of Ultrafine

Particles - Desktop Literature Review and Analysis, Department of the


Murray, A. T., Davis, R., Stimson, R. J., Ferreira, L., 1998. Public Transportation

Access. Transportation Research Part D: Transport and Environment 3(5), 319-

328.

National Atmospheric Emissions Inventory, http://www.naei.org.uk/index.php.

Date verified 28 July 2008.

NEPC, 2000, Proposed Diesel Vehicle Emissions National Environment

Protection Measure Preparatory Work, In-Service Emissions Performance - Phase

2: Vehicle Testing, NEPC, Adelaide, November.

NPI (National Pollutant Inventory), Department of the Environment, Water,

Heritage and the Arts, Australian Government, http://www.npi.gov.au/index.html.


74

Nielsen, O. A., Knudsen, M.A., 2006. Uncertainty in traffic models. European

Transport Conference Strasbourg, France

Ntziachristos, L., Samaras, Z., Eggleston, S., Goriben, N., Hassel, D., Hickman,

A. J., Joumard, R., Rijkeboer, R., White, L., Zierock, K. H., 2000. COPERT III


and emission factors (version 2.1). Technical report prepared by the European

Environment Agency, Copenhagen, Report 49.

Oberdoerster, G., Sharp, Z., Atudorei, V., Elder, A., Gelein, R., Kreyling, W.,

Cox, C., 2004. Translocation of inhaled ultrafine particles to the brain. Inhalation

Toxicology 16, 437-445.

Ortuzar, J., Willumsen, L.G., 2001. Modelling Transport. 3rd edn., John Wiley &

Sons Inc. .

Particle Measurement Programme, DieselNet website,

http://www.dieselnet.com/news/2002/10ricardo.php, verified 9 July 2009.

Particle Measurement Programme (PMP), home page of PMP

http://www.empa.ch/plugin/template/empa/*/20988/---/I=1, verified 9 July 2009.



75

Pirjola, L., Parviainen, H., Hussein, T., Valli, A., Hameri, K., Aalto, P., Virtanen,

A., Keskinen, J., Pakkanen, T., Makela, J., Hillamo, R., 2004. "Sniffer" - A novel

tool for chasing vehicles and measuring traffic pollutants. Atmospheric

Environment 38, 3625-3635.

Pope, C. A., Burnett, R. T., Thun, M. J., Calle, E. E., Krewski, D., Ito, K.,

Thurston, G.D., 2002, Lung cancer, cardiopulmonary mortality, and long-term

exposure to fine particulate air pollution. Journal of the American Medical

Association 287(9), 1132-1141.

Pope, C. A., Dockery, D.W., 2006. Health Effects of Fine Particulate Air



Querol, X., Alastuey, A., de la Rosa, J., Sanchez-de-la-Campa, A., Plana,

F., Ruiz, C.R., 2002. Source apportionment analysis of atmospheric

particulates in an industrialised urban site in southwestern Spain.


Ristovski, Z., Jayaratne, E. R., Lim, M., Ayoko, G. A., Morawska, L., 2006.

Influence of diesel fuel sulphur on the nanoparticle emissions from city buses.

Environmental Science & Technology 40, 1314-1320.

76

Ristovski, Z. D., Morawska, L., 1998. Emission of submicrometer particles from

spark ignition vehicles. Journal of Aerosol Science 29(Supplement 2), S1001-

S1002.

Seaton, A., MacNee, W., Donaldson, K., Godden, D., 1995. Particulate air

pollution and acute health effects. Lancet 345, 176-178.

SEQHTS., 2003-2004. South-East Queensland Household Travel Survey

(SEQHTS) (Brisbane, Gold Coast and Sunshine Coast Area). Brisbane

Queensland Transport

Shi, J., Evans, D., Khan, A., Harrison, R., 2001. Sources and Concentration of

Nanoparticles ( < 10 nm Diameter) in the Urban Atmosphere. Atmospheric


Shi, J., Harrison, R.M., 1999. Investigation of ultrafine particle formation during

diesel exhaust dilution. Environmental Science & Technology 33, 3730-3736.

Shi, J. P., Khan, A. A., Harrison, R.M., 1999. Measurements of ultrafine particle

concentration and size distribution in the urban atmosphere. The Science of the

Total Environment 235, 51-64.

Shifter, I., Diaz, L., Mugica, V., Lopez-Salinas, E., 2005. Fuel-based motor

vehicle emission inventory for the metropolitan area of Mexico city. Atmospheric

Environment 39(5), 931-940.

77

Smit, R., Smokers, R., Rabe, E., 2007, A new modelling approach for road traffic

emissions: VERSIT+. Transportation Research Part D-Transport and


Somers, C. M., McCarry, B.E., Malek, F., Quinn, J. S., 2004, Reduction of

particulate air pollution lowers the rist of heritable mutations in mice. Science,

1008-1010.

Sturm, P. J., Baltensperger, U., Bacher, M., Lechner, B., Hausberger, S., Heiden,

B., Imhof, D., Weingartner, E., Prevot, A.S.H., Kurtenbach, R., Wiesen, P., 2003,

Roadside measurements of particulate matter size distribution. Atmospheric


TNO., 1997. Particulate Matter Emissions (PM10, PM2.5, PM<0.1) in Europe in

1990 and 1993, TNO Report TNO-MEP-R96/472. Netherlands.

Unal, A., Frey, H. C., Rouphail, N.M., 2004. Quantification of Highway Vehicle

Emissions Hot Spots based on on-board measurements. Journal of the Air &

Waste Management Association 54, 130-140.

USEPA., 1993. User's Guide to MOBILE5A, Mobile source emissions

factor model, U.S. Environmental Protection Agency.

USEPA., 1995. Compilation of Air Pollutant Emission Factors, 5th edn, AP-

42, North Carolina.

78

USEPA., 2004. Air quality criteria for particulate matter. Washington DC,

US Environmental Protection Agency, 600/P-99/002aF-bF.

Vogt, R., Kirchner, U., Scheer, V., Hinz, K. P., Trimborn, A., Spengler,

B., 2003. Identification of diesel exhaust particles at an Autobahn, urban

and rural location using single particle mass spectometry. Aerosol Science

& Technology 34, 319-337.

Wahlin, P., Palmgren, F., Van Dingenen, R., 2001. Experimental studies

of ultrafine particles in streets and the relationship to traffic. Atmospheric

Environment 35, S63-S69.

Walker, J. L., Li, J., Srinivasan, S., Bolduc, D., 2008. Travel Demand

Models in the Developed World: Correcting for Measurement Errors.

Transportation Research Board 87th Annual Meeting Washington.

WHO., 2000, Guidelines for Air Quality, World Health Organization, Geneva.

WHO., 2005, Guidelines for Air Quality. World Health Organization,

Geneva.

Willeke, K., Baron, P.A., 1993. Aerosol Measurement: Principles, Techniques,

and Applications. John Wiley & Sons, New York.

79

Young, L. H., Keeler, G.J., 2004. Characterization of ultrafine particle number

concentration and size distribution during a summer campaign in southwest

Detroit. Journal of the Air & Waste Management Association 54(9), 1079-1090.

Zhang, K., 2004. Ambient and Plume Processing of Atmospheric Ultrafine

Particles. PhD thesis, University of California, Davis.

Zhang, K. M., Wexler, A. S., 2004. Evolution of particle number distribution near

roadways - Part I: analysis of aerosol dynamics and its implications for engine

emission measurement. Atmospheric Environment 38(38), 6643-6653.

Zhong, M., Hanson, B. L., 2008. GIS-Based Travel Demand Modeling for

Estimating Traffic on Low-Class Roads. Transportation Research Board 87th

Annual Meeting Washington

Zhu, Y., 2003. Ultrafine particle and freeways. PhD thesis, University of

California, Los Angeles.

80

2.10. BIBLIOGRAPHY

ABS., 2004. Population by Age and Sex. Australian Bureau of Statistics,

Canberra.

ABS., 2004. Survey of Motor Vehicle Use Australia. Australian Bureau of

Statistics, Canberra.

Abu-Allaban, M., 2002. Exhaust particle size distribution measurements at the

Tuscarora Mountain tunnel. Aerosol Science and Technology 36(6), 771-789.

Abu-Allaban, M., Gillies, J. A., Gertler, A.W., 2003. Application of a multi-lag

regression approach to determine on-road PM10 and PM2.5 emission rates.


Abu-Allaban, M., Gillies, J. A., Gertler, A. W., Clayton, R., Proffitt, D., 2003.

Tailpipe, resuspended road dust, and brake-wear emission factors from on-road

vehicles. Atmospheric Environment 37(37), 5283-5293.

Affum, J. K., Brown, A. L., Chan, Y.C., 2003. Integrating air pollution modelling

with scenario testing in road transport planning: the TRAEMS approach. The

Science of The Total Environment 312(1-3), 1-14.

Ahlvik, P., Eggleston, S., Goriben, N., Hassel, D., Hickman, A.J., Joumard, R.,

Ntziachristos, L., Rijkeboer, R., Samaras, Z., Zierock, K. H., 1997. COPERT II

81




Almeida, S. M., Pio, C. A., Freitas, M. C., Reis, M. A., Trancoso, M.A., 2005.

Source apportionment of fine and coarse particulate matter in a sub-urban area at

the Western European Coast. Atmospheric Environment 39(17), 3127-3138.

Anderson, H. R., 2000. Differential epidemiology of ambient aerosols.

Philosophical Transactions of the Royal Society of London Series a-Mathematical




AQEG., 2005. Particulate Matter in the UK. Department for Environment, Food

and Rural Affairs, London.

ARB's., 2002. Study of Emissions from Two "Late Model" Diesel and CNG

Heavy-Duty Transit Buses. California Air Resources Board, 12th CRC On-Road

Vehicle Emissions Workshop, April 15-17, San Diego.

Ayala, A., Kado, N. Y., Okamoto, R.A., 2002. Diesel and CNG Heavy-duty

Transit Bus Emissions over Multiple Driving Schedules: Regulated Pollutants and

Project Overview. Society of Automotive Engineers SAE 2002-01-1722, 1-13.

82

Baron, P.A., Willeke, K., 2001. Aerosol Measurement, Principles, Techniques

and Applications, 2nd edn. New York, John Wiley & Sons, Inc.

Becker, S., Soukup, J. M., Sioutas, C., Cassee, F.R., 2003. Response of human

alveolar macrophages to ultrafine, fine, and coarse urban air pollution particles.

Experimental Lung Research 29(1), 29-44.

Bellasio, R., Bianconi, R., Corda, G., Cucca, P., 2007. Emission inventory for the

road transport sector in Sardinia (Italy). Atmospheric Environment 41, 677-691.

Berner, A., Galambos, Z., Ctyroky, P., Fruhaug, P., Hitzenberger, R., Gomiscek,

B., Hauck, H., Preining, O., Puxbaum, H., 2004. On the correlation of

atmospheric aerosol components of mass size distributions in the larger region of

a central European city. Atmospheric Environment 38(24), 3959-3970.

Bigg, E. K., Turvey, D.E., 1978. Sources of atmospheric particles over Australia.

Atmospheric Environment 12, 1643-1655.

Bin, O., 2003. A logit analysis of vehicle emissions using inspection and

maintenance testing data. Transportation Research Part D: Transport and


83

Birmili, W., Heintzenberg, J., Wiedensohler, A., 1999. Representative

measurement and parameterization of the submicron continental particle size

distribution Journal of Aerosol Science 30, Suppl. 1, S229-S230.

Birmili, W., Wiedensohler, A., Heintzenberg, J., Lehmann, K, 2001. Atmospheric

particle number size distribution in central Europe: Statistical relations to air

masses and meteorology. Journal of Geophysical Research - Atmospheres 106

(D23), 32005-32018.

Boddy, J. W. D., Smalley, R. J., Goodman, P. S., Tate, J. E., Bell, M. C., Tomlin,

A.S., 2005. The spatial variability in concentrations of a traffic-related pollutant

in two street canyons in York, UK-Part II: The influence of traffic characteristics.


Bradley, M. J., 2000. Hybrid-Electric Drive Heavy-Duty Vehicle Testing Project;

Final Emissions Report. Northeast Advanced Vehicle Consortium, Defense

Advanced Research Projects Agency, West Virginia University, USA.

Brodrick, C.J., 2001. Effects of real-world vehicle activities and loads on heavy-

duty diesel vehicle emissions. University of California, Davis.

Brook, R. D., Brooke, J. R., Urch, B. R., Vincent, R., Rajagopalan, S., Silverman,

F., 2000. Inhalation of fine particulate air pollution and ozone causes acute

arterial vasoconstriction in healthy adults. Circulation 105, 1534-1536.

84

Brown, A.L, .Affum, J.K., 2002. A GIS-based environmental modelling system

for transportation planners. Computers, Environment and Urban Systems 26(6),

577-590.

BTRE., 2003. Urban pollutant emissions from motor vehicles: Australian trends

to 2020, Final Draft Report for Environment Australia. Canberra, Bureau of

Transport and Regional Economics.

Bukowiecki, N., Dommen, J., Prevot, A. S. H., Richter, R., Weingartner, E.,

Baltensperger, U., 2002. A mobile pollutant measurement laboratory--measuring

gas phase and aerosol ambient concentrations with high spatial and temporal

resolution. Atmospheric Environment 36(36-37), 5569-5579.

BUWAL., 2004, 1 March. Ordinance on the determination of the particle number

emission level of passenger cars with compression ignition engines, Draft.

http://www.puntofocal.gov.ar/doc/che39.pdf. Date verified 20 February 2008.

Byers, R. J., 1999. Measurement of particulate matter size, concentration and

mass emissions from in-use heavy duty vehicles. United States - West Virginia,

West Virginia University.

Cadle, S. H., Belian, T. C., Black, K. N., Minassian, F., Natarajan, M., Tierney, E.

J., Lawson, D.R., 2005. Real-world vehicle emissions: A summary of the 14th

Coordinating Research Council On-Road Vehicle Emissions Workshop. Journal

of the Air & Waste Management Association 55(2), 130-146.

85

Cadle, S. H., Mulawa, P., Groblicki, P., Laroo, C., Ragazzi, R. A., Nelson, K.,

Gallagher, G., Zielinska, B., 2001. In-use light-duty gasoline vehicle particulate

matter emissions on three driving cycles. Environmental Science & Technology

35(1), 26-32.

Cadle, S. H., Mulawa, P. A., Ball, J., Donase, C., Weibel, A., Sagebiel, J. C.,

Knapp, K. T., Snow, R., 1997. Particulate emission rates from in use high

emitting vehicles recruited in Orange County, California. Environmental Science

& Technology 31(12), 3405-3412.

CARB., 2001. Heavy-Duty Emissions Laboratory, Heavy Duty Testing and Field

Support Section, California Air Resources Board. Report No. 01-01.

CARB., 2002. EMFAC2001/EMFAC200. Calculating emissions inventories for

vehicles in California, User’s Guide California California Air Resources Board.

Chatterjee, S., Conway, R., Lanni, T., Frank, B., Tang, S., Rosenblatt, D., Bush,

C., Lowell, D., Evans, J., McLean, R., Levy, S., 2002. Performance and

Durability Evaluation of Continuously Regenerating Particulate Filters on Diesel

Powered Urban Buses at NY City Transit - Part II. Society of Automotive

Engineers SAE 2002-01-0430.

Chen, C., Niemeier, D., 2005. A mass point vehicle scrappage model.

Transportation Research Part B: Methodological 39(5), 401-415.

86

Chung, A., Chang, D.P.Y., Kleeman, M.J., Perry, K. D., Cahill, T.A., Dutcher, D.,

McDougall, E.M., Stroud, K., 2001. Comparison of real-time instruments used to

monitor airborne particulate matter. Journal of the Air & Waste Management


Clark, N.N., Gautam, M., Rapp, B.L., Lyons, D.W., Graboski, M. S., McCormick,

R. L., Alleman, T.L., Norton, P., 1999. Diesel and CNG Transit Bus Emissions

Characterization by Two Chassis Dynamometer Laboratories: Results and Issues.

Society of Automotive Engineers SAE 1999-01-1469.

Clark, N.N., Lyons, D.W., Bata, R.M., Gautam, M., Wang, W.G., Norton, P.,

Chandler, K., 1997. Natural Gas and Diesel Transit Bus Emissions: Review and

Recent Data. Society of Automotive Engineers Tech. Pap. No. 973203.

Clark, N.N., Lyons, D.W., Rapp, B.L., Gautam, M., Wang, W.G., Norton, P.,

White, C., Chandler, C., 1998. Emissions from Trucks and Buses Powered by

Cummins L-10 Natural Gas Engines. Society of Automotive Engineers Tech. Pap.

No. 981393.

Clarke, A.G., Robertson, L.A., Hamilton, R.S., Gorbunov, B., 2004. A

Lagrangian model of the evolution of the particulate size distribution of vehicular

emissions. Science of the Total Environment 334-35, 197-206.



87







Brussels.

CONCAWE., 1998. A study of the number, size & mass of exhaust particles

emitted from european diesel and gasoline vehicles under steady-state and

european driving cycle conditions. CONCAWE, Brussels Report no. 98/51.

Converse, M.K.N., 2004. Roadside ultrafine and nanoparticle number

distributions in northern Central Valley, California and relationships to

meteorology and traffic. United States -- California, University of California,

Davis.

Cook, R., Touma, J. S., Beidler, A., Strum, M., 2006. Preparing highway

emissions inventories for urban scale modeling: A case study in Philadelphia.


88

Corsmeier, U., Imhof, D., Kohler, M., Kuhlwein, J., Kurtenbach, R., Petrea, M.,

Rosenbohm, E., Vogel, B., Vogt, U., 2005. Comparison of measured and model-

calculated real-world traffic emissions. Atmospheric Environment 39(31), 5760-

5775.

Dickens, C., Booker, D., 1998. Characterisation of vehicle emissions. Journal of

Aerosol Science 29(Supplement 1), 351.


Date verified 20 February 2008.

Dockery, D.W., Pope, C.A., Xu, X., Spengler, J.D., Ware, J.H., Fay, M.E., Ferris,

B.G., 1993. An Association between Air Pollution and Mortality in Six U.S.

Cities. The New England Journal of Medicine 329(24), 1753-1759.



Dora, C., 1999. A different route to health: implications of transport policies

British Medical Journal 318(7199), 1686-1689.

Earnshaw, K., Booker, D.R., 1998. City centre and industrial pollution

measurement using mass- and number-weighted instrumentation. Journal of

Aerosol Science 29(Supplement 1), 591-592.

89

ECJRC., (European Commission Joint Research Centre) 2002. Guidelines for


particulate matter for use in epidemiological studies. European Commission, Italy.

El-Shawarby, I., Ahn, K., Rakha, H., 2005. Comparative field evaluation of

vehicle cruise speed and acceleration level impacts on hot stabilized emissions.


Englert, N., 2004. Fine particles and human health - a review of epidemiological

studies. Toxicology Letters 149(1-3), 235-242.

Eninger, R. M., Rosenthal, F.S., 2004. Heart rate variability and particulate

exposure in vehicle maintenance workers: A pilot study. Journal of Occupational

and Environmental Hygiene 1(8), 493-499.

EPA (Environmental Protection Agency), 2004. Air Emissions Inventory South-

east Queensland Region. Brisbane.

EurActiv.com 2006. EURO 5 emissions standards for cars, EU News, Policy

Positions & EU Actors online. http://www.euractiv.com/en/transport/euro-5-

emissions-standards-cars/article-133325. Date verified 26 June 2008.

90






Feldpausch, P., Fiebig, M., Fritzsche, L., Petzold, A., 2006. Measurement of

ultrafine aerosol size distributions by a combination of diffusion screen separators

and condensation particle counters. Journal of Aerosol Science 37(5), 577-597.

Ferin, J., Oberdoerster, G., Penney, D.P., Soderholm, S.C., Gelein, R., Piper,

H.C., 1990. Increased pulmonary toxicity of ultrafine particles I. Particle

clearance, translocation, morphology. Journal of Aerosol Science. 21(3), 381-384.

Fine, P., Shen, S., Sioutas, C., 2004. Inferring the sources of fine and ultrafine

particulate matter at downwind receptor sites in the Los Angeles basin using

multiple continuous measurements. Aerosol Science and Technology 38(S1), 182-

195.

Gajananda, K., Kuniyal, J. C., Momin, G. A., Rao, P. S. P., Safai, P. D., Tiwari,

S., Ali, K., 2005. Trend of atmospheric aerosols over the north western

Himalayan region, India. Atmospheric Environment 39(27), 4817-4825.

Gehrig, R., Hill, M., Buchmann, B., Imhof, D., Weingartner, E., Baltensperger,

U., 2004. Separate determination of PM10 emission factors of road traffic for

91

tailpipe emissions and emissions from abrasion and resuspension processes.

International Journal of Environment & Pollution 22(3), 312-325.

Gertler, A.W., Gillies, J. A., Pierson, W. R., Rogers, C.F., Sagebiel, J.C., Abu-

Allaban, M., Coulombe, W., Tarnay, L., Cahill, T.A., 2002. Real-World

Particulate Matter and Gaseous Emissions from Motor Vehicles in a Highway

Tunnel. Health Effects Institute Research Report 107.

Gidhagen, L., Johansson, C., Langner, J., Foltescu, V.L., 2005. Urban scale

modeling of particle number concentration in Stockholm. Atmospheric

Environment 39(9), 1711-1725.

Gidhagen, L., Johansson, C., Langner, J., Olivares, G., 2004. Simulation of NOx

and ultrafine particles in a street canyon in Stockholm, Sweden. Atmospheric

Environment 38(14), 2029-2044.

Gidhagen, L., Johansson, C., Omstedt, G., Langner, J., Olivares, G., 2004. Model

simulations of NOx and ultrafine particles close to a Swedish highway.


Gidhagen, L., Johansson, C., Strom, J., Kristensson, A., Swietlicki, E., Pirjola, L.,

Hansson, H., 2003. Model simulation of ultrafine particles inside a road tunnel.


92

Giechaskiel, B., Ntziachristos, L., Samaras, Z., Scheer, V., Casati, R., Vogt, R.,

2005. Formation potential of vehicle exhaust nucleation mode particles on-road

and in the laboratory. Atmospheric Environment 39(18), 3191-3198.

Gillies, J.A., Gertler, A.W., Sagebiel, J. C., Dippel, W.A., 2001. On-road

particulate matter (PM2.5 and PM10) emissions in the Sepulveda Tunnel, Los

Angeles, California. Environmental Science & Technology 35(6), 1054-1063.

Goodwin, J.W.L., Salway, A.G., Murrells, T. P., Dore, C.J., Passant, N.R.,




Gramotnev, G., Brown, R., Ristovski, Z., Hitchins, J., Morawska, L., 2003.

Determination of average emission factors for vehicles on a busy road.


Gramotnev, G., Ristovski, Z., 2004. Experimental investigation of ultra-fine

particle size distribution near a busy road. Atmospheric Environment 38(12),

1767-1776.

Gramotnev, G., Ristovski, Z.D., Brown, R. J., Madl, P., 2004. New methods of

determination of average particle emission factors for two groups of vehicles on a

busy road. Atmospheric Environment 38(16), 2607-2610.

93

Gras, J. L., Ayers, G.P., 1983. Marine aerosol at southern mid-latitudes. Journal

of Geophysical Research 88(C15), 10661-10666.

Graskow, B.R., Kittelson, D. B., Abdul-Khalek, I.S., Ahmadi, M., Morris, J.E.,

1998. Characterisation of Exhaust Particulate Emissions from a Spark Ignition

Engine. SAE Paper 980528, 155-165.

Group, 1999. Source Apportionment of Airborne Particulate Matter in the United




Gutfinger, C., 1996. Aerosol measurement: Principles, techniques, and

applications : edited by K. Willeke and P. A. Baron., Van Nostrand Reinhold,

New York (1993). 876 pp.

Harris, S. J., Maricq, M.M., 2001. Signature size distributions for diesel and

gasoline engine exhaust particulate matter. Journal of Aerosol Science 32, 749-

764.

Harrison, R., Jones, M., Collins, G. 1999. Measurements of the Physical

Properties of Particles in the Urban Atmosphere. Atmospheric Environment 33,

309-321.

94

Harrison, R.M., Jones, M., Collins, G., 1999. Measurements of the physical

properties of particles in the urban atmosphere. Atmospheric Environment 33(2),

309-321.

Harrison, R.M., Shi, J.P., Zi, S., Khan, A., Mark, D., Kinnersley, R., Yin, J., 2000.

Measurement of number, mass and size distribution of particles in the

atmosphere. Philosophical Transactions of the Royal Society A: Mathematical,


Hasegawa, S., Hirabayashi, M., Kobayashi, S., Moriguchi, Y., Kondo, Y.,

Tanabe, K., Wakamatsu, S., 2005. Size distribution and characterization of

ultrafine particles in roadside atmosphere. Journal of Environmental Science and

Health Part a-Toxic/Hazardous Substances & Environmental Engineering, 2671-

2690.

Hausberger, S., Rodler, J., Sturm, P., Rexeis, M., 2003. Emission factors for

heavy-duty vehicles and validation by tunnel measurements. Atmospheric

Environment 37(37), 5237-5245.

Hazi, Y., 2001. Measurements of acidic sulfates and trace metals in fine and

ultrafine ambient particulate matter: Size distribution, number concentration and

source region. United States - New York, New York University.

95

Heintzenberg, J., Birmili, W., Wiedensohler, A., Nowak, A., Tuch, T., 2004.

Structure, variability and persistence of the submicrometer marine aerosol. Tellus

Series B - Chemical and Physical Meteorology 56, 357-367.

Herner, J.D., Aw, J., Gao, O., Chang, D. P., Kleeman, M.J. 2005. Size and

composition distribution of airborne particulate matter in northern California: I-

particulate mass, carbon, and water-soluble ions. Journal of the Air & Waste

Management Association 55(1), 30-51.

Hibberd, M.F., 2005. Vehicle NOx and PM10 Emission Factors from Sydney's

M5-East Tunnel. 17th International Clean Air & Environment Conference

proceedings, Hobart. Clean Air Society of Australia and New Zealand.

Hidy, G.M., 1975. Summary of the California Aerosol Characterization

Experiment. Journal of the Air Pollution Control Association 25, 1106-1114.

Hillamo, R., Kerminen, V. M., Aurela, M., Makela, J., Maenhaut, W., Leck, C.,

2001. Modal structure of chemical mass size distribution in the high Arctic

aerosol. Journal of Geophysical Research - Atmospheres 106(D21), 27555-27571.

Hitchins, J., Morawska, L., Wolff, R., Gilbert, D., 2000. Concentrations of

submicrometre particles from vehicle emissions near a major road. Atmospheric


96

Holmen, B., Chen, Z., Davila, A., Gao, O., Vikara, D.M. 2005. Particulate matter

emissions from Hybrid Diesel-electric and Conventional Diesel Transit Buses:

Fuel and Aftertreatment Effects. The University of Connecticut Report No. JHR

05-304.

Holmes, N.S., Morawska, L., Mengersen, K., Jayaratne, E.R., 2005. Spatial

distribution of submicrometre particles and CO in an urban microscale

environment. Atmospheric Environment 39(22), 3977-3988.

Hoppel, W. A., Larson, R., Vietti, M.A., 1990. Aerosol size distributions and

optical boundaries found in the marine boundary layer over the Atlantic Ocean

Journal of Geophysical Research 95(D4), 3659-3686.

Hueglin, C., Buchmann, B., Weber, R.O., 2006. Long-term observation of real-

world road traffic emission factors on a motorway in Switzerland. Atmospheric

Environment 40(20), 3696-3709.

Hussein, T., Hameri, K., Heikkinen, M., Kulmala, M., 2005. Indoor and outdoor

particle size characterisation at a family house in Espoo, Finland. Atmospheric


Hussein, T., Hameri, K. A., Aalto, P.P., Paatero, P., Kulmala, M., 2005. Modal

structure and spatial-temporal variations of urban and suburban aerosols in

Helsinki - Finland. Atmospheric Environment, 1655-1668.

97

Hussein, T., Puustinen, A., Aalto, P., Makela, J., Hameri, K., Kulmala, M., 2004.

Urban Aerosol Number Size Distributions. Atmospheric Chemistry and Physics

Discussions 4, 391-411.

Imhof, D., Weingartner, E., Ordonez, C., Gehrigt, R., Hill, N., Buchmann, B.,

Baltensperger, U., 2005. Real-world emission factors of fine and ultrafine aerosol

particles for different traffic situations in Switzerland. Environmental Science &

Technology 39(21), 8341-8350.

Imhof, D., Weingartner, E., Prevot, A., Ordonez, C., Kurtenbach, R., Wiesen, P.,

Rodler, J., Sturm, P., McCrae, I., Sjodin, A., Baltersperger, U., 2005. Aerosol and

NOx Emission Factors and Submicron Particle Number Size Distributions in Two

Road Tunnels with Different Traffic Regimes. Atmospheric Chemistry and

Physics Discussions 5, 5127-5166.

Imhof, D., Weingartner, E., Vogt, U., Dreiseidler, A., Rosenbohm, E., Scheer, V.,

Vogt, R., Nielsen, O. J., Kurtenbach, R., Corsmeier, U., Kohler, M.,

Baltensperger, U., 2005. Vertical distribution of aerosol particles and NOx close

to a motorway. Atmospheric Environment 39(31), 5710-5721.

Jaenicke, R., 1993. Tropospheric Aerosols. San Diego, USA, Academic Press.

Jamriska, M., Morawska, L., 2000. The effect of surface deposition, coagulation

and ventilation on submicrometer particles indoors. Clean Air and Envirionment

Conference, Sydney, Australia, 26-30 November 2000.

98

Jamriska, M., Morawska, L., 2001. A model for determination of motor vehicle

emission factors from on-road measurements with a focus on submicrometer

particles. Science of the Total Environment 264(3), 241-255.

Jamriska, M., Morawska, L., 2003. Quantitative Assessment of the Effect of

Surface Deposition and Coagulation on the Dynamics of Submicrometer Particles

Indoors. Aerosol Science and Technology 37(5), 425 - 436.



38(24), 6701-6709.

Jayaratne, E.R., Ristovski, Z.D., Meyer, N., Morawska, L., 2008. Particle and

Gaseous Emissions from Compressed Natural Gas and Ultralow Sulphur Diesel-

Fuelled Buses at Four Steady Engine Loads. Submitted to Science of the Total

Environment. .

Jayaratne, E.R., Verma, T.S., 2001. The impact of biomass burning on the

environmental aerosol concentration in Gaboroen, Botswana Atmospheric




techniques and applications,Van Nostrand Reinhold, New York. 55.

99

Johnson, J.P., Kittelson, D.B., Watts, W.F., Source apportionment of diesel and

spark ignition exhaust aerosol using on-road data from the Minneapolis

metropolitan area. Atmospheric Environment, 2111-2121.

Johnson, J.P., Kittelson, D. B., Watts, W.F., 2005. Source apportionment of diesel

and spark ignition exhaust aerosol using on-road data from the Minneapolis

metropolitan area. Atmospheric Environment 39(11), 2111-2121.

Jones, A.M., Harrison, R.M. 2006. Estimation of the emission factors of particle



Kado, N. Y., Okamoto, R. A., Kuzmicky, P. A., Kobayashi, R., Ayala, A., Gebel,

M. E., Rieger, P. L., Maddox, C., Zafonte, L., 2005. Emissions of toxic pollutants

from compressed natural gas and low sulfur diesel-fueled heavy-duty transit buses

tested over multiple driving cycles. Environmental Science & Technology 39(19),

7638-7649.

Kern, J.M., 2000. Inventory and prediction of heavy-duty diesel vehicle

emissions. United States -- West Virginia, West Virginia University.

Ketzel, M., Berkowicz, R., 2004. Modelling the fate of ultrafine particles from

exhaust pipe to rural background: an analysis of time scales for dilution,

coagulation and deposition. Atmospheric Environment 38(17), 2639-2652.

100

Ketzel, M., Wahlin, P., Berkowicz, R., Palmgren, F., 2003. Particle and trace gas

emission factors under urban driving conditions in Copenhagen based on street

and roof-level observations. Atmospheric Environment 37(20), 2735-2749.

Ketzel, M., Wahlin, P., Kristensson, A., Swietlicki, E., Berkowicz, R., Nielsen, O.

J., Palmgren, F., 2004. Particle size distribution and particle mass measurements

at urban, near-city and rural level in the Copenhagen area and Southern Sweden.

Atmospheric Chemistry and Physics 4, 281-292.

Keywood, M.D., Ayers, G. P., Gras, J.L., Gillett, R.W., Cohen, D.D. 1999.

Relationships between size segregated mass concentration data and ultrafine

particle number concentrations in urban areas. Atmospheric Environment 33(18),

2907-2913.

Khlystov, A., Kos, G.P.A., Ten Brink, H.M., Mirme, A., Tuch, T., Roth, C.,

Kreyling, W.G. 2001. Comparability of three spectrometers for monitoring urban

aerosol. Atmospheric Environment 35(11), 2045-2051.

Kirchstetter, T.W., Harley, R.A., Kreisberg, N.M., Stolzenberg, M.R., Hering,

S.V., 1999. On-road measurement of fine particle and nitrogen oxide emissions

from light- and heavy-duty motor vehicles. Atmospheric Environment 33, 2955-

2968.

101

Kittelson, D., 1998. Engines and Nanoparticles: a Review. Journal of Aerosol

Science 29(5), 575-588.

Kittelson, D., Johnson, J., Watts, W. F., Wei, Q., Drayton, M., Paulsen, D.,

Bukowiecki, N., 2000. Diesel aerosol sampling in the atmosphere. SAE

Technology Paper 2000-01-2212.

Kittelson, D.B., 1998. Engines and nanoparticles: a review. Journal of Aerosol

Science 29(5-6), 575-588.

Kittelson, D.B., Watts, W.F., Johnson, J., 2002. Diesel aerosol sampling

methodology. CRC E-43 Final report.

Kittelson, D.B., Watts, W.F., Johnson, J.P., 2004. Nanoparticle emissions on

Minnesota highways. Atmospheric Environment 38(1), 9-19.

Kousa, A., Kukkonen, J., Karppinen, A., Aarnio, P., Koskentalo, T., 2002. A

model for evaluating the population exposure to ambient air pollution in an urban

area. Atmospheric Environment 36(13), 2109-2119.

Kristensson, A., Johansson, C., Westerholm, R., Swietlicki, E., Gidhagen, L.,

Wideqvist, U., Vesely, V., 2004. Real-world traffic emission factors of gases and

particles measured in a road tunnel in Stockholm, Sweden. Atmospheric


102

Kukkonen, J., Bozo, L., Palmgren, F., Sokhi, R.S., 2003. Particulate matter in

urban air. Air Quality in Cities. Berlin, Springer-Verlag, Berlin 91-120.

Lancaster, M.J., Booker, D.R., 1998. Sampling concentrated aerosols diluter

design for ultrafine particles. Journal of Aerosol Science 29(Supplement 1), 333-

334.

Lanni, T., Frank, B. P., Tang, S., Rosenblatt, D., Lowell, D., 2003. Performance

and Emissions Evaluation of Compressed Natural Gas and Clean Diesel Buses at

New York City's Metropolitan Transit Authority. SAE 2003-01-0300.

Lighty, J.S., Veranth, J.M., Sarofim, A.F., 2000. Combustion Aerosols: Factors

Governing Their Size and Composition and Implications to Human Health.

Journal of the Air & Waste Management Association 50, 1565-1618.

Lowell, D.M., Parsley, W., Bush, C., Zupo, D.. 2003. Comparison of Clean Diesel

buses to CNG Buses. 9th Diesel Engine Emissions Reduction (DEER) Workshop,

Newport, RI, USA, 24-28 August.

Lundgren, D.A., Burton, R.M., 1995. Effect of particle size distribution on the cut

point between fine and coarse ambient mass fractions. Inhalation Toxicology

7(1 ), 131-148.

103

Macharis, C., Mierlo, J. van .Bossche, P. van den., 2007. Transportation Planning

and Technology. Combining Intermodal Transport With Electric Vehicles:

Towards More Sustainable Solutions 30(2-3), 311-323.

Magari, S.R., Hauser, R., Schwartz, J., Williams, P. L., Smith, T. J., Christiani,

D.C., 2001. Association of heart rate variability with occupational and

environmental exposure to particulate air pollution. Circulation 104(9), 986-991.

Makela, J., Koponen, I., Aalto, P., Kulmala, M., 2000. One Year Data of

Submicron Size Modes of Tropospheric Background Aerosols in Southern

Finland. Journal of Aerosol Science 31, 595-611.

Mark, D., Yin, J., Harrison, R., Booker, J., Moorcroft, S., 1998. Measurements of

PM10, PM2.5 particles at four outdoor sites in the UK. Journal of Aerosol Science

29(Supplement 1), 95-96.

Marshall, I.A., Booker, D.R., 1998. An aerosol concentration standard. Journal of

Aerosol Science 29(1-2), 227.

Mazzoleni, C., Kuhns, H. D., Moosmuller, H., Keislar, R.E., Barber, P.W.,

Robinson, N.F., Watson, J.G., 2004. On-road vehicle particulate matter and

gaseous emission distributions in Las Vegas, Nevada, compared with other areas.

Journal of the Air & Waste Management Association 54(6), 711-726.

104

McGregor, F., Ferreira, L., Morawska, L., 2003. Modelling of sub-micrometer

particle concentrations in free-flowing freeway traffic, Brisbane, Australia: some

empirical results. Transportation Research Part D: Transport and Environment

8(3), 229-241.

Mensink, C., De Vlieger, I., Nys, J., 2000. An urban transport emission model for

the Antwerp area. Atmospheric Environment 34(27), 4595-4602.

Meszaros, A., 1977. On the size distribution of atmospheric aerosol particles of

different composition. Atmospheric Environment 11(1075-1081).

Miller, T.L., Davis, W. T., Reed, G.D., Doraiswamy, P., Tang, A., Sanhueza, P.,

2001. Corrections to mileage accumulation rates for older vehicles and the effect

an air pollution emissions. Energy, Air Quality, and Fuels 2001, 49-55.

Mobley, J.D., Cadle, S.H., 2004. Innovative Methods for Emission Inventory



Mohr, M., Lehmann, U., Rutter, J., 2005. Comparison of mass-based and non-

mass-based particle measurement systems for ultra-low emissions from

automotive sources. Environmental Science & Technology 39(7), 2229-2238.

105

Monkkonen, P., Koponen, I., Lehtinen, K., Hameri, K., Uma, R., Kulmala, M.,

2005. Measurements in a highly polluted Asian mega city: Observations of

aerosol number size distribution, modal parameters and nucleation events.


Moosmuller, H., Arnott, W. P., Rogers, C. F., Bowen, J.L., Gillies, J.A., Pierson,

W. R., Collins, J.F., Durbin, T.D., Norbeck, J.M., 2001. Time resolved

characterization of diesel particulate emissions. 1. Instruments for particle mass

measurements. Environmental Science & Technology 35(4), 781-787.



Wiley-VCH, 297-318.

Morawska, L., 2004. Indoor particles, combustion products and fibres, The

Handbook of Environmental Chemistry. Springer-Verlag Heidelberg.

Morawska, L., Zhang, J. 2002. Combustion sources of particles. 1. Health

relevance and source signatures. Chemosphere 49(9), 1045-1058.

Morawska, L., Bofinger, N.D., Kocis, L., Nwankwoala, A., 1998. Submicrometer

and supermicrometer particles from diesel vehicle emissions. Environmental

Science & Technology 32(14), 2033-2042.

106

Morawska, L., Ferreira, L., Thomas, S., Jamriska, M., McGregor, F., 2004.

Quantification and modelling of particle emissions from motor vehicles in urban

environment: Final Report. Queensland University of Technology, Brisbane.

Morawska, L., Jamriska, M., Thomas, S., Ferreira, L., Mengersen, K., Wraith, D.,

McGregor, F., 2005. Quantification of particle number emission factors for motor

vehicles from on-road measurements. Environmental Science & Technology

39(23), 9130-9139.

Morawska, L., Jayaratne, E.R., Mengersen, K., Jamriska, M., Thomas, S., 2002.

Differences in airborne particle and gaseous concentrations in urban air between

weekdays and weekends. Atmospheric Environment 36(27), 4375-4383.

Morawska, L., Johnson, G., Ristovski, Z.D., Agranovski, V., 1999. Relation

between particle mass and number for submicrometer airborne particles.


Morawska, L., Keogh, D. U., Thomas, S. B., Mengersen, K., 2008. Modality in



Morawska, L., Moore, M., Ristovski, Z., 2004. Health impacts of ultrafine

particles: Desktop literature review and analysis. Canberra, Department of

Environment and Heritage.

107

Morawska, L., Ristovski, Z., Ayoko, G.A., Jayaratne, E.R., Lim, M., 2001. Report

of a short investigation of emissions from diesel vehicles operating on low and

ultralow sulphur content fuel. Prepared for BP Australia by Queensland

University of Technology. Queensland University of Technology, Brisbane.

Morawska, L., Ristovski, Z., Jayaratne, E. R., Keogh, D. U., Ling, X., 2008.


characteristics, ambient processing and implications on human exposure

Submitted to Atmospheric Environment.

Morawska, L., Salthammer, T., 2003. Motor vehicle emissions as a source of

indoor particles. Indoor Environment, Airborne Particles and Settled Dust. L.

Morawska and T. Salthammer. Weinheim, Germany, Wiley-VCH.

Morawska, L., Schwela, D., 1998. Airborne particles and health implications:

Directions for the future. Journal of Aerosol Science 29(Supplement 1), 167.

Morawska, L., Thomas, S., Bofinger, N.D., Wainwright, D., Neale, D., 1998.

Comprehensive characterisation of aerosol in a subtropical urban atmosphere:

particle size distribution and correlation with gaseous pollutants. Atmospheric


Morawska, L., Thomas, S., Jamriska, M., Johnson, G., 1999. The modality of

particle size distributions of environmental aerosols. Atmospheric Environment

33(27), 4401-4411.

108

Morawska, L., Zhang, J., 2002. Combustion sources of particles. 1. Health

relevance and source signatures. Chemosphere 49(9), 1045-1058.

Murray, A.T., Davis, R., Stimson, R.J., Ferreira, L., 1998. Public Transportation

Access. Transportation Research Part D: Transport and Environment 3(5), 319-

328.

Nanzetta, M.K., Holmen, B.A., 2004. Roadside particle number, distributions and

relationships between number concentrations, meteorology, and traffic along a

northern California freeway. Journal of the Air & Waste Management Association

54(5), 540-554.

Nazaroff, W., Ligocki, M., Ma, T., Cass, C., 1990. Particle Deposition in

Museums, Comparison of Modelling and Measurement Results. Aerosol Science

& Technology 13(332-348).

NEPC., 2000. Proposed Diesel Vehicle Emissions National Environment



Neususs, C., Wex, H., Birmili, W., Wiedensohler, A., Koziar, C., Busch, B.,

Bruggemann, E., Gnauk, T., Ebert, M., Covert, D.S., 2002. Characterization and

parameterization of atmospheric particle number-, mass, and chemical-size

distributions in central Europe during LACE 98 and MINT - art. no. 8127. Journal

of Geophysical Research - Atmospheres 107(D21), 8127.

109

Nichols, A., Read, P., Booker, D., 1998. Valid analytical measurements:

particulates and aerosols. Journal of Aerosol Science 29(Supplement 2), S863-

S864.

Nielsen, O.A., Knudsen, M.A., 2006. Uncertainty in traffic models. European

Transport Conference Strasbourg, France

Niemeier, D.A., Zheng, Y., Kear, T., 2004. UCDrive: a new gridded mobile

source emission inventory model. Atmospheric Environment 38(2), 305-319.

NPI (National Pollutant Inventory), Department of the Environment, Water, Heritage and the Arts, Australian Government, http://www.npi.gov.au/index.html Date verified 1 July 2008.

Ntziachristos, L., Samaras, Z., Eggleston, S., Goriben, N., Hassel, D., Hickman,

A.J., Joumard, R., Rijkeboer, R., White, L., Zierock, K.H., 2000. COPERT III


and emission factors (version 2.1). Technical report prepared by the European

Environment Agency, Copenhagen, Report 49.

O'Dowd, C., Becker, E., Kulmala, M., 2001. Mid-Latitude North-Atlantic Aerosol

Characteristics in Clean and Polluted Air. Atmospheric Research 58, 167-185.

110

Oberdoerster, G., Ferin, J., Finkelstein, G., Wade, P., Corson, N., 1990. Increased

pulmonary toxicity of ultrafine particles? II. Lung lavage studies. Journal of

Aerosol Science 21(3), 384-387.

Oberdoerster, G., Sharp, Z., Atudorei, V., Elder, A., Gelein, R., Kreyling, W.,

Cox, C., 2004. Translocation of inhaled ultrafine particles to the brain. Inhalation

Toxicology 16, 437-445.

Office of Urban Management, 2004. Draft South East Queensland Regional Plan:

For Consultation. Brisbane Department of Local Government, Planning, Sport &

Recreation, Queensland Government.

Ortuzar, J. de., Willumsen, L.G., 2001. Modelling Transport. John Wiley & Sons

Inc

Paatero, P., Aalto, P., Picciotto, S., Bellander, T., Castano, G., Cattani, G., Cyrys,

J., Kulmala, M., Lanki, T., Nyberg, F., 2005. Estimating time series of aerosol

particle number concentrations in the five HEAPSS cities on the basis of

measured air pollution and meteorological variables. Atmospheric Environment

39(12), 2261-2273.

Pakkanen, T.A., Kerminen, V.M., Loukkola, K., Hillamo, R. E., Aarnio, P.,

Koskentalo, T., Maenhaut, W., 2003. Size distributions of mass and chemical

components in street-level and rooftop PM1 particles in Helsinki. Atmospheric

Environment 37(12), 1673-1690.

111

Pakkanen, T.A., Kerminen, V.M., Korhonen, C.H., Hillamo, R.E., Aarnio, P.,

Koskentalo, T., Maenhaut, W., 2001. Urban and rural ultrafine (PM0.1) particles

in the Helsinki area. Atmospheric Environment 35(27), 4593-4607.

Parkhurst, G., 2004. Air quality and the environmental transport policy discourse

in Oxford. Transportation Research Part D: Transport and Environment 9(6), 419-

436.



Peace, H., Owen, B., Raper, D.W., 2004. Comparison of road traffic emission

factors and testing by comparison of modelled and measured ambient air quality

data. Science of The Total Environment 334-335, 385-395.

Pirjola, L., Parviainen, H., Hussein, T., Valli, A., Hameri, K., Aalto, P., Virtanen,

A., Keskinen, J., Pakkanen, T., Makela, J., Hillamo, R., 2004. "Sniffer" - A novel

tool for chasing vehicles and measuring traffic pollutants. Atmospheric


Pohjola, S. K., Savela, K., Kuusimaki, L., Kanno, T., Kawanishi, M., Weyand, E.,

2004. Polycyclic aromatic hydrocarbons of diesel and gasoline exhaust and DNA

adduct detection in calf thymus DNA and lymphocyte DNA of workers exposed

to diesel exhaust. Polycyclic Aromatic Compounds 24(4-5), 451-465.

112

Pokharel, S.S., Bishop, G.A., Stedman, D.H., 2002. An on-road motor vehicle

emissions inventory for Denver: an efficient alternative to modeling. Atmospheric

Environment 36(33), 5177-5184.

Pope, C.A., Burnett, R.T., Thun, M.J., Calle, E.E., Krewski, D., Ito, K., Thurston,

G.D., 2002. Lung cancer, cardiopulmonary mortality, and long-term exposure to

fine particulate air pollution. Journal of the American Medical Association

287(9), 1132-1141.

Pope, C.A., Dockery, D.W., 2006. Health Effects of Fine Particulate Air




Thurston, G. D., 2002. Lung cancer, cardiopulmonary mortality, and long-term


Association 287(9), 1132-1141.

Porter, J.N., Clarke, A.D., 1997. Aerosol size distribution models based on in situ

measurements. Journal of Geophysical Research 102(D5), 6035-6045.

113

Querol, X., Alastuey, A., de la Rosa, J., Sanchez-de-la-Campa, A., Plana, F.,

Ruiz, C. R., 2002. Source apportionment analysis of atmospheric particulates in

an industrialised urban site in southwestern Spain. Atmospheric Environment

36(19), 3113-3125.

Reddy, M. S., Venkataraman, C., 2002. Inventory of aerosol and sulphur dioxide

emissions from India: I--Fossil fuel combustion. Atmospheric Environment 36(4),

677-697.

Reynolds, A.W., Broderick, B. M., 2000. Predicting real-time traffic-related PM10

and PM2.5 emissions and concentrations in urban areas. Journal of Aerosol

Science 31(Supplement 1), 250-251.

Ristovski, Z., Jayaratne, E.R., Lim, M., Ayoko, G.A., Morawska, L., 2006.

Influence of diesel fuel sulphur on the nanoparticle emissions from city buses


Ristovski, Z., Morawska, L., Ayoko, G. A., Johnson, G., Gilbert, D., Greenaway,

C., 2004. Emissions from a vehicle fitted to operate on either petrol or

compressed natural gas. Science of The Total Environment 323(1-3), 179-194.

Ristovski, Z.D., Jayaratne, E.R., Morawska, L., Ayoko, G. A., Lim, M., 2005.

Particle and carbon dioxide emissions from passenger vehicles operating on

unleaded petrol and LPG fuel. Science of The Total Environment 345(1-3), 93-98.

114

Ristovski, Z.D., Morawska, L., 1998. Emission of submicrometer particles from

spark ignition vehicles. Journal of Aerosol Science 29(Supplement 2), S1001-

S1002.

Ristovski, Z.D., Morawska, L., Ayoko, G. A., Jayaratne, E. R., Lim, M., 2002.

Final report of a comparative investigation of particle and gaseous emissions from

twelve in-service B.C.C. buses operating on 50 and 500 ppm sulphur diesel fuel.

Queensland University of Technology, Brisbane.

Ristovski, Z.D., Morawska, L., Bofinger, N.D., Hitchins, J., 1998. Submicrometer

and supermicrometer particulate emission from spark ignition vehicles.


Romilly, P., 1999. Substitution of bus for car travel in urban Britain: an economic

evaluation of bus and car exhaust emission and other costs. Transportation

Research Part D-Transport and Environment 4(2), 109-125.

Rosenbohm, E., Vogt, R., Scheer, V., Nielsen, O., Drieseidler, A., Baumbach, G.,

Imhof, D., Baltensperger, U., Fuchs, J., Jaeschke, W., 2005. Particulate size

distributions and mass measured at a motorway during the BAB II campaign.


SAE., 2001. Performance and Durability Evaluation of Continuously

Regenerating Particulate Filters on Diesel powered Urban Transit Buses at NY

City Transit. Society of Automotive Engineers SAE 2001-01-0511.

115

SAE., 2002. Performance and Durability of Continuously Regenerating

Particulate Filters on Diesel powered Urban Transit Buses at NY City Transit -

Part II. Society of Automotive Engineers SAE 2002-01-0430.

SAE., 2002. Speciation of Organic Compounds from the Exhaust of Trucks and

Buses: Effect of Fuel and After-treatment on Vehicle Emission Profiles. Society

of Automotive Engineers SAE 2002-01-2873.

SAE., 2002. Year-Long Evaluation of Trucks and Buses Equipped with Passive

Diesel Diesel Particulate Filters. Society of Automotive Engineers SAE 2002-01-

0433.

SAE., 2003. Oxidation catalyst effect on CBG Transit Bus Emissions. Society of

Automotive Engineers SAE 2003-01-1900.

SAE., 2003. Performance and Emissions Evaluation of Compressed Natural Gas

and Clean Diesel Buses at New York City's Metropolitan Transit Authority.


Salma, I., Dal Maso, M., Kulmala, M., Zaray, G., 2002. Modal characteristics of

particulate matter in urban atmospheric aerosols. Microchemical Journal 73, 19-

26.

116

Salma, I., Ocskay, R., Raes, N., Maenhaut, W., 2005. Fine structure of mass size

distributions in an urban environment. Atmospheric Environment 39(29), 5363-

5374.

Scheffe, H., 1959. The Analysis of Variance, John Wiley & Sons, Inc.

Schifter, I., Diaz, L., Mugica, V., Lopez-Salinas, E., 2005. Fuel-based motor



Schmid, H., Pucher, E., Ellinger, R., Biebl, P., Puxbaum, H., 2001. Decadal

reductions of traffic emissions on a transit route in Austria - results of the

Tauerntunnel experiment 1997. Atmospheric Environment 35(21), 3585-3593.

Seaton, A., MacNee, W., Donaldson, K., Godden, D., 1995. Particulate air

pollution and acute health effects. Lancet 345, 176-178.

Sem, G.J., 2002. Design and performance characteristics of three continuous-flow

condensation particle counters: a summary. Atmospheric Research 62(3-4), 267-

294.

SEQHTS., 2003-2004. South-East Queensland Household Travel Survey

(SEQHTS) (Brisbane, Gold Coast and Sunshine Coast Area). Brisbane

Queensland Transport

117

Shah, S.D., Cocker, D.R., Miller, J. W., Norbeck, J.M., 2004. Emission rates of

particulate matter and elemental and organic carbon from in-use diesel engines.


Sharma, M., Agarwal, A. K., Bharathi, K.V.L., 2005. Characterization of exhaust

particulates from diesel engine. Atmospheric Environment 39(17), 3023-3028.

Shi, J., Evans, D., Khan, A., Harrison, R., 2001. Sources and Concentration of

Nanoparticles ( < 10 nm Diameter) in the Urban Atmosphere. Atmospheric


Shi, J., Harrison, R.M., 1999. Investigation of ultrafine particle formation during

diesel exhaust dilution Environmental Science & Technology 33, 3730-3736.

Shi, J.P., Khan, A.A., Harrison, R.M., 1999. Measurements of ultrafine particle

concentrations and size distribution in the urban atmosphere The Science of the


SKM (Sinclair Knight Merz), 2006. Twice the Task: A review of Australia's

freight transport tasks. Melbourne, Victoria, National Transport Commission.




118

Smit, R., Smokers, R., Rabe, E., 2007. A new modelling approach for road traffic



Sokhi, R.S., 2005. Fourth international conference on urban air quality--

measurement, modelling and management, 25-28 March 2003, Prague, Czech

Republic. Atmospheric Environment 39(15), 2695-2696.

Somers, C.M., McCarry, B.E., Malek, F., Quinn, J.S., 2004. Reduction of

particulate air pollution lowers the rist of heritable mutations in mice. Science,

1008-1010.

Stedman, J.R., Linehan, E., Conlan, B., 2000. Receptor modelling of PM10

concentrations at a United Kingdom national network monitoring site in central

London. Atmospheric Environment 35(2), 297-304.

Sturm, P. J., Baltensperger, U., Bacher, M., Lechner, B., Hausberger, S., Heiden,

B., Imhof, D., Weingartner, E., Prevot, A.S.H., Kurtenbach, R., Wiesen, P., 2003.

Roadside measurements of particulate matter size distribution. Atmospheric


Sturm, P.J., Pucher, K., Sudy, C.Almbauer, R.A., 1996. Determination of traffic

emissions--intercomparison of different calculation methods. Science of The

Total Environment 189-190, 187-196.

119

Sundqvist, K. L., Wingfors, H., Brorstom-Lunden, E., Wiberg, K., 2004. Air-sea

gas exchange of HCHs and PCBs and enantiomers of [alpha]-HCH in the Kattegat

Sea region. Environmental Pollution 128(1-2), 73-83.

Swiss Agency for the Environment, Forests and Landscape (SAEFL), 2004. Air

Pollutant emissions from Road Transport 1980-2030 Environmental Series No.

355. Berne SAEFL.

Thomas, S., Morawska, L., 2002. Size-selected particles in an urban atmosphere

of Brisbane, Australia. Atmospheric Environment 36(26), 4277-4288.

TNO., 1997. Particulate Matter Emissions (PM10, PM2.5, PM<0.1) in Europe in


Tran, T. V., Ng, Y. L., Denison, L., 2003. Emission Factors for In-Service

Vehicles Using Citylink Tunnel. Proceedings of the National Clean Air

Conference, Newcastle.

Translink, 2007. Bus patronage and bus fleet statistics. Queensland Transport,

Brisbane.

Tuch, T., Mirme, A., Tamm, E., Heinrich, J., Heyder, J., Brand, P., Roth, C.,

Wichmann, H.E., Pekkanen, J., Kreyling, W.G., 2000. Comparison of two

particle-size spectrometers for ambient aerosol measurements. Atmospheric


120

Tunved, P., Nilsson, E., Hansson, H., Strom, J., 2005. Aerosol characteristics of

air masses in Northern Europe: Influences of location, transport, sinks and

sources. Journal of Geophysical Research - Atmospheres 110(D7), 7201.

Ubanwa, B., Burnette, A., Kishan, S., Fritz, S.G., 2003. Exhaust particulate matter

emission factors and deterioration rate for in-use motor vehicles. Journal of

Engineering for Gas Turbines and Power-Transactions of the Asme 125(2), 513-

523.

Unal, A., Frey, H. C., Rouphail, N.M., 2004. Quantification of Highway Vehicle

Emissions Hot Spots based on on-board measurements. Journal of the Air &

Waste Management Association 54, 130-140.

USEPA., 1993. User's Guide to MOBILE5A, Mobile source emissions factor

model, U.S. Environmental Protection Agency.

USEPA., 1995. Compilation of Air Pollutant Emission Factors, 5th edn, AP-42,

North Carolina.

USEPA., 2004. Air quality criteria for particulate matter. Washington DC, US

Environmental Protection Agency, 600/P-99/002aF-bF.

121

Van Dingenen, R., Putaud, J. P., Martins-Dos Santos, S., Raes, F., 2005. Physical

aerosol properties and their relation to air mass origin at Monte Cimone (Italy)

during the first MINATROC campaign. Atmospheric Chemistry and Physics 5,

2203-2226.

Vardoulakis, S., Gonzalez-Flesca, N., Fisher, B. E. A., Pericleous, K., 2005.

Spatial variability of air pollution in the vicinity of a permanent monitoring

station in central Paris. Atmospheric Environment 39(15), 2725-2736.

Vedal, S., 1997. Ambient particles and health: lines that divide. Journal of the Air

& Waste Management Association 47, 551-581.

Venkatram, A., Fitz, D., Bumiller, K., Du, S. M., Boeck, M., Ganguly, C., 1999.

Using a dispersion model to estimate emission rates of particulate matter from

paved roads. Atmospheric Environment 33(7), 1093-1102.

Vogt, R., Kirchner, U., Scheer, V., Hinz, K. P., Trimborn, A., Spengler, B., 2003.

Identification of diesel exhaust particles at an Autobahn, urban and rural location

using single-particle mass spectrometry. Journal of Aerosol Science 34(3), 319-

337.

Wahlin, P., Palmgren, F., Van Dingenen, R., 2001. Experimental studies of

ultrafine particles in streets and the relationship to traffic Atmospheric


122

Walker, J. L., Li, J., Srinivasan, S., Bolduc, D., 2008. Travel Demand Models in

the Developed World: Correcting for Measurement Errors Transportation

Research Board 87th Annual Meeting Washington.

Watson, J.G., Chow, J. C., Houck, J.E., 2001. PM2.5 chemical source profiles for

vehicle exhaust, vegetative burning, geological material, and coal burning in

Northwestern Colorado during 1995. Chemosphere 43(8), 1141-1151.

Watson, J.G., Zhu, T., Chow, J.C., Engelbrecht, J., Fujita, E. M., Wilson, W.E.,

2002. Receptor modeling application framework for particle source

apportionment. Chemosphere 49(9), 1093-1136.

Wayne, W.S., Clark, N.N., Nine, R.D., Elefante, D., 2004. A comparison of

emissions and fuel economy from hybrid-electric and conventional-drive transit

buses. Energy & Fuels 18(1), 257-270.

Wehner, B., Birmili, W., Gnauk, T., Wiedensohler, A., 2002. Particle number size

distributions in a street canyon and their transformation into the urban-air

background: measurements and a simple model study. Atmospheric Environment

36(13), 2215-2223.

Weijers, E.P., Khlystov, A. Y., Kos, G.P.A., Erisman, J.W., 2004. Variability of

particulate matter concentrations along roads and motorways determined by a

moving measurement unit. Atmospheric Environment 38(19), 2993-3002.

123

Weingartner, E., Nyeki, S., Baltensperger, U., 1999. Seasonal diurnal variation of

aerosol size distribution (10<D<750 nm) at a high-alpine site. Journal of

Geophysical Research - Atmospheres 104(D21), 26809-26820.

WHO., 2005. Guidelines for Air Quality. World Health Organization, Geneva.

Wiedensohler, A., Wehner, B., Birmili, W., 2002. Aerosol number concentrations

and size distributions at mountain rural, urban-influenced rural and urban-

background sites in Germany. Journal of Aerosol Medicine 15(2), 237-243.

Willeke, K., Baron, P.A., 1993. Aerosol Measurement: Principles, Techniques,

and Applications. John Wiley & Sons, New York.

Wingfors, H., Sjodin, A., Haglund, P., Brorstrom-Lunden, E., 2001.

Characterisation and determination of profiles of polycyclic aromatic

hydrocarbons in a traffic tunnel in Gothenburg, Sweden. Atmospheric

Environment 35(36), 6361-6369.

Winther, M., 1998. Petrol passenger car emissions calculated with different

emission models. Science of the Total Environment 224(1-3), 149-160.

Woo, K.S., 2003. Measurement of atmospheric aerosols: Size distributions of

nanoparticles, estimation of size distribution moments and control of relative

humidity. United States -- Minnesota, University of Minnesota.

124

Wu, Y., Hao, J., Fu, L., Wang, Z.. Tang, U., 2002. Vertical and horizontal profiles

of airborne particulate matter near major roads in Macao, China. Atmospheric

Environment 36(31), 4907-4918.

Xie, S., Yu, T., Zhang, Y., Zeng, L., Qi, L., Tang, X., 2005. Characteristics of

PM10, SO2, NOx and O3 in ambient air during the dust storm period in Beijing.

Science of The Total Environment 345(1-3), 153-164.

Young, L.H., Keeler, G.J., 2004. Characterization of ultrafine particle number

concentration and size distribution during a summer campaign in southwest

Detroit. Journal of the Air & Waste Management Association 54(9), 1079-1090.

Zhang, J. J., Morawska, L., 2002. Combustion sources of particles: 2. Emission

factors and measurement methods. Chemosphere 49(9), 1059-1074.

Zhang, K., 2004. Ambient and Plume Processing of Atmospheric Ultrafine

Particles. PhD Thesis, University of California, Davis.

Zhang, K.M., Wexler, A.S., 2004. Evolution of particle number distribution near

roadways--Part I: analysis of aerosol dynamics and its implications for engine

emission measurement. Atmospheric Environment 38(38), 6643-6653.

125

Zhang, K. M., Wexler, A. S., Niemeier, D. A., Zhu, Y. F., Hinds, W. C., Sioutas,

C., 2005. Evolution of particle number distribution near roadways. Part III:

Traffic, analysis and on-road size resolved particulate emission factors.


Zhang, K. M., Wexler, A. S., Zhu, Y. F., Hinds, W. C., Sioutas, C., 2004.

Evolution of particle number distribution near roadways. Part II: the 'Road-to-

Ambient' process. Atmospheric Environment 38(38), 6655-6665.

Zhong, M., Hanson, B.L., 2008. GIS-Based Travel Demand Modeling for

Estimating Traffic on Low-Class Roads Transportation Research Board 87th

Annual Meeting Washington

Zhu, Y. 2003. Ultrafine particle and freeways. PhD Thesis, University of

California, Los Angeles.

Zhu, Y., Hinds, W. C., Kim, S., Shen, S., Sioutas, C., 2002. Study of ultrafine

particles near a major highway with heavy-duty diesel traffic. Atmospheric

Environment 36(27), 4323-4335.

Zhu, Y., Hinds, W. C., Kim, S., Shen, S., Sioutas, C., 2004. Study of ultrafine


Environment 36(27), 4323-4335.

126

Zhu, Y., Hinds, W.C., Kim, S., Sioutas, C., 2002. Concentration and size

distribution of ultrafine particles near a major highway. Journal of the Air &

Waste Management Association 52(9), 1032-1042.

Zhu, Y., Hinds, W.C., Shen, S., Sioutas, C., 2004. Seasonal trends of

concentration and size distribution of ultrafine particles near major highways in

Los Angeles. Aerosol Science & Technology 38(S1), 5-13.

Zhu, Y., Kuhn, T., Mayo, P., Hinds, W.C., 2006. Comparison of daytime and

nighttime concentration profiles and size distributions of ultrafine particles near a

major highway. Environmental Science & Technology 40, 2531-2536.

Zhu, Y.F., Hinds, W.C., 2005. Predicting particle number concentrations near a

highway based on vertical concentration profile. Atmospheric Environment 39(8),

1557-1566.

Zhu, Y.F., Hinds, W.C., Kim, S., Sioutas, C., 2002. Concentration and size


Waste Management Association 52(9), 1032-1042.

Zhu, Y. F., Hinds, W.C., Shen, S., Sioutas, C., 2004. Seasonal trends of

concentration and size distribution of ultrafine particles near major highways in

Los Angeles. Aerosol Science and Technology, 5-13.

127

CHAPTER 3. STATISTICAL TECHNIQUES USED IN

THIS THESIS

3.1. INTRODUCTION

This chapter discusses the statistical methods used in this study, their

strengths and limitations, and outlines alternative methods which are

available but were not considered suitable.

The statistical techniques used in this thesis included:-

(i) Kolmogorov-Smirnov (K-S) test (modality paper, Chapter 4)

(ii) Construction of 95% confidence intervals (modality paper, Chapter 4);

(iii) Trapezoidal rule for integration of the area under a curve (modality

paper, Chapter 4);

(iv) Linear regression for continuous variables (emission factor paper,

Chapter 5.1);

(v) Multifactor Analysis of Variance (ANOVA) for categorical variables

(emission factor paper, Chapter 5.1);

(vi) The stepwise selection technique for statistical model selection

(emission factor paper, Chapter 5.1)

(vii) Scheffe’s multiple comparison tests (emission factor paper,

Chapter 5.1).

128

The nature of air quality data

Aerosol particle size distributions are rarely normally distributed (Baron and

Willeke 2001). These distributions have a tendency to tail off as particle size

increases (Ruzer and Harley 2004) and often their standard deviations are large

compared with the mean particle size, a condition not permitted in a non-negative

normal distribution, hence they are described mathematically more effectively by

lognormal distributions (Baron and Willeke 2001). When particle size

distributions measured in ambient and indoor air environments are plotted on a

logarithmic scale, the distribution is approximated by a Gaussian distribution

(bell-shaped or approximately normal) (Baron and Willeke 2001).

Two basic approaches are currently used for determining sources of air pollution,

including particulate matter, the top-down or receptor-based source

apportionment approach, or the bottom-up or source-based method (Guttikunda et

al. 2008). The receptor-based approach is based on the premise that particulate

matter sources often exhibit characteristic profiles or chemical patterns; whereas

with source-based models the pollution sources are identified and then emission

factors are estimated based on source model information (Guttikunda et al. 2008).

129

3.2. STATISTICAL TECHNIQUES USED IN THIS THESIS

3.2.1. Kolmogorov-Smirnov (K-S) test

This method tests for statistically significant differences between two

cumulative distributions at a chosen significance level (Liao 2002), for

example, between measured aerosol particle size distributions. Unlike other non-

parametric tests, such as the Mann Whitney-U test that assess the equality of two

medians, the K-S test assesses the equivalence of the whole distributions. This

type of test is useful for data with non-standard distributions. For normally

distributed data, for example, the mean and standard deviation would be

considered sufficient statistics for the distribution, but this is not the case for

much of the air quality data considered in this thesis.

The use and applicability of the K-S test for comparing aerosol particle size

distributions which are characterised by a large number of sizing channels (or

bins) has been described by Heitbrink et al. (1991). The K-S test is quite effective

if data are assigned to a large number of intervals, generally (n » 10) which have

the same boundaries (Baron and Willeke 2001). The Aerodynamic Particle Sizer

instrumentation has a large number of sizing channels and Heitbrink et al. (1991)

successfully used the K-S test to analyse its particle count data (Baron and

Willeke 2001). Details of the application of the technique for comparison of

aerosol particle size distributions measured by Scanning Mobility Particle Sizer

instrumentation are provided in Morawska et al. (1999).

130

The non-parametric method selected for use was the K-S test, which is

commonly used in aerosol science, and was employed to examine

statistically significant differences between particle size distributions for

different aerosol types (modality paper, Chapter 4).

3.2.2. Construction of 95% confidence intervals

The reliability of a sample statistic as an estimator of a parameter of interest is

commonly expressed in terms of a confidence interval. A 95% confidence

interval for a population mean, for example, is centred around the

sample mean with upper and lower bounds that are determined by the sample

variance, sample size and 5% Type I error rate.

The interpretation of this interval is such that if samples of n size were repeatedly

taken from a population and corresponding confidence intervals were constructed

for each sample, it could be expected that 95% of the intervals between these

limits would contain the true mean (Sokal and Rohlf 2000). These limits, in this

case, are referred to as the lower and upper 95% confidence limits of the mean,

termed confidence interval (Sokal and Rohlf 2000).

Confidence intervals can be used for examining differences in data sets and when

comparing the means of different groups. For example, a lack of overlap between

confidence intervals for different data sets can confirm the existence of groups

with statistically significantly different means. In this thesis 95% confidence

intervals were constructed for data set clusters situated closest to, on either side,

131

of 1µm (Chapter 4), to determine if an overlap existed between these confidence

intervals, and whether these data sets had statistically significantly different

means.

A 95% confidence interval was constructed to examine whether an overlap

existed between particle mode data sets for different particle metrics located

close to 1 µm, either above or below 1 µm (modality paper, Chapter 4).

3.2.3. Trapezoidal rule for integration of the area under a curve

Classical formulas for numerical integration include Simpson’s rule and the

trapezoidal rule (Dyer and Dyer 2008). Simpson’s rule is discussed later in

Section 3.3 under alternative approaches.

The trapezoidal rule is suitable to apply to sections of different widths (Chapra

2002), such as to different particle bin sizes, which represent particle diameter

size ranges used for analysing aerosol measurement data. For example, the

fractional contributions to different particle size ranges can be calculated by

integrating the area under the curve of each particle size range using the

trapezoidal rule, since this rule takes specific account of the different widths of

the x-axis in each bin. The trapezoidal rule may be applied to either the original

data scale or the log transformed data. The area in each bin size is approximated

to be the shape of a trapezium, and the sum of these areas for all bin sizes

provides an approximation of the total area.

132

The trapezoidal rule was selected as the numerical integration method to

estimate the area under the curve relating to different particle size ranges

(modality paper, Chapter 4).

3.2.4. Development of statistical models using linear regression and

ANOVA

When examining the relationship between a response variable and a number of

independent variables, or covariates, regression and ANOVA (Analysis of

Variance) are considered reasonable models (Hosmer and Lemeshow 2000).

Regression methods are commonly used to explain outcome differences for

differing groups (Liao 2002). Model choice can depend on the type of covariates

(categorical or continuous) and aim of the analysis (for example, test of

significant effect of covariates on the response variable or estimation of

magnitude of the effect).

3.2.5. Linear regression for continuous variables

Regression analysis depends on the general assumption that an underlying

deterministic or systematic relationship exists between the response variable and

the covariates (Olden and Jackson 2000).

In a general linear regression model an assumption is made that the residuals are

normally distributed and independently centred around zero with constant

variance. A residual is the difference between the predicted value under the

model and the observed value. Moreover, in the linear regression model, it is

133

assumed that the outcome variable is continuous (Hosmer and Lemeshow 2000).

The response or covariates can be transformed to meet this assumption, or a

generalisation of the regression model can be used eg., logistic regression for

binary responses, generalised linear model for non-normal data or non-linear

models. The significance of the influence of covariates on the response can be

tested using ANOVA, F and t-tests.

Linear regression was used to develop statistical relationships, and

associated ANOVA, F and t-tests were used to assess the statistical

significance of the covariates on the response of interest (ie., emission

factors) (emission factor paper, Chapter 5.1).

3.2.6. Multifactor Analysis of Variance (ANOVA) for categorical

variables

ANOVA (Analysis of Variance) is a general statistical method which tests

complex hypotheses. Two applications of ANOVA were considered in this thesis

(i) to test the equality of multiple group means and (ii) to test the statistical

significance of relationships between covariates and responses in linear regression

(Liao 2002), where the total variation in the response is split into that explained

by the model and the residual variation. Variation within groups and between

groups, variation between the sample means and the inherent variability within

each sample are compared (Devore and Farnum 2005). In multifactor designs, the

interaction between two or more factors can also be examined (Devore and

Farnum 2005).

134

In the ANOVA model an assumption is made that the data are normally

distributed (Hays and Winkler 1970).

Multifactor ANOVA was used to test the statistical significance of

categorical variables in development of the statistical models.

ANOVA was also used for regression analysis (emission factor

paper, Chapter 5.1).

3.2.7. The stepwise selection technique for statistical model selection

The stepwise selection technique alternates between forward selection and

backward elimination (described later in this Chapter) and at each step of forward

inclusion of an appropriate variable, the model then assesses the significance of

all its variables, and those found to be no longer significant are taken out of the

model before the next forward selection step (Olden and Jackson 2000). The

order in which variables are added does not affect the final model. The problem of

multicollinearity can be avoided to some extent using this method, as the model

will check to ensure that the variable/s is needed in the model.

135

The method employed was stepwise selection to identify the best

statistical model (emission factor paper Chapter 5.1).

3.2.8. Scheffe’s multiple comparison tests

ANOVA (see previous section) only tests whether group means are the same, and

if rejected confirms that at least one mean is different from the others but provides

no information as to which mean this is. Often the aim of a study is to identify

which of the means are different, and hence in these cases multiple comparison

methods are used.

In this thesis the multiple comparison tests were unplanned, and comparison tests

were decided upon once the results of the analysis of the statistical models

developed in this study were completed.

Unplanned multiple comparison tests include comparison tests between all

possible pairs of means. There are two main methods for unplanned tests,

Tukey’s and Scheffe. Both methods have their strengths and weaknesses,

however key differences relate to the fact that Scheffe’s method can undertake

multiple comparison for both equal and unequal sample sizes, whereas Tukey’s

method is based on equal sample sizes (Sahai and Ageel 2000). Secondly,

assumptions of homoscedasticity and normality are more critical for Tukey’s as

opposed to Scheffe’s method (Liao 2002).

136

Scheffe’s method is used for multiple comparison tests (Scheffe 1953, 1959). It is

a powerful method that is suitable to use where unplanned contrasts are sought to

be made among means that have been suggested by experimental results, and is

mathematically equivalent to the sum of squares simultaneous test procedure

(Sokal and Rohlf 2000).

Scheffe’s method was used to identify statistically significant differences

between means of subclasses of different categorical variables (emission

factor paper, Chapter 5.1).

3.3. ALTERNATIVE APPROACHES CONSIDERED BUT NOT USED

IN THIS THESIS

The following approaches were considered but were not used in this thesis study.

3.3.1. Principal Component Analysis

Principal component analysis (PCA) and, more generally factor analysis (FA), are

multivariate statistical techniques used in atmospheric science, where variables

with similar characteristics are grouped together into factors, and the techniques

produce a small set of linear combinations, with the aim of retaining as much of

the original information as is possible (Lin et al. 2008).

137

PCA has been widely used in Europe as a receptor model to suggest likely sources

of particulate matter (Viana et al. 2008). For example, PCA has been used to

analyse concentrations of polycyclic aromatic hydrocarbons in Central Taiwan

(Lin et al. 2008) and long term multisite air quality monitoring site data for PM10,

Nox, CO, SO2 and O3 in Birmingham and Athens (Statheropoulos et al. 1998).

The limitations of PCA with respect to the applications and inference have also

been acknowledged. For example, it has been suggested that in terms of

atmospheric data, factor analysis as a method can attempt to extract more

information than really exists (Henry 1987). In addition, a common problem that

has been reported to occur with components or factors is that these can represent

mixtures of emission sources, instead of representing clear, independent source

profiles (Viana et al. 2008). In its favour, Lioy et al. (1989) found that PCA based

on data which is highly time-resolved can lead to resolving more sources.

PCA/FA was not considered a relevant technique for these research objectives, as

the focus in this PhD research was not on identifying lower dimension indices. In

addition, source apportionment techniques require careful inference interpretation.

Statistical techniques used in this thesis study relate to research which focused on

examination of the characteristic modality within particle size distributions as a

possible basis for developing air quality regulation (Chapter 4); and derivation of

a comprehensive set of particle emission factors for different vehicle types that

can be used in transport modelling and health impact assessments of urban fleets

(Chapter 5.1).

138

3.3.2. Non-parametric methods of statistical comparison

Some major non-parametric methods of statistical comparison include

Kolmogorov-Smirnov (K-S) test and the Mann-Whitney U- test (Liao 2002). Non-

parametric test methods that are used for testing the equality of two sample means

differ from the two sample t-test as the latter makes an assumption that the

difference between the two sample means is normally distributed, whereas this

assumption is not made in non-parametric test methods (Liao 2002).

The Mann-Whitney U-test aims to test the equality of medians from two

populations, and this approach can be applied to ranked or ordered nonmetric data

(Liao 2002). This test was considered less powerful than parametric tests or K-S

tests for the type of data encountered in this thesis.

3.3.3. Techniques for integrating the area under a curve

The Simpson’s rule requires that points be equally spaced (Chapra 2002).

Although this rule can be used for either the original data or the log transformed

data, if the data are highly skewed and highly peaked, it may be considered more

effective to use log scale data which requires less adjustment of the bin sizes.

One of the advantages of using the trapezoidal rule is that this method accounts

for the difference in bin size widths. This feature is not available under Simpson’s

rule, hence the trapezoidal rule was the preferred choice. Bin sizes can vary for

data on the linear scale or log transformed scale, especially for highly skewed

data, such as particle diameters. This can lead to different estimates of area, even

139

under monotone transformation of the data. This concern motivated the choice of

the trapezoidal rule over the Simpson’s rule.

3.3.4. Techniques for selecting variables in statistical model development

Possible approaches for inclusion or exclusion of significant variables in the

development of statistical models include forward selection, backward

elimination and stepwise selection. The forward selection and backward

elimination methods do not allow variables to be added back into a model for

evaluation, however with the stepwise technique variables are able to be added

back into the model for evaluation.

The forward selection technique adds each possible predictor variable in turn to

the model and the most statistically significant, if available, are included in the

model, until no variables remain that are statistically significant (Olden and

Jackson 2000).

The backward elimination technique begins with a model that contains all the

possible predictor variables and the variable that is least significant is deleted,

one by one, until none of the variables are left which are statistically

non-significant (Olden and Jackson 2000).

Often, but not always, all three approaches produce identical models. It is

important to note that under any approach, the model is chosen on the basis of

goodness of fit measures and is not assumed to be the “correct” model.

140

Moreover, it is important to distinguish between a model for description of current

data versus a model used to develop predictions; in the latter case, a model with

poorer goodness of fit to the observed data but greater robustness to alternative

datasets may be preferred.

3.3.5. Multiple comparison methods

Multiple comparison methods include Least Significant Differences Test,

Bonferroni’s method, Tukey’s method and Scheffe’s method. The differences

in these methods relate to two different aspects (i) whether these methods are

suitable for planned or unplanned comparisons; and (ii) their approach for

controlling experimentwise error rate, and the degree of protection against a

Type I error. A Type I error is asserted to have occurred if the null

hypothesis is rejected when it is true (Sokal and Rohlf 2000).

Non-orthoganility exists whenever more tests are conducted than degrees of

freedom existing between groups, and where comparisons are non-orthogonal

they do not exhibit independence (Sokal and Rohlf 2000). For example, when

planned comparisons are found to be non-orthogonal, tests with adjusted

values for Type I error are used (Sokal and Rohlf 2000). In this case a

conservative approach is taken and the Type I error of the statistic of

significance is lowered for each comparison so that the chance of making a

Type I error in the whole series of comparison tests does not exceed the level

of significance (Sokal and Rohlf 2000). This probability is referred to as the

experimentwise error rate (Sokal and Rohlf 2000).

141

Planned comparison tests (or a-priori comparisons) relate to tests designed and

chosen independently of experimental results, they are planned before the

experiment has been conducted (Sokal and Rohlf 2000), and a decision is made

concerning how many tests will be done. Comparison tests suggested as a result

of a completed experiment are referred to as unplanned comparison tests (or

posteriori tests) and include comparison tests between all possible pairs of means

(Sokal and Rohlf 2000).

The Least Significant Differences Test is only valid for planned comparisons,

Bonferroni’s method is a conservative test used for planned comparisons, Tukey’s

method is suitable for unplanned comparisons, and Scheffe’s method is a powerful

unplanned comparison test (Sokal and Rohlf 2000).

The Least Significant Differences Test involves, in the first instance, conducting

an ANOVA to determine whether there are statistically significant differences

among groups, which will be suggested by the results of an overall F-test at a

selected significance level (Liao 2002). Group means are compared and declared

to be significantly different if greater than a specified magnitude that is based on

the desired overall error rate and the number of comparisons to be made.

Bonferroni’s method is used for planned comparisons (Sokal and Rohlf 2000).

This method can be used for linear combinations or pairwise contrasts prior to an

ANOVA with either unequal or equal sample sizes, however, the set of

combinations may not be infinite, as is permitted with Scheffe’s method, but can

142

exceed the number allowed in Tukey’s procedure (Liao 2002). The power of the

Bonferroni method is lowered if more than one null hypothesis is found to be

false (Sokal and Rohlf 2000).

Tukey’s method compares all likely pairwise differences of means (Liao 2002)

and is suitable for unplanned comparison tests (Sahai and Ageel 2000).

However, as discussed in the previous section (3.2.8), Tukey’s method requires

equal sample sizes (Sahai and Ageel 2000) and assumptions of homoscedasticity

and normality are more critical for Tukey’s as opposed to Scheffe’s method (Liao

2002), hence Scheffe was considered a better choice in the thesis for the conduct

of unplanned multiple comparison tests.

143

3.4. REFERENCES

Baron, P. A.,Willeke, K., 2001. Aerosol Measurement, Principles,

Techniques and Applications, 2nd edn. New York, John Wiley & Sons, Inc.

Chapra, S., 2002. Numerical Methods for Engineers with Software and

Programming Applications. New York McGraw Hill.

Devore, J., Farnum, N., 2005. Applied Statistics for Engineers and Scientists.

2nd edn. , Thomson Brooks, California.

Dyer, S. A., Dyer, J.S., 2008. bythenumbers - Numerical integration.

Instrumentation & Measurement Magazine 11(2), 47-49.

Guttikunda, S., Wells, G. J., Johnson, T. M., Artaxo, P., Bond, T. C., Russel, A.

G., Watson, J. G., West, J., 2008. Source Apportionment of Particulate Matter for

Air Quality Management: Review of Techniques and Applications in Developing

Countries Joint UNDP/World Bank Energy Sector Management Assistance

Programme (ESMAP)

Hays, W. L., Winkler, R. L., 1970. Statistics New York, Holt, Reinhart and

Winston

Heitbrink, W., Baron, P., Willeke, K., 1991. Coincidence in time-of-flight aerosol

spectrometers: Phantom particle creation. Aerosol Science and Technology 14,

112-126.

Henry, R.C., 1987. Current factor analysis receptor models are ill-posed.


144

Hosmer, D. W., Lemeshow, S., 2000. Applied Logistic Regression. New

York, John Wiley & Sons, Inc.

Liao, T.M., 2002. Statistical Group Comparison. Canada, John Wiley & Sons

Inc.

Lin, M., Rau, J., Tseng, H., Wey, M., Chu, C., Lin, Y., Wei, M., Lee, C.,

2008. Characterizing PAH emission concentrations in ambient air during a

large-scale joss paper open-burning event Journal of Hazardous Materials

156, 223-229.

Lioy, P. J., Zelenka, M. P., Cheng, M. D., Reiss, N. M., Wilson, W. E., 1989.

The effect of sampling duration of the ability to resolve source types using

factor analysis. Atmospheric Environment 23, 239-254.

Morawska, L., Thomas, S., Gilber, D., Greenaway, C., Rijnders, E., 1999. A

study of the horizontal and vertical profile of submicrometer particles in

relation to a busy road. Atmospheric Environment 33 (8), 1261-1274.

Olden, J. D., Jackson, D. A., 2000. Torturing data for the sake of generality:

How valid are our regression models? Ecoscience 7(4), 501-510.

Ruzer, L. S., Harley, N. H., 2004. Aerosols Handbook Management,

Dosimetry and Health Effects. Florida, USA CRC Press

145

Sahai, H., Ageel, M. I., 2000. The Analysis of Variance: Fied, Random, and

Mixed Models. Boston, Brikhauser.

Scheffe, H., 1953. A method for judging all contrasts in the analysis of variance.

Biometrika 40, 87-104.

Scheffe, H,. 1959. The Analysis of Variance. New York Wiley.

Sokal, R.R., Rohlf, F.J., 2000. Biometry: The Principles and Practice of Statistics

in Biological Research, 3rd edn., W.H. Freeman and Company, New York.

Statheropoulos, M., Vassiliadis, N., Pappa, A., 1998. Principal component and

canonical correlation analysis for examining air pollution and meteorological

data. Atmospheric Environment 32, 1087-1095.

Viana, M., Kuhlbusch, T. A. J., Querol, X., Alastuey, A., Harrison, R. M., Hopke,

P. K., Winiwarter, W., Vallius, M., Szidat, S., Prevot, A. S. H., Hueglin, C.,

Bloemen, H., Wahlin, P., Vecchi, R., Miranda, A. I., Kasper-Giebl, A., Maenhaut,

W., Hitzenberger, R., 2008. Source apportionment of particulate matter in Europe:

A review of methods and results. Journal of Aerosol Science 39, 827-849.

146

CHAPTER 4

MODALITY IN AMBIENT PARTICLE SIZE DISTRIBUTIONS

AND ITS POTENTIAL AS A BASIS FOR DEVELOPING

AIR QUALITY REGULATION

Lidia Morawska1, Diane U. Keogh1, Stephen B. Thomas2,

Kerrie Mengersen3

1 International Laboratory for Air Quality and Health, Queensland

University of Technology, Brisbane, Queensland, Australia

2 ENSR Australia, Fortitude Valley, Queensland, Australia

3 School of Mathematical Sciences, Queensland University of

Technology, Brisbane, Queensland, Australia

(2008) Atmospheric Environment 42 (7), 1617-1628

147

STATEMENT OF JOINT AUTHORSHIP

Title: Modality in ambient particle size distributions and its potential

as a basis for developing air quality regulation

Authors: Lidia Morawska, Diane U. Keogh, Stephen B. Thomas, and

Kerrie Mengersen

Lidia Morawska

Developed the experimental design and scientific method for the South-

East Queensland study and interpreted the data. Wrote the majority of the

manuscript.

Diane U. Keogh (candidate)

Developed the scientific method for the worldwide review of modes. Data

collection, processing, analysis and interpretation of modal data and wrote

this section of the manuscript. Contributed to the manuscript.

Stephen B. Thomas

Contributed to development of the South-East Queensland study

experimental design and scientific method. Conducted the South-East

Queensland measurements and contributed to interpretation of the data.

Contributed to the manuscript.

Kerrie Mengersen

Developed the experimental design of the statistical tests.


148

ABSTRACT

Current ambient air quality standards are mass-based and restricted to PM2.5

and PM10 fractions. The major contribution to both PM2.5 and PM10 fractions is

from particles belonging to the coarse mode and generated by mechanical

processes. These standards are thus unable to effectively control particle

concentrations from combustion sources, such as motor vehicles and power

plants, which tend to emit very small particles that are almost entirely

respirable and in the submicron range, and dominate the nucleation and

accumulation modes, which contribute much less to particle mass

concentration.

The aim of this work was to examine whether PM1 and PM10 would be a more

effective combination of mass standards than PM2.5 (dominant in the nucleation

and accumulation modes) and PM10 (dominant in the coarse mode) in controlling

combustion related ambient particles, as well as those originating from

mechanical processes. Firstly, a large body of data on particle size distributions

in a range of environments in South East Queensland, Australia was analysed,

with an aim of identifying the relation between modality in the distributions and

sources of particles belonging to different modes. The analyses included a matrix

of the following elements: particle volume and number distributions, type of

environment and locations of the modes in the range of PM1, PM2.5 and PM10

fractions.

Secondly, with the same aim, 600 published modal location values relating to

number, surface area, volume and mass size distributions for a range of

environments worldwide, were analysed. The analysis identified a clear and

149

distinct separation between the location of the modes for a substantial number of

environments worldwide and particle metrics, which suggests that modality in

particle size distributions may be a parameter that has potential to be used in the

development of PM1 air quality guidelines and standards. Based on these

analyses, implications for choosing different mass standards for airborne

particulate matter are discussed in the paper.

Keywords: modality of particle distribution, ambient aerosol, PM1, PM2.5,

PM10, air quality regulation.

4.1. INTRODUCTION

Various aspects are considered when developing ambient air quality standards of

which the most important are the exposure-response relationship and the

characteristics of the pollutant, which determine the exposure. Size distribution is

one of the key characteristics of ambient particulate matter, on one hand related to

particle formation and post-formation processes and, on the other hand,

determining the fate of particles in the air and the likelihood of their deposition in

the human respiratory tract. Current ambient air quality standards for PM2.5 and

PM10 fractions are based in part on a scientific basis, but also in part on the data

and limitations of the size ranges measured by equipment at the time of setting the

standards. PM2.5 and PM10 fractions are mass concentration of particles with

aerodynamic diameters smaller than 2.5 and 10 μm, respectively. PM1 is the mass

concentration of particles with aerodynamic diameters smaller than 1 μm.

150

Size-selective inlets which remove particles that exceed a specific aerodynamic

diameter are characterised by sampling effectiveness curves which show the

fraction of particles passing through as a function of aerodynamic diameter.

Sampling effectiveness is summarised by the 50% cut-point (relating to the

diameter that represents half the particles passing through the inlet) and includes a

slope function, representing the contribution from different particle sizes above and

below the 50% cut-point, because an exact sharp cut-point cannot be achieved in

practice (Baron and Willeke 2001).

The PM2.5 fraction is sometimes referred to as fine particles, while the difference

between PM10 and PM2.5 is sometimes referred to as coarse particles. Particles

larger than 10 μm tend to have atmospheric lifetimes that are relatively short

(Harrison et al. 2000) and are of lesser significance from the health point of view

since they are mostly removed by the nose. Prior to setting the PM2.5 standard,

the US EPA conducted an extensive examination of the available data on particle

size distributions. The Air Quality Criteria for Particulate Matter (EPA 1996)

contains a comprehensive discussion of the relative merits of PM1 and PM2.5.

The decision by the US EPA to introduce 2.5 µm as the upper end of the

boundary for fine particles and as a basis for a standard (Reference US Federal

Register) was strongly influenced by the fact that the available epidemiological

data at the time were obtained using PM2.5 measurements (Dockery et al. 1993).

151

An alternative approach in classification of the particles for the purpose of

developing control measures, is to consider location of the modes in particle size

distributions, which relate to the contribution from different pollution sources. A

mode may be defined as a peak in the lognormal function of the number or mass

distribution of an atmospheric aerosol (John 1993). A number of investigations

into the variation of the aerosol size spectrum over a variety of size intervals have

been made. Three terms have been introduced for atmospheric aerosol size

distribution in terms of modal diameters; these classifications focused on particle

size and production mechanism and were the nucleation mode (< 0.1 µm),

accumulation mode (0.1-1 µm) and coarse particle mode (> 1 µm) (Jaenicke

1993).

However it is acknowledged that the location of the modes generally depends on

the metric being referred to, such as particle number, surface area, volume or

mass, and modes will also change depending on the mathematical transformation

method used. For example, Whitby’s model of particle volume size distribution

(1978) was based primarily on atmospheric aerosol number distributions in the

size range 0.01-6 µm, which when transformed to volume distributions, revealed

three modal size ranges, with the nuclei mode (< 0.1 µm), the accumulation mode

(0.1-2 µm) and coarse particle mode (> 2 µm) (Baron and Willeke 2001). More

recently, studies with instruments extending the small size limit to 3 nm have

shown that the nuclei mode needs to be separated into a nucleation mode (< 0.01

µm) and an Aitken nuclei mode (0.01-0.1 µm) (USEPA 2004).

152

In environments affected by anthropogenic influences most of the nucleation

mode particles originate either from the condensation and coagulation of hot,

highly supersaturated vapours released during combustion or arise from the

condensation and coagulation of low vapour pressure materials formed in the

atmosphere by photochemically initiated processes. Coagulation and

heterogeneous nucleation tend to accumulate the aerosol in the accumulation

mode. Nucleation, Aitken, and accumulation modes contain soot, acid

condensates, sulfates and nitrates, as well as trace metals and other toxins. Most

anthropogenic pollution sources are combustion-related and generate particles

with diameters < 1 µm (Jamriska and Morawska 2000). Submicrometer particles

(diameters < 1 µm) represent most particle matter that is dispersed in urban

environments in terms of particle number concentrations (Morawska et al. 1998;

Nazaroff et al. 1990). Almost all particles in the coarse particle mode originate

from natural and anthropogenic mechanical processes, including grinding,

breaking and wear of material and dust resuspension.

The currently accepted division between fine and coarse particles of 2.5 µm does

not follow the natural division between modes attributable to different types of

sources. Instead, it tends to cut through the mode originating from mechanical

processes. It has been shown, however, that there is usually a clear separation

between the accumulation and coarse modes around 1 µm or somewhat above,

where the mass of particles belonging to these two modes is at a minimum

(Lundgren and Burton 1995). Therefore the rationale behind the classification of

one micrometer as a division between fine and coarse particles in particle mass

and particle volume size distributions would be that it constitutes a natural

division between particles generated mainly from combustion and photochemical

153

processes and particles generated from mechanical processes. Obviously this

definition, as any, would still be somewhat arbitrary, as nature itself does not

provide a perfect division.

Knowledge and understanding of the presence and location of modes in particle

distributions is of importance not only for understanding the mechanisms of

atmospheric processes, but also, importantly, for exposure and risk assessment,

particularly for setting standards and guidelines for air quality. The disparity

between what the standards divide into fine and coarse particles and what nature

divides into modes originating from different sources may make control of

particles more difficult and in fact may also be less desirable from the health

point of view.

The aim of the work reported in this paper was to analyse the available

information on modal locations in ambient particle size distributions and, based

on this, to explore the potential for PM1 as an effective mass standard together

with PM10 in controlling contributions from different types of air pollution

sources.

4.2. METHODS AND TECHNIQUES

The analysis conducted within the scope of this work was divided into two steps.

Firstly, characteristics of the modality in particle size distributions for a range of

environments in South East Queensland, Australia were investigated to examine

the relationship between fractional contribution of mass from different modes in

particle size distribution (and thus from different sources) to PM1, PM2.5 and

PM10. South East Queensland was chosen because for this environment the

154

authors have detailed information available on particle size distributions, with

thousands of spectra collected. The conclusions as to modality of particle size

distributions reached by Morawska et al. (1999), as well as the averaged size

distributions obtained, served as bases for the analyses presented in this paper.

Particle modal characteristics, their dependence on local conditions in South East

Queensland and their variability with time were reviewed by Morawska et al.

(1999). This paper also provided a detailed analysis of the modal characteristics

of over 6,000 particle size spectra collected over a period of three years for a

range of environments, including marine, modified background, suburban

background, traffic influenced, urban influenced and vegetation burning. Details

concerning the classification of these environments are provided in Morawska et

al. (1999). Measurements of size distributions in the size range 0.016 to 30 µm

were conducted using SMPS and APS instrumentation. Spectra corresponding to

one sample were combined, normalised and smoothed using a chi-square fitting

procedure to give one distribution, and Kolmogorov-Smirnov (K-S) tests were

used to compare measured aerosol size distributions (For details see Morawska et

al. 1999). The aim of the analysis was to combine the distributions from two

instruments measuring submicrometer and supermicrometer particle size

distributions for the calculation of the volume size distributions and to allow

interpretation of the modal characteristics for each environment studied. The

focus of that work was on source identification and identification of the

relationship between the sources and size distribution of particles generated. For

each environment there was a clear division between the accumulation and coarse

modes, but not between the nucleation and accumulation modes. As the densities

of the aerosols were not known, only volume and not mass distributions were

155

calculated. As there is, however, a direct correlation between mass and volume

distributions, where density acts as a scaling factor, modality displayed by

volume and mass distributions are the same.

For each distribution referred to above, in this work, the fractional contribution of

N+A (nucleation and accumulation) and C (coarse) modes to volumes of PM1,

PM2.5 and PM10 were calculated. Ultrafine particles (diameters of < 0.1 µm) tend

to dominate particle number and make a significant contribution to surface area but

little to mass, with the cube dependence of volume (and therefore mass) resulting

in significantly different particle size distributions for particle number and mass

distributions (Harrison et al. 2000). In simple terms, it is likely that the majority of

particle number is in the transient nucleation and Aitken modes, particle surface

area in the accumulation mode, and volume and mass divided between the

accumulation and coarse particle modes (Harrison et al. 2000).

The relative contributions were calculated by summing the volumes under the

peaks of the modes with the boundary between the accumulation and coarse

modes being taken as the sharp visible division on the figures. The total volume

of the individual modes was not calculated, which could have been done by

extrapolating the curves that describe the mode down to zero on the horizontal

axis or fitting a statistical mixture model. Instead the contributions to PM1, PM2.5

and PM10 were calculated assuming sharp cut-offs. While due to the limitations in

the measurement techniques these cut-offs are not sharp, it was considered that

for the purpose of the assessment conducted in this work this assumption would

not affect the overall outcome of the assessment, but would significantly simplify

the calculations. Moreover, where modes overlap, the concentration levels are

156

usually a few orders of magnitude lower than in the peaks and therefore the

contribution from the volumes not included was considered to be negligible.

Fractional contributions to the modes were calculated by integrating the area

under the curve using the trapezoidal rule. The trapezoidal rule takes account of

the different width of the x-axis in each bin, and may be applied to either the

original data scale or the log scale. Log scale values were used in our

calculations and calculations were undertaken using Origin (Version 6.0).

Secondly, an analysis of modal locations reported in international literature was

conducted to determine whether a clear and distinct separation occurs in the log-

transformed data between the modes around 1 µm, in different environments and

for different metrics. This was evaluated by constructing a 95% confidence

interval for the mean of those modal values lying below 1 µm, and a second 95%

confidence interval for the mean of those values lying above 1 µm. The means of

the two groups were asserted to be significantly different if these two confidence

intervals did not overlap. Moreover, the value of 1 µm was determined to be an

effective threshold if it separated the two intervals, so that it was larger than the

confidence interval for the smaller mean, and smaller than the confidence interval

for the larger mean. This was evaluated by testing for the existence of two modal

groups with significantly different means, as determined by non-overlapping

95% confidence intervals. This analysis was used to ascertain which

environments and metrics may possibly be suited to PM1 standards.

157

4.3. RESULTS AND DISCUSSION

4.3.1. Contribution of the modes in South East Queensland to PM1, PM2.5,

PM10

Figure 4.1 presents averaged size distributions for different types of

environments in South East Queensland in terms of both number and volume size

distribution. To enable the distinctions between the modes and identification of

the size boundaries of the modes, both types of spectra are presented in double

logarithmic scale. A vertical line shows the location of the division according to

the boundary of PM2.5 and coarse particles.

Figure 4.1. Normalised number and volume size distributions in South East Queensland, Australia (a) traffic influenced aerosol (b) urban influenced aerosol (c) vegetation burning influenced aerosol (d) marine influenced aerosol (e) modified background aerosol (f) suburban background aerosol. N + A (nucleation and accumulation modes), Coarse (coarse mode). 157

158

A general conclusion that can be made from inspection of the distributions

presented in Figure 4.1 is that in all of the environments there is a good

separation between accumulation and coarse particle modes, but that this

separation occurs at or below 1 μm. Harrison et al. (2000) found a similar

separation at around 1 μm in measured particle size distributions from suburban

Birmingham, United Kingdom in terms of number, surface area and volume. It

can be seen in the South East Queensland environments that in all cases the

division at 2.5 µm cuts across the coarse particle mode, close to its peak.

Inspection of the spectra presented in Figure 4.1 reveals that for traffic influenced

aerosols as well as for urban influenced, suburban background and modified

background aerosols both number and volume distributions are bimodal, with the

majority of particle number being associated with the fine particle mode

(nucleation, Aitken and accumulation regions, N+At +A), while most of the mass

is associated with the coarse particle mode (C). Since our instruments only

extended to 16 nm, we do not have information on the nucleation mode and the

Aitken mode and the accumulation mode are not clearly separated in all size

distributions. Similarity between the modal locations in the N+At and A region in

these environments leads to the conclusion that in this urban environment

automobile exhausts are the major contributors.

The coarse particles, on the other hand, may more likely result from a number of

different sources and not just the predominant road dust source for aerosols sampled

adjacent to the freeway, as indicated by differences between the shapes of the size

distribution curves. For example, the distributions for modified background aerosols

are considered representative of the influences by biogenic sources, with the broad

159

width of the coarse particle mode being a reflection of the presence of particles

originating from plant emissions in the aerosols. The existence of several more

peaks in the suburban aerosol is likely the result of several background aerosol

sources, which in urban type locations are usually masked by the presence of much

stronger sources, such as traffic emissions.

There is also close similarity between the shapes of the size distribution curves of

vegetation burning influenced aerosols and the traffic and urban aerosols in South

East Queensland. There is a difference, however, in the width of the modes, with

the N+A mode of the vegetation burning influenced aerosol being at a larger

particle size than the other two aerosols typically encountered in urban

environments. There are a number of peaks present within the N+A modes of the

marine influenced aerosol as presented in Figure 4.1. They include free

troposphere nuclei mode, effects related to the influence of cloud processing of

coagulating nuclei and the sea salt component of marine aerosols. While the

majority of the particles in the number size distribution are smaller than 1 μm

diameter, the majority of the volume is in fact occupied by particles with

diameters greater than 1 µm.

For each of the distributions presented in Figure 4.1, fractional contribution of

N+A and C modes to the volumes of PM1, PM2.5 and PM10 was calculated,

assuming, as discussed above, sharp cut-offs of 1, 2.5 and 10 µm. These

contributions are shown in Table 4.1 and form the basis of our study conclusions.

The most obvious conclusion from Table 4.1 is that PM10 volume in all

environments, except vegetation burning, can be attributed mainly to particles

from the coarse mode (C), that is, particles generated from mechanical processes.

160

Contribution from combustion processes to PM10 is negligible. Volume from

N+A modes for vegetation burning contributes about 50% to PM10 volume.

Similarly to PM10, in most of the environments the Coarse (C) mode has the

strongest contribution to PM2.5 volume. However, in traffic influenced and

vegetation burning the contribution from N+A is substantial. In the case of

vegetation burning, N+A volume has a stronger contribution to PM2.5 volume

compared to C volume and to its contribution to PM10 volume. Contribution

from N+A mode volume to PM1 is dominant for traffic influenced, vegetation

burning, marine influenced and modified background.

Table 4.1 Percent contribution of N+A and C modes by mass to

PM1, PM2.5 and PM10 in South East Queensland, Australia

PM1

% contribution

(by mass)

PM2.5

% contribution

(by mass)

PM10

% contribution

(by mass)

Environment

Type N+A C N+A C N+A C

Traffic Influenced 99 1 61 39 24 76

Urban Influenced 49 51 3 97 < 1 > 99

Vegetation burning 100 0 90 10 52 48

Marine influenced 82 18 2 98 <1 >99

Modified

background

88 12 13 87 < 1 > 99

Suburban

background

38 62 1 99 < 1 > 99

161

4.3.2. Modal locations in the published literature

The review of published studies revealed 605 modes reported for particle

number, surface area, volume and mass size distributions. Since access to the

data used by other authors was not available, the examination focused on the

location of these reported modes. Moreover, for the purposes of this study, only

modal location values ≤ 10 µm were extracted. Of the 605 modes identified, five

occurred at ≥ 10 µm in particle volume and were not included in the review,

leaving a total of 600 examined in our study. Particle concentrations and their

relative variations were not considered in this study. The published values

spanned diverse environments, and included background, central european

aerosol, desert, fires, forest, high alpine, marine and modified marine, modified

background, north-west Himalayas, rural/continental, suburban, traffic-

influenced, urban-influenced, urban background and vegetation burning

environments. Tables 4.2-4.5 present listings of the international studies

reviewed.

Figure 4.2 presents a compilation of all 600 modal location values from the

analysis for a range of environments and metrics. Modal location value ranges for

the different metrics spanned from 0.006 to 3 µm for number; 0.02 to 3.5 µm for

surface area; 0.008 to 10 µm for volume and 0.06 to 7.8 µm in mass particle size

distributions. Approximately 98% of number modal location values occurred at ≤ 1

µm. Surface area modal locations showed a similar pattern to mass but were

shifted to the right, to the larger size ranges.

162

Table 4.2 International literature reviewed to identify the location of the modes in a number of different environments worldwide for particle number size distributions

Condition Researchers Particle size range measured

(μm )

Location

Central European Aerosol Neususs et al. 2002 0.003-10

Leipzig and Berlin, Germany

Continental background Birmili et al. 1999 & 2001 0.003-0.8 Melpitz, Germany Continental background Wiedensohler et al. 2002 0.003-0.8 Melpitz, Germany

Forest Makela et al. 2000 0.003-0.5 Southern Finland

Forest Tunved et al. 2005 0.01-0.5 Hyytiala, Matorova Station, Varrio, Finland

High Alpine Weingartner et al. 1999 0.018-0.75 Jungfrauhoch, Switzerland, 3580m

Marine Heintzenberg et al. 2004 0.0031-0.65 Cape Grim, Australia

Marine Heintzenberg et al. 2004 0.0031-0.79 Sagres, Portugal

Marine Heintzenberg et al. 2004 0.003-0.9

N/S Atlantic, Indian Ocean, Pacific, Yellow Sea, Sea of Japan

Marine & modified marine a Morawska et al. 1999 0.016-30 Brisbane, Australia Marine & polluted air masses O'Dowd et al. 2001 0.005-150 Mace Head, Ireland

Modified background Morawska et al. 1999 0.016-30 Brisbane, Australia

Rural Tunved et al. 2005 0.01-0.452 Aspvreten, Sweden

Suburban Hussein et al. 2005 0.003-0.6 Finland

Suburban background Morawska et al. 1999 0.016-30 Brisbane, Australia

Traffic-influenced Morawska et al. 1999 0.016-30 Brisbane, Australia

Traffic-influenced Pirjola et al. 2004 0.007-10 Helsinki, Finland

Traffic-influenced Rosenbohm et al. 2005

0.0107-10 northside) 0.0202-10

(southside) Heidelberg, Germany Traffic-influenced

Zhu et al. 2002 a,b & 2004 6-220 Los Angeles, USA Traffic-influenced

Zhu et al. 2006 a 7-300 Los Angeles, USA Transition zone between continental boundary layer and free troposphere Van Dingenen et al. 2005 b 0.006-10

Monte Cimone Observatory, Italy

Urban Hussein et al. 2004 0.008-0.4 Kumpula and Siltavuori, Finland

Urban Hussein et al. 2005 0.003-0.6 Siltavuori and Pasila, Finland

Urban Monkkonen et al. 2005 0.003-0.8 New Delhi, India

Urban Wehner et al. 2002 0.003-0.8 Leipzig, Germany

Urban Wiedensohler et al. 2002 0.003-0.8 Leipzig, Germany

Urban Fine and Sioutas 2004 0.0141-2.5 LA Basin, USA

Urban Salma et al. 2002 0.01-10 Budapest, Hungary

Urban Morawska et al. 1999 0.016-30 Brisbane, Australia

Vegetation burning Morawska et al. 1999 0.016-30 Brisbane, Australia a Night-time data

163

Table 4.3 International literature reviewed to identify the location of the

modes in a number of different environments worldwide for particle

surface area distributions

Condition Researchers

Particle size range

measured (μm ) Location

Transition zone between continental Boundary layer and free troposphere Van Dingenen et al. 2005 a 0.006-10


Urban

Salma et al. 2002 0.01-10 Budapest, Hungary

Vegetation burning Jayaratne & Verma 2001 0.1-5 Gaborone, Botswana, Southern Africa

a Night-time data

Table 4.4 International literature reviewed to identify the location of the

modes in a number of different environments worldwide for particle

volume size distributions


Particle size range

measured (μm ) Location

Background Hidy 1975 0.015-30 Southern California, USA

Central European Aerosol Neususs et al. 2002 0.003-10 Leipzig and Berlin, Germany

Desert Hidy 1975 0.015-30 Southern California, USA Marine Hidy 1975 0.015-30 Southern California, USA

Marine and modified marine a Hoppel et al. 1990 0.006-2.2 Wallops Island, USA

Marine and modified marine a Gras and Ayers 1983 0.0025-5 Tasmania, Australia

Marine and modified marine a Porter and Clarke 1997 0.17-7.5 Tasmania, Australia

Marine and modified marine a Porter and Clarke 1997 0.17-7.5 Hawaii, USA

Modified marine a Morawska et al. 1999 0.016-30 Brisbane, Australia Marine and polluted air masses O'Dowd et al. 2001 0.005-150 Mace Head, Ireland

Modified background Morawska et al. 1999 0.016-30 Brisbane, Australia

Suburban Meszaros 1977 0.020-100 Budapest, Hungary

Suburban background Morawska et al. 1999 0.016-30 Brisbane, Australia

Traffic-influenced Hidy 1975 0.015-30 Southern California, USA

Traffic-influenced Morawska et al. 1998 0.016-30 Brisbane, Australia

Transition zone between continental boundary layer & free troposphere

Van Dingenen et al. 2005 b 0.006-10


Urban Morawska et al. 1999 0.016-30 Brisbane, Australia

Vegetation burning Jayaratne & Verma 2001 0.1-5

Gaborone, Botswana Southern Africa

Vegetation burning Morawska et al. 1999 0.016-30 Brisbane, Australia

a Modified marine in these cases refers to marine aerosol influenced by continental air parcels. b Night-time data

164

Table 4.5 International literature reviewed to identify the location of

the modes in a number of different environments worldwide for particle

mass size distributions


Particle size range measured

(μm ) Location

Himalayas Gajananda et al. 2005 0.08-9 North-west Himalayas, India

Marine Hillamo et al. 2001 0.045-10 High Arctic, remote boundary layer

Rural Berner et al. 2004

0.06-16 Vienna, Austria

Traffic Berner et al. 2004 0.06-16 Vienna, Austria

Urban Berner et al. 2004 0.06-16 Vienna, Austria



Three conclusions can be made from inspection of the results presented in

Figure 4.2. Firstly, it can be seen that there is a clear and distinct separation

between the modes at 1 µm for all worldwide environmental data reviewed for

surface area, volume and mass size distributions. The one exception is a volume

size distribution mode identified in marine and modified marine in Tasmania at

1 μm by Gras and Ayers (1983) where the salt component was found to comprise

more than 95% of the total volume. Secondly, it can be seen in Figure 4.2 that

clusters of modal values appear for each metric. Finally, the figure shows that

number and volume size distribution modal location values for South East

Queensland generally fell within the modal size ranges reported for the worldwide

environments.

165

1 10 100 1000 10000

Number

Surface Area

Volume

Mass

Figure 4.2. Published modal location values relating to particle size distributions for South East Queensland, Australia (marked x) and for a range of environments worldwide and metrics (n=600). Vertical dashed lines indicate the 95% confidence interval upper bounds for modal value clusters to the left of 1 μm and 95% confidence interval lower bounds for modal value clusters to the right of 1 µm in particle volume and mass size distributions, these modal value clusters are circled above

166

The effect of relative humidity on particle size under certain circumstances is

important and has been the topic of many investigations. For example,

Mobility Analysers can change relative humidity conditions during sampling,

and in many cases heat the sampled air sufficiently to reduce the size of the

particles. Mass size distributions measured at high relative humidity or in

clouds or fog show considerable fine particulate matter above 1 µm.

However, the papers reviewed by this study comprised a very wide range of

conditions, including studies related to high humidity conditions. Overall this

has not had an impact on the inferences. In fact the study region in South-East

Queensland experiences an annual average relative humidity of between 60-

73% in the mornings and 49-60% in the afternoons. Therefore it appears that

in the majority of cases (or under most circumstances) humidity is not a factor

changing the location of the mode according to the conclusions discussed

here. This is an important conclusion in relation to considerations in setting

standards, as these need to account for the majority of cases, especially in

relation to anthropogenic contributions.

4.3.3. Separation between modal location values in mass and volume

particle size distributions at around 1 µm

Of the 600 modal values examined in this study in particle number, surface

area, volume and mass size distributions 87 modes (15%) were clustered

closely on either side of 1 µm, and five modes were found at exactly 1 µm.

As clearly indicated in Figure 4.2, the modal values formed two distinct

subgroups above and below 1 μm. The upper 95% confidence bound of the

167

smaller mean, and the lower 95% confidence bound of the larger mean for

particle volume and mass size distributions, are displayed by vertical dashed

lines in Figure 4.2. The lack of overlap between these confidence intervals is

apparent, confirming the existence of two groups with statistically different

means. The confidence intervals were calculated for modal value clusters at

between 0.197 and 0.5 µm and 1.84 and 8 µm in volume; and 0.43 and 0.65

µm and 3.16 and 5.06 µm in mass size distributions. To facilitate comparison

between South East Queensland and modal values reported elsewhere in the

world, confidence intervals for South East Queensland modal values in

volume size distribution were not calculated.

When considering all the modal location values depicted in Figure 4.2, a

distinct gap was found between the location of the modes at both below and

above 1 µm. This distinct gap occurred at between 0.65 and 2 µm in mass

particle size distributions; between 0.3 and 2.2 µm in surface area; between

0.5 and 1.8 µm in volume and between 0.8 and 1.2 µm for number. It should

be noted that two distinct modal location values present at 1 µm and 2 µm in

Figure 4.2 related to marine environments. These were a mode found at 2 µm

in particle mass, which related primarily to sea salt particles in the remote

marine boundary layer in the high Arctic over the central Arctic Ocean

(Hillamo et al. 2001) and a mode at 1 µm in volume in an undisturbed marine

environment in the southern mid-latitudes, west coast of Tasmania, Australia,

where the salt component made up more than 95% of the total volume (Gras

and Ayers 1983).

168

4.4. CONCLUSIONS

The relation between fractional contribution to volume and mass from

different modes in the particle size distribution (and thus from different

sources) to PM1, PM2.5 and PM10 was examined in this paper, based on a large

body of data on ambient particle size distributions from the measurements

conducted in South East Queensland, Australia. The conclusions from the

analyses in relation to developing air quality regulations are as follows.

Firstly, PM10 measurements provide information almost entirely on particles

generated from mechanical processes and belonging to the coarse mode. In an

urban environment this could also mean particles resuspended by the vehicular

traffic and mechanical wear and tear of the tyres, but not emitted from motor

vehicles.

Secondly, PM2.5 measurements (coarse mode) also provide information mainly

on particles generated by mechanical processes, but the contribution from

combustion process modes (nucleation and accumulation modes) becomes

significant for some environments. Thus interpretation of PM2.5 data could

become very complex in order to distinguish the contribution from different

types of sources. It follows that the application of this PM2.5 parameter, as a

basis for standards may not adequately facilitate control of particle emissions

and concentrations.

169

Thirdly, PM1 measurements (nucleation and accumulation modes) provide

very good information about contributions from combustion processes and

enable a much better distinction to be made between combustion and

mechanically generated aerosols. It would thus appear that PM1 and PM10

mass standards would be most desirable from the legislation point of view.

The review of 600 modal location values for particle number, surface area,

volume and mass size distributions in a wide range of environments

worldwide revealed a clear and distinct separation around 1 µm. A similar

separation was found in all the South-East Queensland environments

examined in terms of the separation between accumulation and coarse modes

for volume and number size distributions, which occurred at around 1 µm.

We conclude that examination of the location of the modes in particle size

distributions has potential as a basis for developing air quality standards and

guidelines as modes provide useful information about contributions from

different pollution sources and particle mechanisms. Therefore, based on both

the local South East Queensland study and the other studies conducted around

the world, it is concluded that PM1 and PM10 offer greater potential as a

combination for particle mass standards than the current mass standards of

PM2.5 and PM10.

Two additional points need to be discussed when considering a PM1 standard.

Firstly, while at the moment very little data are available on PM1

concentrations, there are measurement technologies available to undertake

170

these measurements, which are very similar to those used for PM2.5

monitoring. Secondly, in addition to particle mass concentration standards,

future legislations may also consider number concentration standards, which

would be focused on submicrometer or even smaller, ultrafine particles. In

urban areas, for example, motor vehicles are the major emitter of ultrafine

particles, which are very small and prolific in terms of particle number, but

have negligible mass. The rapid progress in the monitoring technologies

available to measure particle number concentration currently makes such

measurements possible. While this paper considered only the rationale for the

most advantageous combination of particle mass standards from the legislative

point of view, more discussion should be conducted to consider the best

combination of particle mass and number concentration standards.

171

4.5. REFERENCES

Baron, P.A., Willeke, K., (Eds.) 2001. Aerosol Measurement, Principles,

Techniques and Applications, 2nd edn, John Wiley & Sons, Inc., New York.

Berner, A., Galambos, Z., Ctyroky, P., Fruhauf, P., Hitzenberger, R.,

Gomiscek, B., Hauck, H., Preining, O., Puxbaum, H., 2004. On the correlation

of atmospheric aerosol components of mass size distributions in the larger

region of a central European city. Atmospheric Environment 38 (24), 3959-

3970.

Birmili, W., Wiedensohler, A., Heintzenberg, J., Lehmann K., 2001.

Atmospheric particle number size distribution in central Europe: Statistical

relations to air masses and meteorology. Journal of Geophysical Research-

Atmospheres 106 (D23), 32005-32018.

Birmili, W., Heintzenberg, J., Wiedensohler, A., 1999. Representative

measurement and parameterization of the submicron continental particle size

distribution. Journal of Aerosol Science 30, Suppl. 1, S229-S230.

Dockery, D.W., Pope, C.A., Xu, X., Spengler, J.D., Ware, J.H, Fay., M.E,

Ferris., B.G., Speizer, F.E., 1993. An Association between Air Pollution and

Mortality in Six U.S. Cities. The New England Journal of Medicine, 329 (24),

1753-9.

172

Fine, PM., Shen, S., Sioutas, C., 2004. Inferring the sources of fine and

ultrafine particulate matter at downwind receptor sites in the Los Angeles

basin using multiple continuous measurements. Aerosol Science and

Technology 38, 182-195 Suppl. 1.

Gajananda, K., Kuniyal, J.C., Momin, G.A., Rao, P.S.P., Safai, P.D., Tiwari,

S., Ali, K., 2005. Trend of atmospheric aerosols over the north western

Himalayan region, India. Atmospheric Environment 39 (27), 4817-4825.

Gras, J.L., Ayers, G.P., 1983. Marine aerosol at southern mid-latitudes.

Journal of Geophysical Research 88(C15), 10661-10666.

Harrison, R.M., Shi, J.P., Zi, S., Khan. A., Mark, D., Kinnersley, R., Yin, J.,

2000. Measurement of number, mass and size distribution of particles in the

atmosphere, Philosophical Transactions of the Royal Society A:

Mathematical, Physical and Engineering Sciences, 358 (1775), 2567-2580.

Heintzenberg, J., Birmili, W., Wiedensohler, A., Nowak, A., Tuch, T., 2004.

Structure, variability and persistence of the submicrometre marine aerosol,

Tellus Series B-Chemical and Physical Meteorology 56 (4), 357-367.

Hidy, G.M., 1975. Summary of the California Aerosol Characterization

Experiment. Journal of the Air Pollution Control Association 25, 1106-1114.

173

Hillamo, R., Kerminen, V.M., Aurela, M., Makela, T., Maenhaut, W., Leck,

C., 2001. Modal structure of chemical mass size distribution in the high

Arctic aerosol, Journal of Geophysical Research-Atmospheres 106 (D21),

27555-27571.

Hoppel, W.A., Larson, R., Vietti, M.A., 1990. Aerosol size distributions and

optical boundaries found in the marine boundary layer over the Atlantic

Ocean. Journal of Geophysical Research 95(D4), 3659-3686.

Hussein, T., Hameri, K., Heikkinen, M.S.A., Kulmala, M. 2005. Indoor and

outdoor particle size characterization at a family house in Espoo-Finland,


Hussein, T., Puustinen A., Aalto P.P., Makela, J.M., Hameri K., Kulmala, M.

2004. Urban aerosol number size distributions. Atmospheric Chemistry and

Physics 4, 391-411.

Jamriska, M., Morawska, L., 2000. The effect of surface deposition,

coagulation and ventilation on submicrometer particles indoors. Clean Air

and Environment Conference, Sydney, Australia, 26-30 November 2000.

Jaenicke, R., 1993. Tropospheric aerosols. In Hobbs, P.V. (ed) Aerosol-

Cloud-Climate Change Interactions, Academic Press, San Diego, USA, 1-31.

174

Jayaratne, E.R., Verma, T.S., 2001. The impact of biomass burning on the

environmental aerosol concentration in Gaborone, Botswana, Atmospheric


John, W., 1993. The characteristics of environmental and laboratory

generated-aerosols, in: Willeke and Baron (Eds.), Aerosol measurement:

Principles, techniques and applications,Van Nostrand Reinhold, New York,

55.

Lundgren, D.A., Burton, R.M., 1995. Effect of particle size distribution on

the cut point between fine and coarse ambient mass fractions, Inhalation

Toxicology 7 (1), 131-148.

Makela, J.M., Koponen, I.K., Aalto, P., Kulmala., M., 2000. One-year data of

submicron size modes of tropospheric background aerosol in Southern

Finland. Journal of Aerosol Science 31 (5), 595-611.

Meszaros, A., 1977. On the size distribution of atmospheric aerosol particles

of different composition. Atmospheric Environment 11, 1075-1081.

Monkkonen, P., Koponen, I.K., Lehtinen, K.E.J., Hameri, K., Uma, R.,

Kulmala, M., 2005. Measurements in a highly polluted Asian mega city:

observations of aerosol number size distribution, modal parameters and

nucleation events. Atmospheric Chemistry and Physics 5, 57-66.

175

Morawska, L., 2004. Indoor particles, combustion products and fibres. The

Handbook of Environmental Chemistry , Springer-Verlag Heidelberg, 4F,

117-147.

Morawska, L., Thomas, S., Jamriska, M., 1999. The modality of particle size

distributions of environmental aerosols, Atmospheric Environment 33: 4401-

4411.

Morawska, L., Thomas, S., Bofinger, N.D., Wainwright, D., Neale D., 1998.

Comprehensive characterisation of aerosols in a subtropical urban

atmosphere: particle size distribution and correlation with gaseous pollutants.

Atmospheric Environment 32, 2467–2478.

Nazaroff, W., Ligocki, M., Ma, T., Cass, G., 1990. Particle Deposition in

Museums, Comparison of Modelling and Measurement Results, Aerosol

Science and Technology 13, 332-348.

Neususs, C., Wex. H., Birmili, W., Wiedensohler, A., Koziar, C., Busch, B.,

Bruggemann, E., Gnauk, T., Ebert, M., Covert, D.S. 2002. Characterization

and parameterization of atmospheric particle number-, mass-, and chemical-

size distributions in central Europe during LACE 98 and MINT - art. no.

8127. Journal of Geophysical Research-Atmospheres 107 (D21), 8127-8127.

O'Dowd, C.D., Becker, E., Kulmala, M., 2001. Mid-latitude North-Atlantic

aerosol characteristics in clean and polluted air. Atmospheric Research 58

(3), 167-185.

176

Pirjola, L., Parviainen, H., Hussein, T., Valli, A., Hameri, K., Aaalto, P.,

Virtanen, A., Keskinen, J., Pakkanen, T.A., Makela, T., Hillamo, R.E., 2004.

"Sniffer" - a novel tool for chasing vehicles and measuring traffic pollutants,

Atmospheric Environment 38 (22), 3625-3635.

Porter, J.N., Clarke A.D., 1997. Aerosol size distribution models based on in

situ measurements. Journal of Geophysical Research 102(D5), 6035-6045.

Rosenbohm, E., Vogt, R., Scheer, V., Nielsen, O.J., Dreiseidler, A.,

Baumbach, G., Imhof, D., Baltensperger, U., Fuchs, J., Jaeschke, W., 2005.

Particulate size distributions and mass measured at a motorway during the

BAB II campaign. Atmospheric Environment 39 (31), 5696-5709.

Salma, I., Ocskay, R., Raes, N., Maenhaut, W., 2005. Fine structure of mass

size distributions in an urban environment. Atmospheric Environment 39 (29),

5363-5374.

Salma, I., Dal Maso, M., Kulmala, M., Zaray, G., 2002. Modal characteristics

of particulate matter in urban atmospheric aerosols. Microchemical Journal 73

(1-2), 19-26.

Tunved, P., Nilsson, E.D., Hansson, H.C., Strom, J., 2005. Aerosol

characteristics of air masses in northern Europe: Influences of location,

transport, sinks, and sources - art. no. D07201. Journal of Geophysical

Research-Atmospheres 110 (D7), 7201-7201.

177

USEPA., 2004. Air Quality Criteria for Particulate Matter 2004. U.S.

Environmental Protection Agency, Washington, DC, EPA 600/P-99/002aF-

bF, 2004.

Van Dingenen, R., Putaud, J.P., Martins-Dos Santos, S., Raes, F., 2005.

Physical aerosol properties and their relation to air mass origin at Monte

Cimone (Italy) during the first MINATROC campaign. Atmospheric

Chemistry and Physics 5, 2203-2226.

Wehner, B., Birmili, W., Gnauk T., Wiedensohler, A., 2002. Particle number

size distributions in a street canyon and their transformation into the urban-air

background: measurements and a simple model study. Atmospheric

Environment 36 (13), 2215-2223.

Weingartner, E., Nyeki, S., Baltensperger, U., 1999. Seasonal and diurnal

variation of aerosol size distributions (10 < D < 750 nm) at a high-alpine site

(Jungfraujoch 3580 m asl). Journal of Geophysical Research-Atmospheres

104 (D21), 26809-26820.

Wiedensohler, A., Wehner, B., Birmili, W., 2002. Aerosol number

concentrations and size distributions at mountain-rural, urban-influenced

rural, and urban-background sites in Germany. Journal of Aerosol Medicine-

Deposition Clearance and Effects in the Lung 15 (2), 237-243.

178

Zhu, Y., Hinds, W.C., Kim, S., Sioutas C., 2002a. Concentration and size


Waste Management Association 52 (9), 1032-1042.

Zhu, Y., Hinds, W.C., Kim, S., Shen, S, Sioutas, C., 2002b. Study of ultrafine


Environment 36 (27), 4323-4335.

Zhu, Y., Hinds, W.C., Shen, S., Sioutas, C., 2004, Seasonal trends of

concentration and size distribution of ultrafine particles near major highways

in Los Angeles, Aerosol Science & Technology 38 (S1), 5-13.

Zhu, Y., Kuhn, T., Mayo, P., Hinds, W.C., 2006. Comparison of daytime and

nighttime concentration profiles and size distributions of ultrafine particles

near a major highway. Environmental Science & Technology 40, 2531-2536.

179

CHAPTER 5

DERIVATION OF MOTOR VEHICLE TAILPIPE PARTICLE EMISSION FACTORS

180

5. OVERVIEW OF CHAPTERS 5.1. AND 5.2.

Current knowledge concerning which are the most suitable emission factors to use in

transport modelling is patchy and ill-defined.

This Chapter presents a rigourous method developed to derive a comprehensive set of

particle emission factors that can be used in transport modelling and health impact

assessments to quantify inventories for motor vehicle fleets, and also presents the

outputs from the statistical models developed to derive these average emission

factors.

181

CHAPTER 5.1.

This Chapter presents the second paper of the PhD project. This paper presents a

comprehensive set of particle emission factors for urban motor vehicle fleets, which

can be used in transport modelling and health impact assessments to derive size-

resolved inventories of tailpipe particle emissions covering the full size range of

particles emitted from vehicles, and includes emission factors for particle number,

particle volume, PM1, PM2.5 and PM10 for different Vehicle and Road Type

combinations.

The paper discusses the method used to derive average emission factors and the

rationale for selection of the most suitable emission factors. The approach included

development of five statistical models that produced average emission factors, based

on a statistical analysis of 667 particle emission factors in the international published

literature. The paper also identified a number of gaps in our current knowledge about

motor vehicle emission factors related to exhaust and non-exhaust emissions.

The emission factors considered the most suitable to use in transport modelling

presented in this paper are suitable for deriving inventories for urban fleets in other

developed countries, and have particular application for areas which may have no, or

insufficient measurement data, upon which to derive emission factors.

182

CHAPTER 5.2.

Chapter 5.2 presents the outputs of the five statistical models developed to produce

average emission factors; and results of statistical tests that examined differences in

average emission factors related to categorical variables examined in the statistical

models. These two topics are also discussed in the paper presented in Chapter 5.1.

The statistical model outputs presented in this Chapter include the explanatory model

variables and their average emission factors, associated 95% confidence intervals and

standards errors, for particle number, particle volume, PM1, PM2.5 and PM10. From

these statistical model outputs, the most suitable emission factors for different Vehicle

and Road Type combinations for different particle metrics were selected. These are

shown in Tables 5.2.1-5.2.5 in bold italics shaded gray in Chapter 5.2, and are also

summarised in Table 5.1.4 in Chapter 5.1.

Chapter 5.2 also presents a multiple comparison plot, Figure 5.2.1, which depicts the

statistical relationships between average values of published emission factors for

categorical variables examined in the statistical analysis. The results depicted in this

plot are also referred to in the paper presented in Chapter 5.1.

183

CHAPTER 5.1

DERIVATION OF MOTOR VEHICLE TAILPIPE PARTICLE

EMISSION FACTORS SUITABLE FOR MODELLING URBAN

FLEET EMISSIONS AND AIR QUALITY ASSESSMENTS

Diane U. Keogh1, Joe Kelly2, Kerrie Mengersen2, Rohan Jayaratne1, Luis Ferreira3,

Lidia Morawska1

1 International Laboratory for Air Quality and Health, Queensland University of

Technology, Gardens Point, Brisbane, Australia

2 School of Mathematical Sciences, Queensland University of Technology,

Gardens Point, Brisbane, Australia

3 School of Urban Development, Queensland University of Technology,

Gardens Point, Brisbane, Australia

Environmental Science and Pollution Research – International. Published online, doi

0.1007/s11356-009-0210-9.

184


Title: Derivation of motor vehicle particle tailpipe particle emission

factors suitable for modelling urban fleet emissions and air

quality assessments

Authors: Diane U. Keogh, Joe Kelly, Kerrie Mengersen, Rohan Jayaratne,

Luis Ferreira, and Lidia Morawska


Developed the experimental design and scientific method of the study. Data

collection, interpretation and processing. Analysis and interpretation of statistical

model outputs. Wrote the majority of the manuscript.

Joe Kelly

Developed the statistical models and conducted the statistical tests.

Kerrie Mengersen

Contributed to the experimental design and scientific method for the

statistical models and statistical tests. Contributed to the manuscript.

Rohan Jayaratne

Assisted with the manuscript

Luis Ferreira

Reviewed the manuscript.

Lidia Morawska

Contributed to the design in relation to classification of emission factors

relating to measurement methodology used.

185

ABSTRACT

Background, aim, and scope Urban motor vehicle fleets are a major source of

particulate matter pollution, especially of ultrafine particles (diameters < 0.1 µm), and

exposure to particulate matter has known serious health effects. A considerable body

of literature is available on vehicle particle emission factors derived using a wide

range of different measurement methods for different particle sizes, conducted in

different parts of the world. Therefore the choice as to which are the most suitable

particle emission factors to use in transport modelling and health impact assessments

presented as a very difficult task. The aim of this study was to derive a

comprehensive set of tailpipe particle emission factors for different vehicle and road

type combinations, covering the full size range of particles emitted, which are suitable

for modelling urban fleet emissions.

Materials and methods A large body of data available in the international literature

on particle emission factors for motor vehicles derived from measurement studies was

compiled and subjected to advanced statistical analysis, to determine the most

suitable emission factors to use in modelling urban fleet emissions.

Results This analysis resulted in the development of five statistical models which

explained 86%, 93%, 87%, 65% and 47% of the variation in published emission

factors for particle number, particle volume, PM1, PM2.5 and PM10 respectively. A

sixth model for total particle mass was proposed but no significant explanatory

variables were identified in the analysis. From the outputs of these statistical models,

the most suitable particle emission factors were selected. This selection was based on

186

examination of the statistical robustness of the statistical model outputs, including

consideration of conservative average particle emission factors with the lowest

standard errors, narrowest 95% confidence intervals and largest sample sizes, and the

explanatory model variables, which were Vehicle Type (all particle metrics),

Instrumentation (particle number and PM2.5), Road Type (PM10) and Size Range

Measured and Speed Limit on the Road (particle volume).

Discussion A multiplicity of factors need to be considered in determining emission

factors that are suitable for modelling motor vehicle emissions, and this study derived

a set of average emission factors suitable for quantifying motor vehicle tailpipe

particle emissions in developed countries.

Conclusions The comprehensive set of tailpipe particle emission factors presented in

this study for different vehicle and road type combinations enable the full size range

of particles generated by fleets to be quantified, including ultrafine particles

(measured in terms of particle number). These emission factors have particular

application for regions which may have a lack of funding to undertake measurements,

or insufficient measurement data upon which to derive emission factors for their

region.

Recommendations and perspectives In urban areas motor vehicles continue to be a

major source of particulate matter pollution and of ultrafine particles. It is critical

that in order to manage this major pollution source methods are available to quantify

187

the full size range of particles emitted for transport modelling and health impact

assessments.

Keywords: ANOVA; emission factors; linear regression; motor vehicles;

multiple comparison; particle mass; particle number; Scheffe; ultrafine

particles.

5.1. BACKGROUND, AIM AND SCOPE

In urban areas motor vehicle fleets are the main source of particulate matter pollution,

and these particles span a very broad size range (diameters 0.003–10 µm); however

most are ultrafine size and measured in terms of particle number (number

concentration of particles with diameters < 0.1 µm) (Harrison et al. 1999; Shi and

Harrison 1999; Shi et al. 1999; Shi et al. 2001; Morawska 2003; Wahlin et al. 2001).

For this reason, it is critical that particle number emissions be included in

development of motor vehicle inventories and health impact assessments.

Emission factors are used in combination with transport data to develop inventories,

and a very large body of data on emission factors derived from measurements is

available in the international literature. These relate to measurement studies of

vehicles under different driving conditions conducted on dynamometers in

laboratories, on or near roads, and in tunnels. A wide range of different measurement

methods have been used for different particle sizes, conducted in different parts of the

world, and a multiplicity of issues need to be considered and resolved in order to

188

derive emission factors. Factors can include vehicle type, fuel type, vehicle age,

technologies fitted, speed and load, road environment characteristics, driving cycles,

driving patterns, method and instrumentation used and size range measured, to name

a few. This extensive body of data on particle emission factors has never been

comprehensively analysed, and the question that remains is - Which tailpipe particle

emission factors are the most suitable to use in transport modelling and health

impact assessments of motor vehicle fleets?

Many mobile emission source models are available in developed countries which

utilise performance-based emission factors (related to emissions generated per

vehicle per kilometre derived from measurement data), for example, the average

speed models MOBILE (USEPA 1993), EMFAC (CARB 2002), COPERT (Ahlvik et

al. 1997; Ntziachristos et al. 2000; Bellasio et al. 2007); and VERSIT+ LD (Smit et

al. 2007) which considers actual driving pattern data. Most of these models provide

estimates for PM10, and to a lesser extent PM2.5. COPERT IV, however, has available

a small suite of solid particle number emission factors for different vehicle types

derived from dynamometer measurements (Samaras et al. 2005).

In developing countries access to land use and transport network data is often rare

(Walker et al. 2008) and hence more indirect methods for estimating emissions are

commonly used, such as basing emission factors on estimated total fuel consumed or

on remotely sensed data. Emission estimates based on remotely sensed data usually

provide a snapshot of emissions relating to a limited number of locations, and may

189

not be representative of activity patterns for a typical trip in a region (Frey et al.

2002b); and the accuracy of fuel-based models can depend on how well the driving

modes, vehicle and age distribution from which the emission factors were derived

represent the study region (Frey et al. 2002 a,b).

The aim of this work was to identify the most suitable tailpipe particle emission

factors to use in transport modelling and health impact assessments to quantify motor

vehicle fleet particle emissions in terms of particle number, particle volume, PM1,

PM2.5 and PM10 emissions, based on analysis of emission factors derived from

measurement data. Emission factors for brake and tyre wear, road dust and particle

surface area emissions were not considered in this analysis as only limited data exists

in the literature.

5.2. MATERIALS AND METHODS

An extensive review was conducted of emission factors published in the international

literature for particle number, particle volume, total particle mass, PM1, PM2.5 and

PM10 for motor vehicle tailpipe emissions. Details of the literature reviewed and

studies from which emission factor data was sourced for this study are outlined in

Table 5.1.1. Based on this review, statistical models were developed and emission

factor data classified and grouped into relevant sub-classes within each model

variable class. Statistical model output data were analysed and a rationale developed

to identify the most suitable average emission factors to use in modelling urban motor

vehicle emissions.

190

Table 5.1.1 Source of tailpipe particle emission factors examined in the statistical

analysis to derive average emission factors for different vehicle and road type

combinations.

Particle metric

Researchers

Country of Study

Study Location

Size Range Measured

(nm) af

Instrumentation bd

Vehicle Type Emission Factors e

Particle number

(Cadle et al, 2001)

USA

Dynamometer

> 3

ELPI, UCPC

LDV

(CONCAWE, 1998) Belgium Dynamometer 10-237.2 DMA LDV

15.7-685.4 SMPS, DMPS LDV 10-1000 EAA LDV

(Morawska et al, 2001) Australia Dynamometer 15-700 SMPS HDV

(Ristovski et al, 2002) Australia Dynamometer 8-400 SMPS Bus (Diesel)

(Abu-Allaban, 2002) USA Tunnel 10-400 SMPS Fleet

(Gertler et al, 2002) USA Tunnel 10-500 SMPS Fleet

(Gidhagen et al, 2003) Sweden Tunnel

< 10, 10 -29, 29-109, 109-900, 3-900 DMPS HDV, LDV

(Imhof et al, 2005b)

Austria & UK Tunnel

18-50, 18-100,

18-300, 18-700 SMPS

Fleet, HDV, LDV

(Jamriska et al, 2004) Australia Tunnel 17-890 SMPS Bus (Diesel)

(Kristensson et al, 2004) Sweden Tunnel 3-900 DMPS Fleet

(Corsmeier et al, 2005) Germany

Vicinity of the road 30-10,000 ELPI

Fleet, HDV, LDV

3-900, 10-

400 SMPS Fleet

(Gidhagen et al, 2004a) Sweden

Vicinity of the road 7-450 CPC, DMPS Fleet

(Gidhagen et al, 2004b) Sweden

Vicinity of the road > 3 CPC, DMPS HDV, LDV

(Gramotnev et al, 2003) Australia

Vicinity of the road 15-700 SMPS Fleet

(Gramotnev et al, 2004) Australia


(Hueglin et al, 2006) Switzerland

Vicinity of the road 7-3000 CPC Fleet

(Imhof et al, 2005c) Germany

Vicinity of the road 30-10,000 ELPI

Fleet, HDV, LDV

(Imhof et al, 2005a) Switzerland

Vicinity of the road > 7 CPC

Fleet, HDV, LDV

18-50, 18-100,

18-300 SMPS Fleet, HDV, LDV

191

Particle metric

Researchers

Country of Study

Study Location

Size Range Measured

(nm) af

Instrumentation bd


Particle Number (c’td)

(Jamriska and Morawska, 2001) Australia


(Jones and Harrison, 2006) UK

Vicinity of the road

11-30, 30-100,

11-450, 101-450 SMPS HDV, LDV

(Ketzel et al, 2003) Denmark

Vicinity of the road 10-700 CPC, DMPS Fleet

(Kittelson et al, 2004) USA


3-1000 CPC Fleet

(Morawska et al, 2005) Australia

Vicinity of the road 17-890 SMPS

Fleet, HDV, LDV

700-20,000 APS Fleet, HDV, LDV

(Zhu and Hinds, 2005) USA

Vicinity of the road > 6 CPC Fleet

Particle volume

(Imhof et al, 2005b)

Austria & UK Tunnel

18-50, 18-100,

18-300, 18-700 SMPS

Fleet, HDV, LDV

(Corsmeier et al, 2005) Germany

Vicinity of the road 30-10,000 ELPI HDV, LDV

(Imhof et al, 2005c) Germany

Vicinity of the road 29-250 ELPI

Fleet, HDV, LDV

29-640, 29-

1000 ELPI Fleet


Vicinity of the road

18-50, 18-100,

18-300 SMPS Fleet, HDV, LDV

Total Particle mass

(Ayala et al, 2002) USA Dynamometer

MOUDI, ELPI, SMPS

Bus (Diesel & CNG)

(Chatterjee et al, 2002) USA Dynamometer not reported Bus (Diesel)

(Clark et al, 1997 & 1998) USA Dynamometer not reported

Bus (Diesel & CNG)

(Clark et al, 1999) USA Dynamometer not reported

Bus (Diesel & CNG)

(CONCAWE, 1998) Belgium Dynamometer 17.9-16,000

Berner impactor & filter paper LDV

(Kado et al, 2005) USA Dynamometer not reported

Bus Diesel & CNG)

(Lanni et al, 2003) Canada Dynamometer Pallflex filters

Bus Diesel & CNG)

(Lowell et al, 2003)

USA & Canada Dynamometer not reported

Bus( Diesel, CNG, LNG)

192

Particle metric

Researchers

Country of Study

Study Location

Size

Range Measured

(nm) af

Instrumentation bd


Total Particle mass (c’td)

(Morawska et al, 1998) Australia Dynamometer 8-300 SMPS Bus Diesel

(Bradley, 2000) USA Dynamometer Fibreglass filters Bus (Diesel, CNG, Hybrid)

(SAE, 2001; 2002a,b; 2003a, b; cited in Lowell et al, 2003)

USA & Canada Dynamometer not reported

Bus (Diesel, CNG)

(CARB, 2001; ARB's, 2002 cited in Lowell et al, 2003) USA Dynamometer not reported

Bus (Diesel & CNG)

(Ubanwa et al, 2003) USA Dynamometer not reported

HDV, Bus (Diesel)

(Wayne et al, 2004) USA Dynamometer not reported

Bus (Diesel, LNG, Hybrid)

(Jamriska et al, 2004) Australia Tunnel 17-700 SMPS Bus (Diesel)

(Holmen et al, 2005) USA

Vicinity of the road Telfon filters

Bus (Diesel), Hybrid Bus

(Kittelson et al, 2004) USA


(Mazzoleni et al, 2004) USA

Vicinity of the road Remote sensing Fleet

(Shah et al, 2004) USA

Vicinity of the road Teflon filters Fleet, HDV

(Zhang et al, 2005) USA

Vicinity of the road 6-220 inverse modelling

HDV, LDV

> 220c inverse modelling HDV, LDV

(Abu-Allaban et al, 2003b) USA

Vicinity of the road Chemical balance HDV, LDV

193

Particle metric

Researchers

Country of Study

Study

Size Range Measured

(nm) af

Instrumentation bd


PM1 (DOEH, 2003) Australia Dynamometer sm APS HDV, LDV

(Imhof et al, 2005b) Austria & UK Tunnel sm Kleinfiltergerate

Fleet, HDV, LDV

(Gehrig et al, 2004) Switzerland

Vicinity of the road sm Beta-ray

Fleet, HDV, LDV


Vicinity of the road sm Betameter

Fleet, HDV, LDV

PM2.5 (NEPC, 2000) Australia Dynamometer sm APS HDV, LDV

(Wayne et al, 2004) USA Dynamometer sm Glass-fibre filter

Bus (Diesel & LNG), Hybrid Bus

(Gertler et al, 2002) USA Tunnel sm

IMPROVE sampler

Fleet, HDV, LDV

(Gillies et al, 2001) USA Tunnel sm

Medium-volume samplers Fleet

(Imhof et al, 2005b) UK Tunnel sm TEOM

Fleet, HDV, LDV

(Jamriska et al, 2004) Australia Tunnel sm TEOM, DustTrak Bus (Diesel)

(Kristensson et al, 2004) Sweden Tunnel sm TEOM & DMPS Fleet

(Tran et al, 2003) Australia Tunnel sm Teflon filters HDV, LDV

(Abu-Allaban et al, 2003a) USA

Vicinity of the road sm DustTrak

HDV, LDV, Bus

(Morawska et al, 2004) Australia

Vicinity of the road sm DustTrak Fleet

(Abu-Allaban et al, 2003b) USA

Vicinity of the road sm Chemical balance HDV, LDV

PM10

(Cadle et al, 1997)

USA

Dynamometer

sm

Teflon & Quartz filters LDV

(Cadle et al, 2001) USA Dynamometer sm MOUDI LDV

(Lowell et al, 2003)

USA & Canada

Dynamometer

sm

not reported

Bus (Diesel)

(NEPC, 2000) Australia Dynamometer sm APS HDV, LDV

(Romilly, 1999) UK Dynamometer sm not reported

LDV, Bus, Midibus, Minibus

(SAE, 2001; SAE, 2002a) cited in Lowell et al, 2003)

USA & Canada Dynamometer sm not reported

Bus (Diesel & CNG)

(Wayne et al, 2004) USA Dynamometer sm not reported

Bus (Diesel & LNG), Hybrid Bus

(Gertler et al, 2002) USA Tunnel sm DustTrak

Fleet, HDV, LDV

194

Particle metric

Researchers

Country of Study

Study

Size Range Measured

(nm) af

Instrumentation bd


PM10 (cont’d)

(Gillies et al, 2001) USA Tunnel sm

Medium-volume samplers Fleet

(Hibberd, 2005) Australia Tunnel sm

Statistical analysis c Fleet, HDV, LDV

(Imhof et al, 2005b) Austria Tunnel sm TEOM Fleet, HDV

(Kristensson et al, 2004) Sweden Tunnel sm TEOM & DMPS Fleet

(Schmid et al, 2001) Austria Tunnel sm Quartz filters Fleet, HDV, LDV

(Tran et al, 2003) Australia Tunnel sm Teflon filters LDV

(Abu-Allaban et al, 2003a) USA

Vicinity of the road sm DustTrak HDV, LDV, Bus

(Venkatram et al, 1999) USA

Vicinity of the road sm Teflon filters Fleet

(Gehrig et al, 2004)

Switzerland

Vicinity of the road sm Beta-ray Fleet, HDV, LDV

(Imhof et al, 2005a)

Switzerland

Vicinity of the road sm Betameter Fleet, HDV, LDV

a 1000 nm is equivalent to 1 µm. These units refer to particle diameter. b Instrumentation (in alphabetical

order) - Aerodynamic Particle Sizer (APS), Berner low pressure Impactor, Beta-ray absorption monitors,

Betameter, Chemical Mass Balance, Condensation Particle Counter (CPC), Differential Mobility Analyzer

(DMA), Differential Mobility Particle Sizer (DMPS), Dynamometer, DustTrak, Electrical Aerosol Analyser

(EAA), Electrical Low Pressure Impactor (ELPI), Filters (Fibreglass, Glass fibre, Teflon, Quartz),

Kleinfiltergerate, LIDAR-based VERSS and remote sensing, Mass Single Stage Multidilutor, MOUDI

(Micro-Orifice Uniform Deposit Impactor), Samplers (IMPROVE, high volume, medium volume), Scanning

Mobility Particle Sizer (SMPS), Tapered Element Oscillating Microbalances(TEOM) and Ultrafine

Condensation Particle Counter (UCPC). c Fit log-normal functions to extrapolate concentrations beyond >

220nm. Statistical analysis of in-stack pollution monitoring data and hourly vehicle counts. d Not reported –

dynamometer studies which did not provide further information on Instrumentation used. e Vicinity of the

road studies refer to studies conducted on or near the road, near a kerb, upwind or downwind of the road. f

sm – refers to Size Range Measured and relates to particles with diameters < 1 µm, < 2.5 µm and < 10 µm

(known as PM1, PM2.5 and PM10 respectively). g LDV (Light duty vehicles), HDV (Heavy duty vehicles) –

refer Table 5.1.2. for further detail.

195

5.2.1. Model variables examined

From an original population of over 900 emission factors reviewed in this study, the

final emission factor sample size obtained was 667. This occurred due to the high

number of missing data in the studies, as not all studies reported the same

information. The model variables developed for the statistical analysis were based on

data commonly reported in studies.

Data relating to a total of 667 particle emission factors were examined grouped into

relevant sub-classes within each model variable class. The categorical model

variables developed were Particle Metric, Country of Study, Study Location,

Instrumentation, Vehicle Type, Fuel Type, Road Type, Road Class; and the

continuous model variables were Size Range Measured, Average Vehicle Speed,

Speed Limit on the Road, Average Number of Vehicles travelling in a fleet per day,

Drive Cycles, Engine Power, Heavy Duty Vehicle (HDV) Share, Number of HDVs

travelling in a fleet per day. These model variables are described in Table 5.1.2, and

the sample size of emission factors relating to these model variables are shown in

Table 5.1.3.

196

Table 5.1.2 Model variables examined in the statistical analysis to derive average

emission factors to use in transport modelling and health impact assessments, to

quantify tailpipe particle emissions generated by motor vehicle fleets

Model Variable Name

Model Variable Sub-classes

Particle Metric Particle number, particle volume, total particle mass, PM1, PM2.5, PM10

Country of Study Australia; USA/Canada; Other Countries (Austria, Belgium, Denmark, Germany, Sweden, Switzerland, UK) a

Study Location Dynamometer (in a laboratory), tunnel or in the vicinity of a road b Road Type Boulevard, freeway, highway, motorway, rural area, tunnel, urban c Speed Limit on the Road

The reported Speed Limit on the Road d

Road Class Urban and Non-Urban roads; Highway and Non-Highways roads e Average Number of Vehicles Per Day

The average number of vehicles travelling in a vehicle fleet per day f

Heavy Duty Vehicle Share

Percentage of heavy duty vehicles (HDVs) travelling in a vehicle fleet per day g

Number of HDVs Per Day

Number of HDVs travelling in a vehicle fleet per day h

Vehicle Type Fleet, light duty vehicles (LDVs), heavy duty vehicles (HDVs) Bus i Fuel Types Diesel, Gasoline, Compressed Natural Gas, Liquefied Natural Gas j Drive Cycles Drive Cycles for Buses, Trucks and Other vehicles k Average Vehicle Speed

Average Vehicle Speed tested on a dynamometer or reported in a tunnel or vicinity of the road study l

Engine Power Reported for two bus studies m Instrumentation 20 different types of Instrumentation n Size Range Measured

Size Range Measured by Instrumentation o

a Groups based on numbers of studies found. b Vicinity of the road - on or near the road, near a curb,

upwind, downwind of a road. c Urban Drive Cycle data classed as urban Road Type. d Few studies

reported, where reported was Boulevard 82, highway 82 and 100, freeway 100, motorway 120, tunnel

60, 64, 80, 89, urban 50 and 57 km/hr. e Road Class based on either the reported Speed Limit on the

Road, or the speed limit that would most likely be associated with the Road Type. < 60 road classed

Urban; ≥ 60 non-Urban; ≥ 80 Highway; < 80 km/hr non-Highway. Insufficient data were available to

examine individual speeds or other specific speed ranges. f Ranges 13,128-103,080 per day particle

number; 23,000-30,000 particle volume; 12,540-12,900 total particle mass; 20,000-69,816 PM1;

20,000-69,816 per day PM10. 5 buses/minute particle number and PM2.5. g Ranges 5-100% particle

197

number, 7-60% particle volume; 1-100% total particle mass, PM2.5; 6.1-18% PM1; 2.6-83% for PM10.

h Derived where data for both Average Number of Vehicles Per Day and Heavy Duty Vehicle Share (%)

were available. i Based on author classifications, including HDV (number of axles, gross vehicle mass

or length); LDV (wheel pair distance, vehicle length or weight). LDVs included cars and trucks with

specified vehicle weights; and HDVs with gross vehicle mass ranging from 3.5-12 tonne to > 25 tonne.

j Few reported diesel fuel sulphur content, where reported was < 15ppm, < 30 ppm Ultralow sulphur

diesel (ULSD) HDV; 300ppm Low sulphur diesel (LSD) for Bus, 24-480ppm for LDV and HDV.

Diesel, ULSD and LSD classed as diesel Fuel Type. k Buses - Bus Route, Central Bus District, Central

Business District – Aggressive Driving, Composite, CUEDC cycle, Manhattan, New York Bus,

Orange County Transit Authority, Route 22, Route 77, UDDS and Urban. Other vehicles - CUEDC

cycle, FTP, HHDDT; Hot UC, Hot Cycle, Cold Cycle, REP05, Steady State, UC and Urban. Trucks -

CBD–CBD14, HDCC. l Ranges < 50, 50-120 particle number, 86-113 particle volume; 80-120 total

particle mass; 30-90 PM1; 45-91 PM2.5; < 65 and 45-91 km/hr for PM10. m Engine Power: Reported in

two diesel bus studies (Jamriska et al. 2004; Ristovski et al. 2002). Instrumentation (in alphabetical

order) Aerodynamic Particle Sizer, Berner low pressure Impactor, Betameter, Beta-ray absorption

monitors, Chemical Mass Balance, Condensation Particle Counter, Differential Mobility Analyzer,

Differential Mobility Particle Sizer, DustTrak, Electrical Aerosol Analyser, Electrical Low Pressure

Impactor, Filters (Fibreglass, Glass fibre, Teflon, Quartz), Kleinfiltergerate, LIDAR-based VERSS and

remote sensing, Mass Single Stage Multidilutor, Micro-Orifice Uniform Deposit Impactor, Samplers,

Scanning Mobility Particle Sizer, Tapered Element Oscillating Microbalances, Ultrafine Condensation

Particle Counter. o Particle number 0.003-1 µm (dynamometer), 0.01-0.9 µm (tunnel), 0.003-20 µm

(vicinity of the road); particle volume 0.018-10 µm. Ranges particle number 0.003-1 µm

(dynamometer), 0.01-0.9 µm (tunnel studies), 0.003-20 µm (vicinity of the road), total particle number

count > 3 nm; 0.018-10 µm (particle volume). Few size ranges reported in total particle mass studies,

where reported 0.008-16 µm (dynamometer), 0.017-0.7 µm (tunnel), 0.008-0.3 µm, > 0.22 µm vicinity

of the road.

198

Table 5.1.3 Sample size of emission factors for different model variables examined in the statistical analysis, listed by particle

metric

Sample sizes related to Study Location and Road Environment statistical model variables Particle metric

Country of Study a

Study Location

Road Type

Speed Limit

Road Class

km/hr

Road Class

km/hr

Average No Vehicles per day d

HDV Share,

% d

Australia

Other b

USA & Canada

Dyno

Tunnel

Vicinity of

road

≤ 60

> 60

< 80

≥ 80

On-road fleets

On-road fleets

P Number 26 109 21 15 50 91 149 99 36 114 48 102 104 100 P Volume -- 57 -- -- 23 34 57 55 9 48 21 36 52 28 PM1 10 34 -- 10 9 25 34 15 11 31 30 12 34 25 PM2.5 18 7 60 17 18 50 72 c 20 26 52 31 38 7 38 PM10 19 50 57 45 23 58 96 c 33 58 31 47 40 38 54 Total Mass 3 12 184 165 2 32 119 c 8 97 65 97 65 2 18 TOTAL 76 269 322 252 125 290 240 230 237 341 274 293 237 263 Sample sizes related to Vehicle Type and Instrumentation statistical model variables

Particle metric

Vehicle Type

Fuel Type Reported e

Drive Cycle

Average Vehicle

Speed

Engine Power

Instrumentation

Size Range Measured g

P Number

156

34

6

13

2

156

156 (lower) f; 137(upper)

P Volume 57 -- -- 2 -- 57 57 (lower & upper ) PM1 44 16 17 16 -- 44 Particles with diameters < 1 µm PM2.5 85 33 17 26 4 85 Particles with diameters < 2.5 µm PM10 126 37 31 14 nr 126 Particles with diameters < 10 µm Total Mass 199 173 150 17 2 199 15 (lower & upper) TOTAL 667 293 221 88 8 667 232 (lower) f; 207 (upper)

199

a Country of Study is considered to have limited relevance for dynamometer measurements, except for Urban Drive Cycles, which were classed Urban Road Type

(see c below). b Other Countries included studies from Austria, Belgium, Denmark, Germany, Sweden, Switzerland and the United Kingdom. c Within these total

Road Type sample sizes, 92 emission factors related to total particle mass, 16 to PM2.5 and 23 to PM10 which were dynamometer measurements using an Urban

Drive Cycle. These data were classified in the statistical models as Urban Road Type. d Average Number of Vehicles Per Day and Heavy Duty Vehicle Share sample

sizes related to on-road vehicle fleets, and where data was available in studies for both these variables, the additional model variable Number of HDVs Per Day was

derived. e Not all studies reported vehicle Fuel Type, particularly studies of on-road vehicle fleets. f Some particle number studies reported only the lower Size

Range Measured, such as where total particle count was measured. Lower & upper – represent the lower size range and upper size ranges measured.

200

5.2.2 Statistical analysis of variables

We considered the relationship between reported tailpipe particle emission factors for

different particle sizes to the various study-specific explanatory variables (Table

5.1.3) using linear models. In particular, the model for particle number (here denoted

Yi) in study i is related to Vehicle Type (j=1,2,3) and Instrumentation (k=1, …10):

= and Yi = + ei

where is the intercept, is the effect of Vehicle Type j, and is the effect of

instrumentation k, and ei ~ N (0, σ2). A similar model applies on changing the

response (Yi) with different explanatory variables (Xi).

A separate statistical model was developed for each of the six particle metrics

examined in this study and the proportion of variation explained was calculated using

R2 = 1 – ∑ ei2 / Var (Yi). This is the fraction of variability in the dependent variable

(the emission factor) that may be accounted for, or explained, by variation in the

independent variable or variables, where the Var (Yi) is the usual sample variance of

Yi.

In this study the analysis of the data for the categorical variables involved fitting a

univariate general linear model (a multi-factor ANOVA). A stepwise technique, using

both forward and backward elimination, was then used to select the best model. For

the continuous variables linear regression analysis was undertaken with the variables

added as independent explanatory covariates in the general linear model. All analysis

201

was undertaken at a 5% level of significance. Statistically significant variables were

identified through ANOVA tests and post-hoc Scheffe multiple comparisons (Scheffe,

1959). The multiple comparison statistical tests were conducted at a 95% confidence

level for all categorical variables and their sub-classes to determine whether, within

each class of categorical variable, there were statistically significant differences

between the average published emission factor values for different sub-classes of

variable.

Analyses were undertaken in SPSS (SPSS Version 14.0) and from these average

particle emission factors for the different particle metrics, together with their standard

error and 95% confidence interval values, were derived. A separate statistical model

was developed for each of the six particle metrics examined in this study and model

coefficients of determination derived (R2), which provided information about the

fraction of variability in the dependent variable (the emission factor) that may be

accounted for, or explained, by variation in the independent variable or variables.

The statistical models produced average particle emission factors, and their

associated standard error and 95% confidence intervals. The standard error value

provides an indication of how reliable the model is as a means of predicting the

average particle emission factor for the particular combination of values of the

independent variables it relates to. The lower the standard error value, in relation to

its associated average emission factor, the more reliable the predicted average

emission factor may be considered. The lower and upper bound 95% confidence

202

interval values produced by the statistical models for each average emission factor

represent the range within which we can be 95% confident the true value will lie. In

some statistical models combinations of dependent and independent variables

produced high standard error values, and a lower bound 95% confidence interval

value which, although physically uninterpretable, can be obtained as a consequence

of the normal assumptions underlying the models, where these lower bound values

were obtained they were not reported.

5.2.3 Basis for selection of the most suitable emission factors

The wide range of different capabilities of Instrumentation used to derive emission

factors were not evaluated as an aim of this study. The rationale for selection of the

most suitable tailpipe particle emission factors to use in transport modelling and

health impact assessments from the five statistical model outputs was based on the

statistical robustness of the statistical model outputs, including consideration of

conservative average particle emission factors with the lowest standard errors,

narrowest 95% confidence intervals and largest sample sizes. Other factors taken into

account were the explanatory variables found for the statistical models. In

considering the explanatory variable Size Range Measured the focus was on

Instrumentation that measured the widest size ranges, including down to the lowest

size range.

203

5.3. RESULTS

This section presents the tailpipe particle emission factors considered the most

suitable to use in transport modelling and health impact assessments.

5.3.1 Sample sizes of emission factors examined in the statistical models

All average emission factors predicted by the statistical models and presented in this

paper are expressed in particle emissions generated per vehicle per kilometre driven.

It is important to note when considering the sample sizes of emission factors

examined in this study, that a single emission factor may represent one individual

vehicle (or group of vehicles) tested on a dynamometer, or be the average emission

factor derived for a vehicle type (eg., light duty vehicles) travelling in a vehicle fleet

on a road or in a tunnel. Hence, the total sample size examined in this study of 667

emission factors represents a relatively very large sample of motor vehicles.

5.3.2 Statistical models developed to derive average emission factors

Six statistical models were proposed for particle number, particle volume, total

particle mass, PM1, PM2.5 and PM10. The analysis revealed that the statistical models

developed for particle number, particle volume, PM1 and PM2.5 were robust, and

explained 86%, 93%, 87% and 65% respectively of the variation in published

emission factors. However the PM10 model was found to be less robust as it

explained only 47% of the variation in published emission factor values. PM10

emission factors derived from studies conducted on or near roads may have been

204

influenced by varying quantities of resuspended road dust occurring at the PM10 size

range, leading to higher values than those derived from dynamometer and tunnel

studies, and which may have confounded the ability of the statistical model to

explain the variation in published emission factors.

The sixth statistical model for total particle mass was found to be a null model, as no

explanatory variables were identified. This result is likely to be attributed to the fact

that most of the studies simply measured total particle mass, and not different

subsets of particle mass size fractions which typically have differing proportions of

particle mass associated with them.

The final set of average tailpipe particle emission factors considered the most

suitable for use in transport modelling and health impact assessments for different

vehicle and road type combinations, together with their 95% confidence interval

values, are presented in Table 5.1.4. Aspects related to their selection are discussed

below.

205

Table 5.1.4. Tailpipe particle emission factors for motor vehicles considered the most suitable to use in transport modelling and health

impact assessments, derived based on statistical analysis in this study of 667 emission factors in the international published literature

Particle metric

Emission unit per vehicle

Explanatory variables (in bold font)

Fleet emission factor

95% confidence interval

HDV emission factor


LDV emission factor


Bus emission factor


Vehicle Type & Instrumentation

CPC c 7.26 3.85-10.66 65 60.19-69.81 3.63 a-9.85 -- --

Particle number

1014 particles per km

SMPS c -- -- -- -- -- -- 3.08b a-9.30 Particle volume

Cubic cm per km

Vehicle Type, Size Range Measured & Speed Limit on the Road

18-300nm, <= 60 km/hr 0.07 a-0.19 0.93 0.81-1.06 0.03 a-0.15 -- -- 18-700nm, > 60 km/hr 0.04 a-0.16 0.41 0.32-0.49 0.05 a-0.3 -- -- PM1 mg per km Vehicle Type & Fuel

Type Fuel not specified Fuel not specified & diesel Combined

36

--

2-70

--

--

287

--

257-317

16

--

a-50

--

--

--

--

-- PM2.5 mg per km Vehicle Type &

Instrumentation

TEOM & DMPS c DustTrak All Instrumentation

60 -- --

a-166 -- --

-- --

302

-- --

236-367

-- 33 --

-- a-80

--

-- 299b

--

-- 205-394

-- PM10 mg per km Vehicle Type &

Road Type

Boulevard -- -- 4815 3459-6171 454 a-1413 4130ce 2774-5486 Urban 688 a-1546 538 a-1145 156 a-635 1089ce 306-1872 Freeway 200 a-2118 2500 1144-3856 285 a-1244 -- -- Highway 66 a-1421 840 a-1947 141 a-924 -- -- Motorway 77 a-1432 213 a-1568 63 a-1419 -- -- Rural Area 67 a-1984 394 a-2312 46 a-1964 -- -- Tunnel

Dynamometer e 306 a-884 1019 236-1802 14 a- 797 --

313ce --

a-753

206

a The lower bound 95% confidence interval value calculated to be negative and therefore is not valid. These values, although physically uninterpretable, can be

obtained as a consequence of the normal assumptions underlying the models, and hence are not reported. b Diesel buses. c Buses – Fuel not specified (can be

assumed to be Diesel-fuelled due to the timing and location of the studies), principally Diesel-fuelled buses. d Condensation Particle Counter (CPC), Scanning

Mobility Particle Sizer (SMPS), Tapered Element Oscillating Microbalances(TEOM) and Differential Mobility Analyzer (DMA). e The average dynamometer

emission factor for buses for PM10 is also presented; as the on-road boulevard and urban Road Type studies were reported to be affected by very high levels of

resuspended road dust and the influence of variation in acceleration and speed (Abu-Allaban et al. 2003).

207

5.4. DISCUSSION

This section discusses the tailpipe particle emission factors considered the most

suitable to use in transport modelling and health impact assessments for different

particle metrics; and the results of statistical tests that examined differences in

mean values of published emission factors.

5.4.1. Statistical models used to derive average emission factors

These are discussed below for different particle metrics.

Particle number model: This statistical model explained 86% of the variation in

published emission factors (n=156). Vehicle Type and Instrumentation were the

explanatory model variables and emission factors were available for 10 different

Instrumentation. In selecting the most suitable emission factors, it was important

to consider Instrumentation that measured the lowest possible size range,

including down to 0.003 µm where particle number emissions tend to be very

prolific. This lower limit size range is commonly measured by the Condensation

Particle Counter (CPC), which estimates particle count, and emission factors

based on CPC measurements were available in the literature for Fleet, light duty

vehicles (LDV) and heavy duty vehicles (HDV). However particle number

emission factors for Diesel buses were restricted to those derived from Scanning

Mobility Particle Sizer (SMPS) measurements.

The SMPS focuses on estimating particle size distribution (as opposed to total

particle count) and does not measure the lower size range of the nucleation mode

< 0.01 µm. The lower size window for the SMPS is commonly set higher than for

208

the CPC, usually in the range 0.01-0.02 µm, whereas for the CPC the range is

usually 0.002-0.01 µm, which means that generally the CPC measures the lower

size range of the nucleation mode and the SMPS does not.

Particle volume model: This statistical model explained 93% of the variation in

published emission factors (n=57) and the explanatory model variables included

Vehicle Type, Speed Limit on the Road and Size Range Measured. Consideration

was given to selecting emission factors which related to the broadest size ranges

measured, including down to the lowest possible size range, and to different

reported Speed Limits on the Road. Most of the average particle volume emission

factors, and their 95% confidence interval values, produced by the statistical

model were less than 1 cm3 per vehicle per kilometre. For almost all the particle

volume emission factors Speed Limit on the Road or in the tunnel was reported,

and the availability of this data may have contributed to the statistical model’s

high R2 value of 0.93.

PM1 model: The explanatory variables for this statistical model were Vehicle

Type and Fuel Type, which explained 87% of the variation in published emission

factors (n=44). Emission factors examined in the analysis included those derived

for diesel vehicles measured on a dynamometer; and from studies conducted on or

near roads or in tunnels where the Fuel Type was not specified. The literature

review revealed that at the time of this study the majority of LDVs were petrol-

fuelled and HDVs diesel-fuelled, hence it can be assumed that these were the

dominant Fuel Types in the vehicle fleets studied.

209

Few data are available in the literature for PM1 emission factors, and given that

most motor vehicle particle emissions are < 1 µm (dominated by ultrafine

particles) this is an important size range to have a comprehensive database for.

Recent research found that a combination of PM1 and PM10 mass ambient air

quality standards are likely to be more suitable to control combustion and

mechanically-generated sources, such as motor vehicles, than the current

standards of PM2.5 and PM10 (Morawska et al. 2008), further emphasising the

importance of deriving PM1 emission factors.

PM2.5 model: Sixty-five percent of the variation in published emission factors

(n=85) was explained by this statistical model, and its explanatory variables were

Vehicle Type and Instrumentation. Emission factors were examined for 8

different Instrumentation reported in the literature.

PM10 model: For PM10 the explanatory variables were Vehicle Type and Road

Type, and this statistical model explained 47% of the variation in published

emission factors (n=126). This low value for R2 is reflected in the large values for

standard errors (in relation to the predicted average emission factor) and high

values for upper bound 95% confidence intervals produced by the statistical

model. The presence of varying amounts of resuspended road dust at the PM10

size range are likely to have influenced emission factors derived in on-road

studies, as compared to those derived from dynamometer and tunnel studies, and

is likely to have confounded the explanation of variation. Few methods are

available for discriminating road dust from tailpipe emissions, particularly at the

PM2.5 and PM10 size ranges, and quantities of road dust can vary depending on the

210

construction material of road surfaces and their maintenance, climatic conditions,

and other factors such as vehicle speed and traffic volumes.

Few bus emission factors are available derived from on-road measurements and

those available and included in the statistical model related to measurements on

boulevard and urban Road Types in the US (Abu-Allaban et al. 2003a). However

the authors of this study considered their high PM10 emission factors were

influenced by significantly high contributions from resuspended road dust and,

within each vehicle category, by the effects of speed and acceleration (Abu-

Allaban et al. 2003a). For this reason the average emission factor for buses

derived from dynamometer measurements is also presented as a suitable emission

factor in Table 5.1.4, in addition to average emission factors for bus for urban and

boulevard Road Types, as it is considered more conservative and unlikely to be

affected by high rates of resuspended road dust. This average dynamometer

emission factor for bus included emission factors for a wide range of different

urban bus Drive Cycles.

211

Total particle mass model: No statistically significant variables were identified

for this statistical model. The sample size was 199 and overall mean from this null

model was 158 mg/km for all combined Vehicle Types; 158 mg/km for bus, and

91 mg/km for Fleet, 380 mg/km for heavy duty vehicles (HDV) and 32 mg/km for

light duty vehicles (LDV). The inability to identify relationships in this statistical

model may stem from the fact that these studies measured a broad range of

different particle sizes, and most emission factors were not derived segregated by

different subsets of particle mass fractions, but simply measured total particle

mass.

5.4.2 Statistical differences between published emission factors

Post-hoc Scheffe’s multiple comparison statistical tests (Scheffe 1959) were used

to investigate the differences in means between levels corresponding to sub-

classes within all categorical variables, irrespective of whether they had a

significant effect on the response variable (the published emission factor value), at

a 95% confidence level. The findings of these statistical tests are discussed

below.

Country of Study; Study Location; Road Types vs Dynamometer: It was found

that the variables Country of Study and Study Location (dynamometer, on or near

the road, tunnel) were not statistically significant in explaining the variation in the

means of published emission factors for most particle metrics. When comparing

the means for different Road Types with those derived from dynamometer

measurements, statistically significant differences were only found between

dynamometer and motorway (PM1) and dynamometer and boulevard Road Types

212

(PM10). These differences, however, are likely to have been influenced by high

speed scenarios, as the PM1 study measured emissions on a motorway in

Switzerland with a speed limit of 120km/hr (Imhof et al, 2005a) and the PM10

study in the US attributed the significantly high PM10 emission rates to

contributions from resuspended road dust and to the influence of variation in

acceleration and speed (Abu-Allaban et al. 2003a).

Vehicle Type and Fuel Type: For Vehicle Type statistically significant

differences were found between the means for Fleet and HDV for particle

number, PM1 and PM2.5; and between the means for Fleet and LDV for PM2.5.

The means for LDV and HDV were found to be statistically significantly different

for all particle metrics. No statistically significant differences were found between

the means of different Fuel Types for particle number, or between the means for

different Fuel Types for total particle mass. However statistically significant

differences were found between the means for petrol and diesel Fuel Types for

PM10.

Instrumentation: No statistically significant differences were found between

mean values measured by different Instrumentation for PM2.5 and total particle

mass. However, a significant difference was found between the mean value for

published emission factors for particle number derived from the Condensation

Particle Counter (CPC) of 22.69 x 1014 particles per vehicle per km and the

Scanning Mobility Particle Sizer (SMPS) of 2.08 x 1014 particles per vehicle per

km, highlighting a major difference between the results of these two measurement

techniques, which requires investigation as a broader issue.

213

Statistically significant differences were found for PM1 between the means for the

Aerodynamic Particle Sizer (APS) and Betameter and between the APS and Beta-

ray absorption monitors, however these differences are likely to be influenced by

the fact that the PM1 measurements related exclusively to diesel vehicles (LDVs

and HDVs) tested on dynamometers in Australia. Higher values of emission

factors are likely to be associated with diesel-fuelled vehicles as compared to

petrol and other fuelled-vehicles.

Size Range Measured for particle number: In relation to the Size Range

Measured for particle number, no statistically significant differences were found

for the lower and upper size ranges measured for particle number between the

average emission factors for the various levels of each of the categorical

variables, after accounting for the associated variability of these estimates.

Emission factors derived using the CPC for total particle count which reported

only the lower size ranges measured (and did not report the upper size range

measured) were unable to be included in these statistical tests. Their inclusion

may have led to a different result as the CPC generally measures down to 0.002

µm, where particle numbers are very prolific.

5.4.3 Relevance and application of the average particle emission factors

presented in this study

A general conclusion from examination of the results of the post-hoc multiple

comparison tests discussed above is that these findings support the relevance and

applicability of using the average emission factors derived in this study for

modelling tailpipe particle emissions from urban fleets in developed countries.

214

Where statistically significant differences were found these were generally

associated with emission factors for diesel-fuelled vehicles, or related to high

speed scenarios or to conditions with significantly high levels of resuspended road

dust.

It is suggested that when using the average emission factors presented in this

study, that three calculations be made. Firstly, a calculation using the relevant

average emission factor, and two further calculations using the lower and upper

bound 95% confidence interval values associated with the average emission factor

(where available). It should be noted that where a single, individual road is

concerned, the lower and upper bound 95% confidence interval values will be

more widely distributed than those reported in this study.

5.5. CONCLUSIONS

This paper presents a comprehensive set of tailpipe particle emission factors,

covering the full size range of particles emitted by motor vehicles, which are

suitable for use in transport modelling and health impact assessments of urban

fleet emissions in developed countries. These emission factors were derived for

different Vehicle and Road Type combinations based on advanced statistical

analysis of a large body of data on emission factors derived from measurement

studies, and include emission factors for particle number and different fractions of

particle mass.

215

The average emission factors were derived from statistical models which were

found to explain 86%, 93%, 87% and 65% of the variation in published emission

factors for particle number, particle volume, PM1, and PM2.5 respectively, and

hence are concluded to have been derived from robust models. The statistical

model for PM10, however, explained only 47% of the variation in published

emission factors and it is likely may have been confounded by the effects of

resuspended road dust at this size range.

The explanatory variables identified in the statistical models included Vehicle

Type (all particle metrics), Instrumentation (particle number and PM2.5), Fuel

Type (PM1), Road Type (PM10) and Size Ranged Measured and Speed Limit on

the Road (particle volume), and we conclude that these are important variables to

consider in design and interpretation of data in emission factor studies for

different particle metrics.

The relevance and suitability of the derived set of tailpipe particle emission

factors for use in urban areas in developed countries is supported by the findings

from the statistical analysis of published emission factors in the international

literature, which were as follows.

First, statistical analysis of published emission factors revealed that few

statistically significant differences were found between the mean values for

different particle metrics for Country of Study and Study Location (dynamometer,

on or near a road, tunnel).

216

Second, few statistically significant differences were found between the means of

published emission factors derived in dynamometer studies and those derived for

different Road Types, except under high speed scenarios or conditions with

significantly high levels of resuspended road dust, suggesting that for most

particle metrics the two methods provide generally similar results.

Third, statistically significant differences were found between mean published

emission factors for LDVs and HDVs for all particle metrics; and between petrol

and diesel-fuelled vehicles for PM10, consistent with higher emission rates that

would be expected from diesel-fuelled vehicles, as compared to petrol and other

fuelled vehicles.

5.6. RECOMMENDATIONS AND PERSPECTIVES

The average emission factors presented in this study are suitable for developing

road-link based inventories, quantifying the spatial distribution of particle

concentrations and for developing health impact assessments, covering the full

size range of particles emitted by fleets. They are particularly useful for regions

which may have insufficient funding to conduct measurements, or little or no data

upon which to derive emission factors for their local region.

Better scientific techniques and tools are needed to produce data that can be used

to model fleet emissions, as variations were found between different

Instrumentation and methods used to derive emission factors. For example,

statistically significant differences were found between the mean values of

published emission factors for particle number measured by the Condensation

217

Particle Counter of 22.69 x 1014 particles per vehicle per km, as compared to

Scanning Mobility Particle Sizer (SMPS) Instrumentation of 2.08 x 1014 particles

per vehicle per km, a difference which requires further investigation as a broader

issue. Particle number emission factors for buses are rare and limited to estimates

derived from SMPS measurements, which generally do not measure down to the

lower size range of 0.002 µm in the nucleation mode where particle number tends

to be very prolific.

While this study examined available tailpipe particle emission factors in the

international literature, more studies are needed that derive speed-related particle

emission factors for on-road and tunnel studies, particularly for speeds less than

50 km/hr to model congestion. More studies are also needed to derive emission

factors for particle number for buses, and for different subsets of particle number

< 1 µm, such as for ultrafine and nanoparticles (diameters < 0.05 µm), where

particle number tends to be very prolific, for different Vehicle Types. Limited

particle emission factor data are available for motor vehicles for particle volume,

particle surface area, PM1, brake and tyre wear, road grade, engine power, and for

buses measured on different Road Types.

218

5.7. REFERENCES

Abu-Allaban, M., 2002. Exhaust particle size distribution measurements at the

Tuscarora Mountain tunnel. Aerosol Science and Technology 36(6), 771-789.

Abu-Allaban, M., Gillies, J.A., Gertler, A.W., 2003a. Application of a multi-lag



Abu-Allaban, M., Gillies, J.A., Gertler, A.W., Clayton, R., Proffitt, D., 2003b.

Tailpipe, resuspended road dust, and brake-wear emission factors from on-road

vehicles. Atmospheric Environment 37(37), 5283-5293.

Ahlvik P, Eggleston S, Goriben N, Hassel D, Hickman AJ, Joumard R,

Ntziachristos L, Rijkeboer R, Samaras Z, Zierock K.H (1997) COPERT II




ARB's, 2002. Study of Emissions from Two "Late Model" Diesel and CNG

Heavy-Duty Transit Buses. California Air Resources Board, 12th CRC On-Road

Vehicle Emissions Workshop, April 15-17, San Diego.

219

Ayala, A., Kado, N.Y., Okamoto, R.A., 2002. Diesel and CNG Heavy-duty

Transit Bus Emissions over Multiple Driving Schedules: Regulated Pollutants and

Project Overview. Society of Automotive Engineers SAE 2002-01-17221-13.

Bradley, M.J., 2000. Hybrid-Electric Drive Heavy-Duty Vehicle Testing Project;

Final Emissions Report. Northeast Advanced Vehicle Consortium, Defense

Advanced Research Projects Agency, West Virginia University, USA.

Bellasio R, Bianconi R, Corda G, Cucca P (2007) Emission inventory for the road

transport sector in Sardinia (Italy). Atmospheric Environment 41, 677-691.

Cadle, S.H., Mulawa, P.A., Ball, J., Donase, C., Weibel, A., Sagebiel, J. C.,



& Technology 31(12), 3405-3412.

Cadle, S.H., Mulawa, P., Groblicki, P., Laroo, C., Ragazzi, R. A., Nelson, K.,



35(1), 26-32.

CARB., 2001. Heavy-Duty Emissions Laboratory, Heavy Duty Testing and Field

Support Section, California Air Resources Board. Report No. 01-01.

220

CARB, 2002. EMFAC2001/EMFAC200. Calculating emissions inventories for

vehicles in California, User’s Guide, California California Air Resources Board.

Chatterjee, S., Conway, R., Lanni, T., Frank, B., Tang, S., Rosenblatt, D., Bush,

C., Lowell, D., Evans, J., McLean, R., Levy, S., 2002. Performance and

Durability Evaluation of Continuously Regenerating Particulate Filters on Diesel

Powered Urban Buses at NY City Transit – Part II. Society of Automotive

Engineers SAE 2002-01-0430.

Clark, N.N., Lyons, D.W., Bata, R.M., Gautam, M., Wang, W.G., Norton, P.,

Chandler, K., 1997. Natural Gas and Diesel Transit Bus Emissions: Review and

Recent Data. Society of Automotive Engineers Tech. Pap. No. 973203.

Clark, N.N., Lyons, D.W., Rapp, B.L., Gautam, M., Wang, W.G., Norton, P.,

White, C., Chandler, C., 1998. Emissions from Trucks and Buses Powered by

Cummins L-10 Natural Gas Engines. Society of Automotive Engineers Tech. Pap.

No. 981393.

Clark, N.N., Gautam, M., Rapp, B.L., Lyons, D.W., Graboski, M.S., McCormick,

R. L., Alleman, T. L., Norton, P., 1999. Diesel and CNG Transit Bus Emissions

Characterization by Two Chassis Dynamometer Laboratories: Results and Issues.


221



european driving cycle conditions. CONCAWE, Brussels Report no. 98/51.

Corsmeier, U., Imhof, D., Kohler, M., Kuhlwein, J., Kurtenbach, R., Petrea, M.,

Rosenbohm, E., Vogel, B., Vogt, U., 2005. Comparison of measured and model-

calculated real-world traffic emissions. Atmospheric Environment 39(31), 5760-

5775.



Frey HC, Unal A, Chen J (2002a) Recommended strategy for on-board emission

data analysis and collection for the new generation model. Prepared for Office of

Transportation and Air Quality, US Environmental Protection Agency.

Frey HC, Unal A, Chen J, Li S, Xuan C (2002b) Methodology for developing

modal emission rates for EPA's multi-scale motor vehicle and equipment

emission estimation system, North Carolina State University for the Office of

Transportation and Air Quality, US Environmental Protection Agency.





222

Gertler, A.W., Gillies, J.A., Pierson, W.R., Rogers, C.F., Sagebiel, J. C., Abu-





Hansson, H.C., 2003. Model simulation of ultrafine particles inside a road tunnel.


Gidhagen, L., Johansson, C., Langner, J., Olivares, G., 2004a. Simulation of NOx

and ultrafine particles in a street canyon in Stockholm, Sweden. Atmospheric

Environment 38(14), 2029-2044.

Gidhagen, L., Johansson, C., Omstedt, G., Langner, J., Olivares, G., 2004b.

Model simulations of NOx and ultrafine particles close to a Swedish highway.


Gillies, J.A., Gertler, A.W., Sagebiel, J.C., Dippel, W.A., 2001. On-road

particulate matter (PM2.5 and PM10) emissions in the Sepulveda Tunnel, Los

Angeles, California. Environmental Science & Technology 35(6), 1054-1063.


Determination of average emission factors for vehicles on a busy road.


223

Gramotnev, G., Ristovski, Z.D., Brown, R.J., Madl, P., 2004. New methods of

determination of average particle emission factors for two groups of vehicles on a

busy road. Atmospheric Environment 38(16), 2607-2610.

Harrison R, Jones M, Collins G (1999) Measurements of the Physical Properties

of Particles in the Urban Atmosphere. Atmospheric Environment 33, 309-321.




Holmen, B., Chen, Z., Davila, A., Gao, O., Vikara, D.M., 2005. Particulate matter

emissions from Hybrid Diesel-electric and Conventional Diesel Transit Buses:

Fuel and Aftertreatment Effects. The University of Connecticut Report No. JHR

05-304.

Hueglin, C., Buchmann, B., Weber, R. O., 2006. Long-term observation of real-

world road traffic emission factors on a motorway in Switzerland. Atmospheric

Environment 40(20), 3696-3709.


Baltensperger, U., 2005a. Real-world emission factors of fine and ultrafine

aerosol particles for different traffic situations in Switzerland. Environmental


224


Rodler, J., Sturm, P., McCrae, I., Sjodin, A., Baltersperger, U., 2005b. Aerosol

and NOx Emission Factors and Submicron Particle Number Size Distributions in

Two Road Tunnels with Different Traffic Regimes. Atmospheric Chemistry and

Physics Discussions 55127-55166.

Imhof, D., Weingartner, E., Vogt, U., Dreiseidler, A., Rosenbohm, E., Scheer, V.,

Vogt, R., Nielsen, O.J., Kurtenbach, R., Corsmeier, U., Kohler, M.,

Baltensperger, U., 2005c. Vertical distribution of aerosol particles and NOx close


Jamriska, M., Morawska, L., 2001. A model for determination of motor vehicle

emission factors from on-road measurements with a focus on submicrometer

particles. Science of the Total Environment 264(3), 241-255.



38(24), 6701-6709.




225

Kado, N.Y., Okamoto, R.A., Kuzmicky, P.A., Kobayashi, R., Ayala, A., Gebel,

M. E., Rieger, P.L., Maddox, C., Zafonte, L., 2005. Emissions of toxic pollutants

from compressed natural gas and low sulfur diesel-fueled heavy-duty transit buses

tested over multiple driving cycles. Environmental Science & Technology 39(19),

7638-7649.

Ketzel, M., Wahlin, P., Berkowicz, R., Palmgren, F., 2003. Particle and trace gas

emission factors under urban driving conditions in Copenhagen based on street

and roof-level observations. Atmospheric Environment 37(20), 2735-2749.


Minnesota highways. Atmospheric Environment 38(1), 9-19.

Kristensson, A., Johansson, C., Westerholm, R., Swietlicki, E., Gidhagen, L.,

Wideqvist, U., Vesely, V., 2004. Real-world traffic emission factors of gases and

particles measured in a road tunnel in Stockholm, Sweden. Atmospheric


Lanni, T., Frank, B. P., Tang, S., Rosenblatt, D., Lowell, D., 2003. Performance

and Emissions Evaluation of Compressed Natural Gas and Clean Diesel Buses at

New York City's Metropolitan Transit Authority. SAE 2003-01-0300.

Lowell, D.M., Parsley, W., Bush, C., Zupo, D., 2003. Comparison of Clean Diesel

buses to CNG Buses. 9th Diesel Engine Emissions Reduction (DEER) Workshop,

Newport, RI, USA, 24-28 August.

226

Mazzoleni, C., Kuhns, H.D., Moosmuller, H., Keislar, R.E., Barber, P.W.,

Robinson, N. F., Watson, J.G., 2004. On-road vehicle particulate matter and

gaseous emission distributions in Las Vegas, Nevada, compared with other areas.

Journal of the Air & Waste Management Association 54(6), 711-726.

Morawska, L., Bofinger, N.D., Kocis, L., Nwankwoala, A., 1998. Submicrometer

and supermicrometer particles from diesel vehicle emissions. Environmental





University of Technology, Brisbane.

Morawska L, Salthammer T (2003) Chapter 3: Motor Vehicle Emissions as a


Wiley-VCH; 297-318.

Morawska L, Moore M R, Ristovski ZD (2004) Health Impacts of Ultrafine



227




39(23), 9130-9139.

Morawska L, Keogh DU, Thomas SB, Mengersen K (2008) Modality in ambient

particle size distributions and its potential as a basis for developing air quality

regulation. Atmospheric Environment 42 (7), 1617-1628.




Ntziachristos L, Samaras Z, Eggleston S, Goriben N, Hassel D, Hickman AJ,

Joumard R, Rijkeboer R, White L, Zierock K H (2000) COPERT III Computer

programme to calculate emissions from road transport: methodology and emission

factors (version 2.1). Technical report prepared by the European Environment

Agency, Copenhagen, Report 49.

Ristovski, Z.D., Morawska, L., Ayoko, G.A., Jayaratne, E.R., Lim, M., 2002.




228




SAE., 2001. Performance and Durability Evaluation of Continuously

Regenerating Particulate Filters on Diesel powered Urban Transit Buses at NY

City Transit. Society of Automotive Engineers SAE 2001-01-0511.

SAE., 2002a. Performance and Durability of Continuously Regenerating

Particulate Filters on Diesel powered Urban Transit Buses at NY City Transit -

Part II. Society of Automotive Engineers SAE 2002-01-0430.

SAE., 2002b. Year-Long Evaluation of Trucks and Buses Equipped with Passive

Diesel Diesel Particulate Filters. Society of Automotive Engineers SAE 2002-01-

0433.

SAE., 2003a. Oxidation catalyst effect on CBG Transit Bus Emissions. Society of

Automotive Engineers SAE 2003-01-1900.

SAE., 2003b. Performance and Emissions Evaluation of Compressed Natural Gas

and Clean Diesel Buses at New York City's Metropolitan Transit Authority.


229

Samaras Z, Ntziachristos L, Thompson N, Hall D, Westerholm R, Boulter P

(2005). Characterisation of Exhaust Particulate Emissions from Road Vehicles,

PARTICULATES program, European Commission. Contract No 2000-

RD.11091, source http://lat.eng.auth.gr/particulates/downloads.htm.

Scheffe H (1959) The Analysis of Variance, John Wiley & Sons, Inc.




Shah, S.D., Cocker, D.R., Miller, J.W., Norbeck, J.M., 2004. Emission rates of

particulate matter and elemental and organic carbon from in-use diesel engines.


Shi J, Harrison RM (1999) Investigation of ultrafine particle formation during

diesel exhaust dilution. Environmental Science & Technology 33, 3730-3736.

Shi J P, Khan AA, Harrison RM (1999) Measurements of ultrafine particle

concentration and size distribution in the urban atmosphere. The Science of the


Shi J, Evans D, Khan A, Harrison R (2001) Sources and Concentration of

Nanoparticles (<10 nm Diameter) in the Urban Atmosphere. Atmospheric


230

Smit R, Smoker, R, Rab, E (2007) A new modelling approach for road traffic






Ubanwa, B., Burnette, A., Kishan, S., Fritz, S.G., 2003. Exhaust particulate matter

emission factors and deterioration rate for in-use motor vehicles. Journal of

Engineering for Gas Turbines and Power-Transactions of the Asme 125(2), 513-

523.

USEPA (1993) User's Guide to MOBILE5A, Mobile source emissions factor


Wahlin P, Palmgren F, Van Dingenen R (2001) Experimental studies of ultrafine

particles in streets and the relationship to traffic. Atmospheric Environment 35,

S63-S69.

Venkatram, A., Fitz, D., Bumiller, K., Du, S.M., Boeck, M., Ganguly, C., 1999.

Using a dispersion model to estimate emission rates of particulate matter from

paved roads. Atmospheric Environment 33(7), 1093-1102.

231

Walker JL, Li J, Srinivasan S, Bolduc D (2008) Travel Demand Models in the

Developed World: Correcting for Measurement Errors Transportation Research

Board 87th Annual Meeting Washington.




Zhang, K.M., Wexler, A.S., Niemeier, D.A., Zhu, Y.F., Hinds, W. C., Sioutas, C.,

2005. Evolution of particle number distribution near roadways. Part III: Traffic,

analysis and on-road size resolved particulate emission factors. Atmospheric

Environment 39(22), 4155-4166.

Zhu, Y. F., Hinds, W. C., 2005. Predicting particle number concentrations near a

highway based on vertical concentration profile. Atmospheric Environment 39(8),

1557-1566.

232

CHAPTER 5.2

DERIVATION OF MOTOR VEHICLE

PARTICLE EMISSION FACTORS - STATISTICAL

MODEL OUTPUTS

5. INTRODUCTION

This Chapter presents the outputs of five statistical models which were developed

to produce average emission factors for different Vehicle Types, and identify the

most suitable particle emission factors to use in transport modelling and health

impact assessments. The method for developing these statistical models is

discussed in detail in the paper presented in Chapter 5.1.

A multiple comparison plot is also presented in this Chapter which depicts the

statistical relationships between the average values of published emission factors

in terms of categorical variables examined in the statistical analysis. The results

shown in this plot are commented on in the paper presented in Chapter 5.1.

Additional comments discussing average emission factors derived by the

statistical models for particle volume and PM10, and the rationale for selection of

LDV, HDV and bus PM10 emission factors used in developing the urban South-

East Queensland inventory (presented in paper three, Chapter 6) are also

discussed.

233

5.1. STATISTICAL MODEL OUTPUTS

The statistical model outputs shown in Tables 5.2.1-5.2.5 present the explanatory

model variables, their average emission factors and corresponding 95%

confidence intervals and standards errors for particle number, particle volume,

PM1, PM2.5 and PM10.

From these statistical model outputs, the most suitable emission factors for

different Vehicle Types and different particle metrics were selected. These were

selected based on their statistical characteristics, including consideration of

conservative average particle emission factors with the lowest standard errors,

narrowest 95% confidence intervals and largest sample sizes. Other factors

considered for some particle metrics were Size Range Measured and Road Type.

The most suitable emission factors to use in transport modelling are shown in

Tables 5.2.1-5.2.5 in bold italics shaded gray (and summarised in Table 5.1.4,

Chapter 5.1).

234

Emission factor sample sizes in the statistical models

It should be noted that when considering the sample size of emission factors in

Tables 5.2.1-5.2.5, that one single emission factor may represent emissions from

one individual vehicle (or group of vehicles) tested on a dynamometer, or

represent the average emissions of an entire vehicle fleet measured on or near a

road or in a tunnel. A single emission factor may also represent an average

emission factor derived for a vehicle class, such as for all LDVs, HDVs or buses,

travelling in a large vehicle fleet on a road or in a tunnel. Therefore the total

sample size examined in this study of 667 emission factors represents a relatively

very large sample of motor vehicles.

235

Table 5.2.1. Particle number model explanatory variables and average particle number emission factors

Sample

Size

95% Confidence Interval

Vehicle Type

Average particle

emission factor value

1014 particles per vehicle per

km

Standard Error

Lower Bound

Upper Bound

HDV 1 0.02 5.44 (a) 10.77

Instrumentation, Size Range Measured* and Study Location(b) APS 700-20,000nm Vicinity of the road Combined 1 0.02 5.44 (a) 10.77

Fleet 10 7.26 1.72 3.85 10.66 HDV 5 65.00 2.43 60.19 69.81 LDV 3 3.63 3.14 (a) 9.85

CPC > 6nm, > 7nm, 3-3000nm Vicinity of the road

Combined 18 25.30 1.44 22.44 28.15 Fleet 2 3.35 3.85 (a) 10.96 HDV 1 52.00 5.44 41.24 62.76 LDV 1 1.40 5.44 (a) 12.16

CPC, DMPS >3nm, 10-700nm Vicinity of the road

Combined 4 18.92 2.87 13.25 24.59 LDV 3 1.06 3.14 (a) 7.27 DMA 10-237.2nm

Dynamometer Combined 3 1.06 3.14 (a) 7.27 Fleet 5 5.40 2.43 0.59 10.21 HDV 5 29.32 2.43 24.51 34.13 LDV 8 2.82 1.92 (a) 6.63

DMPS <10nm, 3-900nm Tunnel

Combined 18 12.51 1.31 9.92 15.11 LDV 2 1.42 3.85 (a) 9.02 EAA 10-1000nm

Dynamometer Combined 2 1.42 3.85 (a) 9.02 Fleet 2 1.80 3.85 (a) 9.41 HDV 2 7.79 3.85 0.18 15.40 LDV 2 1.22 3.85 (a) 8.83

ELPI 30-10,000nm Vicinity of the road

Combined 6 3.60 2.22 (a) 8.00 LDV 6 0.68 2.22 (a) 5.07 ELPI, UCPC

> 3nm Dynamometer Combined 6 0.68 2.22 (a) 5.07

BUS – Diesel (e)

3 3.08 3.14 (a) 9.30

Fleet 41 1.28 0.85 (a) 2.96 HDV 26 4.86 1.07 2.75 6.97

SMPS 3-900nm Mix of Tunnel, Vicinity of the road & Dynamometer(c)

LDV 26 0.46 1.07 (a) 2.57 Combined 96 2.42 0.90 0.65 4.19

LDV 2 1.22 3.85 (a) 8.82 SMPS, DMPS 15.7-685.4nm Dynamometer

Combined

2 1.22 3.85 (a) 8.82 BUS 3 3.08(d) 3.14 (a) 9.30 Fleet 60 3.18(d) 1.12 0.97 5.40 HDV 40 26.50(d) 1.56 23.42 29.57 LDV 53 1.39(d) 1.15 (a) 3.67

*ALL

Combined 156 8.48(d) 0.72 7.06 9.91

236

(a) The lower bound 95% confidence interval value calculated to be negative

and therefore is not valid.

(b) The minimum and maximum size range measured by the Instrumentation.

(c) Includes 1 HDV and 1 Bus Dynamometer measurement.

(d) Based on modified population marginal mean.

(e) 300 Bus trips were measured in a tunnel and 12 Buses on dynamometer.

Instrumentation: CPC - Condensation Particle Counter; DMA - Differential

Mobility Analyser ; DMPS - Differential Mobility Particle Sizer ; EAA –

Electrical Aerosol Analyser ; ELPI – Electrical Low Pressure Impactor ;

SMPS – Scanning Mobility Particle Sizer; UCPC – Ultrafine Condensation

Particle Counter.

*ALL – Instrumentation and Vehicle Types. ** ALL emission factors

Combined.

237

Table 5.2.2. Particle volume model explanatory variables and average particle volume emission factors

95% Confidence

Interval

Vehicle Type

Study Location & Size Range Measured, (nm)

Urban vs Non-Urban Roads, Speed Limit on the Road =<60km/hr, >60km/hr

Sample Size

Average particle

emission factor value

per cubic cm per km

Standard Error

Lower Bound

Upper Bound

=<60

1 0.01 0.06 (a) 0.13

>60 4 0.01 0.03 (a) 0.07

18-100 Tunnel and Vicinity of the road

Combined

5 0.01 0.03 (a)

0.08 =<60 1 0.07

0.06

(a) 0.19

>60 4 0.05 0.03 (a) 0.12

18-300 Vicinity of the road =<60; Tunnel > 60

Combined

5 0.06 0.03

(a) 0.13 =<60 1 0.00 0.06 (a) 0.13 >60 4 0.00 0.03 (a) 0.06

18-50 Tunnel and Vicinity of the road Combined 5 0.00 0.03 (a) 0.07

>60 1 0.04 0.06 (a) 0.16 18-700 Tunnel Combined 1 0.04 0.06 (a) 0.16

>60 1 0.09

0.06

(a) 0.22

29-1000 Vicinity of the road

Combined

1 0.09 0.06

(a)

0.22 >60

1 0.06

0.06

(a) 0.19

29-250 Vicinity of the road Combined 1 0.06 0.06 (a) 0.19

>60 1 0.07 0.06 (a) 0.20 29-640 Vicinity of the road Combined 1 0.07 0.06 (a) 0.20

=<60 3 0.03 0.04 (a) 0.10 >60 16 0.05 0.02 0.01 0.09

FLEET

ALL - Fleet and Size Ranges Measured

Combined

19 0.04 0.02 0.01 0.07

238

95% Confidence

Interval

Vehicle Type



Sample Size

Average particle

emission factor

value per cubic cm per km

Standard Error

Lower Bound

Upper Bound

=<60 1 0.09 0.06 (a) 0.22 >60 4 0.03 0.03 (a) 0.10


=<60 1 0.93

0.06

0.81

1.06

>60 4 0.21 0.03 0.15 0.27

18-300 Vicinity of the road =<60; Tunnel > 60 Combined 5 0.57 0.03 0.50 0.64

=<60 1 0.01 0.06 (a) 0.14 >60 4 0.01 0.03 (a) 0.07

18-50 Tunnel and Vicinity of the road Combined 5 0.01 0.03 (a) 0.08 18-700 Tunnel

>60 2 0.41

0.04

0.32

0.49

Combined 2 0.41 0.04 0.32 0.49 >60 1 0.41 0.06 0.29 0.54 29-250

Vicinity of the road Combined 1 0.41 0.06 0.29 0.54

>60 1 0.41 0.06 0.29 0.54 30-10000(b) Vicinity of the road Combined 1 0.41 0.06 0.29 0.54

=<60 3 0.35 0.04 0.27 0.42 >60 16 0.25 0.02 0.21 0.28

HDV

ALL - HDV and Size Ranges Measured Combined 19 0.28 0.02 0.24 0.31

=<60 1 0.00 0.06 (a) 0.13 >60 4 0.01 0.03 (a) 0.07


=<60 1 0.03

0.06

(a) 0.15

>60 4 0.03 0.03 (a) 0.09

18-300 Vicinity of the road =<60; Tunnel > 60 Combined 5 0.03 0.03 (a) 0.10

=<60 1 0.00 0.06 (a) 0.13 >60 4 0.00 0.03 (a) 0.06

18-50 Tunnel and Vicinity of the road Combined 5 0.00 0.03 (a) 0.07 18-700 Tunnel

>60 2 0.05

0.04

(a) 0.13

Combined 2 0.05 0.04 (a) 0.13 >60 1 0.03 0.06 (a) 0.16 29-250

Vicinity of the road Combined 1 0.03 0.06 (a) 0.16

>60 1 0.03 0.06 (a) 0.16 30-10000(b) Vicinity of the road Combined 1 0.03 0.06 (a) 0.16

=<60 3 0.01 0.04 (a) 0.08

LDV

>60 16 0.02 0.02

(a) 0.06

ALL - LDV and Size Range Measured Combined 19 0.02 0.02 (a) 0.05

239

95% Confidence

Interval

Vehicle Type



Sample Size

Average particle

emission factor

value per cubic cm per km

Standard Error

Lower Bound

Upper Bound

=<60 3 0.03 0.04 (a) 0.11 >60 12 0.02 0.02 (a) 0.05


=<60 3 0.34 0.04 0.27 0.42 >60 12 0.10 0.02 0.06 0.13

18-300 Vicinity of the road =<60; Tunnel > 60

Combined 15 0.22 0.02 0.18 0.26

=<60 3 0.01 0.04 (a) 0.08 >60 12 0.00 0.02 (a) 0.04

18-50 Tunnel and vicinity of the road Combined 15 0.00 0.02 (a) 0.04

>60 5 0.16 0.03 0.10 0.22 18-700 Tunnel Combined 5 0.16 0.03 0.10 0.22


>60 3 0.17 0.04 0.10 0.24

*ALL

29-250 Vicinity of the road Combined 3 0.17 0.04 0.10 0.24


>60 2 0.22 0.04 0.13 0.31 30-10000(c) Vicinity of the road Combined 2 0.22 0.04 0.13 0.31

=<60 9 0.13 0.02 0.09 0.17 >60 48 0.10 0.01 0.08 0.12

Combined*

Combined** 57 0.11 0.01 0.09 0.13

(a) The lower bound 95% confidence interval value calculated to be negative and

therefore is not valid. (b) This study measured a vehicle fleet that comprised 60%

HDVs. (c) Based on modified population marginal mean.

*All emission factors for Vehicle Types and Size Ranges Measured Combined.

** All emission factors Combined.

240

Table 5.2.3. PM1 model explanatory variables and average PM1 emission factors

Fuel Type and Study Location

Vehicle Type

Sample Size

Average particle emission factor value, per vehicle mg/km

Standard Error


Lower Bound

Upper Bound

Fleet 11 36 17 2 70 HDV 12 289 16 256 321 LDV 11 16 17 (a) 50

Fuel not Specified Vicinity of the road and tunnel

Combined

34

114

22

69

158

HDV 5 285 25 235 336 LDV 5 306 25 255 356

Diesel Dynamometer

Combined 10 622 35 551 693 Fleet 11(b) 36(b) 17 2 70 HDV 17 287 15 257 317 LDV 16 161 15 130 191

*ALL

Combined

44 186(b)

9

168

205

(a) The lower bound 95% confidence interval value calculated to be negative and therefore is not valid.

(b) Based on modified population marginal mean.

* ALL - Diesel and Fuel not Specified.

241

Table 5.2.4. PM2.5 model explanatory variables and average PM2.5 emission factors

Instrumentation & Study Location

Vehicle Type

Sample

size

Average particle emission factor value per vehicle, mg/km

Standard Error


Lower Bound

Upper Bound

HDV – Diesel

5 286 47 192 380

LDV – Diesel

5 306 47 212 401

APS Dynamometer

Combined 10 296 33 229 363 HDV 8 182 37 106 256

LDV 10 23 33 (a) 89

Chemical balance Vicinity of the road Combined 18 102 25 52 152

Bus (c) 5 299 47 205 394 DustTrak Vicinity of the road and tunnel

Fleet 2 15 74 (a) 164

HDV 7 301 40 221 380 LDV 20 33 24 (a) 80 Combined 34 162 24 112 212

Bus (d) 7 42 40 (a) 122 Glass-Fibre Filter Dynamometer Combined 7 42 40 (a) 122

HDV 1 526 106 315 737 Teflon Filters Tunnel LDV 1 7 106 (a) 218 Combined 2 267 74 117 416

Fleet 3 60 61 (a) 182 Samplers Tunnel

HDV 1 135 106 (a) 346 LDV 1 14 106 (a) 225 Combined 5 70 54 (a) 177

Bus – Diesel (e)

2 234 75 84 383 TEOM Tunnel

Fleet 1 49 106 (a) 260 HDV 1 381 106 170 592 LDV 1 19 106 (a) 230 Combined 5 171 49 72 269

Fleet 4 60 53 (a) 166 TEOM & DMPS Tunnel Combined 4 60 53 (a) 166

Bus 14 192(b) 32 127 256 Fleet 10 46(b) 38 (a) 122 HDV 23 302 33 236 367 LDV 38 67 32 3 131

*ALL

Combined 85 156(b) 17 122 191

242


therefore is not valid.


(c) Relate to 300 Diesel Bus trips measured in a tunnel, and Buses, fuel not specified,

tested in the vicinity of the road.

(d) Buses included 3 hybrid buses (2 fitted with catalysed particulate filters); 3 Buses

fuelled with Diesel (fitted with oxidation catalysts) and 1 Bus fuelled with liquified natural

gas.

(e) TEOM equivalent data, where the correlation between TEOM and DustTrak response

to diesel emissions was assessed and the DustTrak results were recalculated into TEOM

equivalent data.

Instrumentation: APS – Aerodynamic Particle Sizer; DMPS – Differential Mobility Particle

Sizer; TEOM – Tapered Element Oscillating Microbalances.

* ALL – Instrumentation and Vehicle Types.

243

Table 5.2.5. PM10 model explanatory variables and average PM10 emission factors

Sample

Size

Average particle

emission factor per

vehicle


Vehicle Type

Road Type

mg/km Standard

Error Lower Bound

Upper Bound

Bus BOULEVARD(c) 2 4130 684 2774 5486 URBAN (d) 6 1089 395 306 1872 Dynamometer

(e) 19

313 222 (a) 753 Combined 27 1844 273 1302 2386 Fleet FREEWAY 1 200 967 (a) 2118 HIGHWAY 2 66 684 (a) 1421 MOTORWAY 2 77 684 (a) 1432 RURAL AREA 1 67 967 (a) 1984 TUNNEL 11 306 291 (a) 884 URBAN 5 688 432 (a) 1546 Combined 22 234 292 (a) 814 HDV BOULEVARD 2 4815 684 3459 6171 FREEWAY 2 2500 684 1144 3856 HIGHWAY 3 840 558 (a) 1947 MOTORWAY 2 213 684 (a) 1568 RURAL AREA 1 394 967 (a) 2312 TUNNEL 6 1019 395 236 1802 URBAN 10 538 306 (a) 1145 Dynamometer 2 259 684 (a) 1615 Combined 28 1322 229 867 1777 LDV BOULEVARD 4 454 483 (a) 1413 FREEWAY 4 285 483 (a) 1244 HIGHWAY 6 141 395 (a) 924 MOTORWAY 2 63 684 (a) 1419 RURAL AREA 1 46 967 (a) 1964 TUNNEL 6 14 395 (a) 797 URBAN 16 156 242 (a) 635 Dynamometer 10 47 306 (a) 653 Combined 49 151 191 (a) 529

BOULEVARD 8 3133(b) 360 2418 3848 FREEWAY 7 995(b) 426 149 1841 HIGHWAY 11 349(b) 322 (a) 988 MOTORWAY 6 117(b) 395 (a) 900 Dynamometer 31 206(b) 260 (a) 723 RURAL AREA 3 169(b) 558 (a) 1276 TUNNEL 23 446(b) 210 30 863 URBAN 37 618 176 269 966

*ALL

Combined 126 749 123 505 993

244


therefore is not valid.


(c) Fuel not specified by the studies (can be assumed to be Diesel-fuelled due to the

timing and location of the study).

(d) Five of the 6 buses were tested on the CBD (Central Business District) Urban bus

Drive Cycle, and 1 bus (Fuel not specified, can be assumed to be Diesel-fuelled due to

the timing and location of the study) was tested on an Urban road. In this analysis CBD

Drive Cycle emission factors were classed as Urban Road Type, due to the scarcity of

studies available that have measured PM10 for buses in on-road measurement

campaigns, and as this Drive Cycle closely emulates urban driving conditions. Of the 5

buses tested on the CBD Drive Cycle - 3 were Diesel-fuelled (1 Low Sulphur Diesel

(LSD) with an oxidation catalyst and 2 Ultralow Sulphur Diesel (ULSD) - one with an

oxidation catalyst and 1 with both an oxidation catalyst and a particle filter) and 2 Diesel

Hybrids (with catalysed particle filters).

(e) These 19 buses emulated a mixed bus fleet. They comprised 9 buses where the fuel

was not specified (these may have been Diesel-fuelled, however the Fuel Types were not

reported); 5 buses were Diesel-fuelled fitted with oxidation catalysts (of these 1 LSD; and

1 ULSD with a particle filter); 2 Diesel Hybrids (with oxidation catalysts; one also had a

catalysed particle filter); and 3 LNG-fuelled with no aftertreatment devices. Although

removal of the 3 LNG bus emission factors would have increased the overall average

emission factor produced by the PM10 statistical model for buses for dynamometer from

313 to 371 mg/km, these LNG emission factors were not removed because for 9 of the

19 buses tested the fuel used was not specified and was unable to be determined.

245

5.2 STATISTICAL RELATIONSHIPS BETWEEN CATEGORICAL

VARIABLES

Figure 5.2.1 presents a multiple comparison plot depicting the statistical

relationships between the average values of published emission factors in terms of

categorical variables examined in the statistical analysis. These related to the

results of post-hoc Scheffe’s multiple comparison statistical tests that investigated

the differences in means between levels corresponding to all categorical variables

(Scheffe 1959), irrespective of whether they had a significant effect on the

response variable (the published emission factor value).

In Figure 5.2.1 variables whose mean values are statistically similar are connected

by joined-lines, and those also annotated with an ‘X’ indicate the variable marked

‘X’ is statistically similar to the variables to which it is joined. Variables without

joined-lines between them have statistically significant relationships at a 95%

confidence level. These results are commented on in the paper presented in

Chapter 5.1.

246

Figure 5.2.1. Multiple comparison plot showing the nature of the statistical

relationship between the categorical model variables for different metrics

Variables connected by a joined-line are statistically similar and those marked X

show the variable marked X and the variables to which it is joined are statistically

similar. Variables without a joined-line are statistically significantly different at a

95% confidence level.

(a) Country of Study

PM1 (a)

Australia Other Countries USA PM2.5 (b)

Australia Other Countries USA, Canada

PM10

Australia Other Countries USA, Canada

Total P mass

Australia Other Countries USA P number

Austria Germany Switzerland UK P volume

(a) Post-hoc multiple comparison tests were not performed for PM1 as there were

fewer than 3 Country of Study groups.

(b) Australian PM2.5 emission factors related mainly to diesel-fuelled vehicles, which

could generally be expected to produce higher values than other fuelled vehicles.

247

(b) Study Location

Vicinity of the Road (a)

Tunnel Dynamometer

PM1 (b)

PM2.5

PM10

P number

Total p mass

P volume (c)

(a) Vicinity of the Road relates to measurements on or near the road (near a curb,

upwind and downwind, downwind only, using vehicle chasing or on-road mobile

laboratories)

(b) The PM1 dynamometer measurements related exclusively to diesel LDVs and

HDVs tested in Australia. Diesel vehicles would generally produce higher emission

factors than other fuelled vehicles.

(c) No dynamometer values were available for particle volume, and there were less

than 3 groups so post-hoc multiple comparison tests were not able to be performed.

248

(c) Dynamometer and Road Types

Motorway Rural area Highway Tunnel Dynamometer Urban Road

PM1

Dynamometer Boulevard Highway Freeway Tunnel Urban Road

PM2.5

Boulevard Highway Motorway Dynamometer Urban Road Tunnel

PM10

PM10

Boulevard

Rural area

Freeway

Dynamometer Boulevard Highway Freeway Tunnel Urban Road Total mass

Dynamometer Highway Freeway Motorway Tunnel Urban Road

Particle number

Highway Motorway Tunnel Urban Road

Particle volume

X

249

(d) Dynamometer and Road Classes

Dyno Speed

Limit on the Road < 80 km/hr

Speed Limit on the Road ≥ 80 km/hr

Dyno Speed Limit on the Road ≤ 60km/hr

Speed Limit on the Road > 60km/hr

PM1

PM2.5

PM10

Total Particle

mass

(a)

(a)

(a)

Particle number

Particle volume

(b) (b)

(a) Total particle mass was not tested as sample sizes were too small. (b) No dynamometer emission factors were available for particle volume.

250

(e) Vehicle Type

Fleet LDV HDV PM1

Bus Fleet LDV HDV PM2.5 (b)

Fleet HDV LDV PM2.5 Bus HDV LDV Fleet P number (a)

Fleet LDV HDV P number

Bus Fleet LDV HDV PM10 (b) PM10

LDV HDV PM10

Bus Fleet LDV HDV Total P mass

Fleet LDV HDV P volume

(a) The statistical similarity between average emission factors for bus and HDV, bus and LDV and

bus and Fleet may be influenced by the fact that bus measurements for particle number related

exclusively to measurements undertaken using a Scanning Mobility Particle Size (SMPS), whereas

the sample of emission factors for HDV, LDV and Fleet included measurements undertaken using a

Condensation Particle Counter (CPC). The CPC measures the nucleation mode (where particle

number tend to be very prolific), and the SMPS does not, which may result in lower value

measurements. (b) In terms of particle mass, PM1 is considered more relevant for buses than

PM2.5 and PM10.

X

X

X

X

251

(f) Fuel Type

PM1 (a)

PM2.5 (a)

Fuel not Specified

Diesel Petrol

PM10

Diesel Petrol PM10 Particle number

Fuel not Specified

Diesel Petrol

CNG Diesel Petrol LNG ULSD LSD Total particle mass

P volume (b)

(a) Post-hoc multiple comparison tests were not performed for PM1 and PM2.5 as

there were fewer than 3 groups.

(b) Fuel Types were not reported for particle volume.

Fuel Types: CNG – Compressed Natural Gas, LNG – Liquified Natural Gas, ULSD –

Ultralow Sulphur Diesel, LSD – Low Sulphur Diesel.

X

252

(g) Instrumentation

Beta-ray Betameter Kleinfiltergerate APS

PM1

APS Telfon filters

Chemical balance

DustTrak Glass fibre filter

Sampler TEOM TEOM, DMPS

PM2.5

Chemical balance

Filters Impactor MOUDI, ELPI, SMPS

Remote sensing

SMPS SMPS & Others

Total mass

CPC SMPS APS (a) CPC, DMPS DMA ELPI EAA ELPI & UCPC

Particle number

SMPS DMPS

Particle number

Particle volume (b)

PM10 (c)

253

(a) The small sample size (n=7) and very small mean value for APS in particle number of

0.002 x 1014 particles per vehicle per kilometre hampered a meaningful comparison with

mean values for CPC and SMPS, which both had substantially larger sample sizes and

larger mean values of CPC 22.69 x 1014 particles per vehicle per kilometre (n=18) and

SMPS 2.083 x 1014 particles per vehicle per kilometre (n=96). The APS also measures a

vastly different particle size range to CPC and SMPS, as shown in Table 5.2.1.

b) Post-hoc multiple comparison tests were not performed as there were fewer than 3

groups.

(c) PM10 Instrumentation sample sizes were too small to test for significant differences

between the means.

Instrumentation - APS Aerodynamic Particle Sizer; CPC - Condensation Particle Counter;

DMA - Differential Mobility, DMPS – Differential Mobility Particle Sizer; EAA – Electrical

Aerosol Analyser ; ELPI – Electrical Low Pressure Impactor ; SMPS – Scanning Mobility

Particle Sizer. TEOM – Tapered Element Oscillating Microbalances.

254

5.3. ADDITIONAL COMMENTS RELATED TO PARTICLE VOLUME

AND PM10 EMISSION FACTORS

The majority of emission factors used to develop the urban South-East

Queensland (SEQ) inventory were sourced from those identified as the most

suitable to use in transport modelling and health impact assessments presented in

this Chapter.

Particle volume emission factors

It should be noted that average emission factors for particle volume were not used

in developing the urban SEQ inventory. This decision was taken because a direct

correlation exists between particle mass and volume distributions, where density

acts as a scaling factor (Morawska et al. 2008), and it was considered that a more

detailed understanding of particle mass emission rates could be obtained by using

average emission factors for subsets of mass fractions for PM1, PM2.5 and PM10.

In addition, these mass fractions have greater relevance to current mass-based

ambient air quality standards.

Possible influence of road dust in LDV and HDV PM10 emission factor

measurements

Considerable variation can be seen between the values of average emission

factors produced by the statistical models for PM10 for LDVs, HDVs and buses

for different Road Types, as compared to tunnel and dynamometer average

emission factors (Table 5.2.5).

255

For example, the LDV statistical model (Table 5.2.5) had a sample size of 49

emission factors and produced average emission factors that ranged from 14

mg/km for tunnel (n=6) to 47 mg/km (n=10) for dynamometer studies; and for

different Road Types ranged from 46 mg/km for a rural area road (n=1) to 63

mg/km for motorway (n=2); 141 mg/km (n=6) for highway; 156 mg/km (n=16)

for urban road; 285 mg/km (n=4) for freeway; and 454 mg/km (n=4) for

boulevard. Even greater variation can be seen between PM10 average emission

factors for HDVs for tunnel, dynamometer and studies of different Road Types

(Table 5.2.5). Differences in average bus emission factors are discussed in

Section 5.5. below.

The PM10 statistical model showed the lowest correlation coefficient, as compared

to the statistical models developed for the other particle metrics, of 0.47 (see

Chapter 5.1). This low correlation coefficient may be influenced by the presence

of varying quantities of road dust occurring at the PM10 size range found near

roads or in tunnels that were stirred up and resuspended by vehicle traffic. These

levels of road dust can vary depending on the road surface and climatic

conditions, as well as the influence of other factors such as traffic volumes,

vehicle speed, and congested driving conditions, where there may be more

stopping and starting, such as on urban roads. Few methods are presently

available for discriminating resuspended road dust from tailpipe vehicle particle

emissions; and as LDVs usually make up the major proportion of vehicles

travelling in on-road fleets, the presence of road dust is likely to more heavily

influence the LDV emission factor derived from measured fleet emissions, than

256

that for HDVs, as these usually only comprise a small proportion of the fleet in

terms of their numbers.

5.4. ADDITIONAL COMMENTS ON PM10 EMISSION FACTORS

USED IN THE URBAN SEQ INVENTORY

The basis for selection of the average PM10 emission factors to use in developing

the urban SEQ inventory, from those produced by the statistical models, is

generally discussed below.

LDV and HDV emission factors for PM10

The average emission factors for LDVs and HDVs for PM10 selected from

the statistical model outputs presented in this Chapter to use in developing

the urban SEQ inventory were those for urban and highway Road Types

(Table 5.1.4, Chapter 5.1).

These average emission factors generally had lower standard errors and

lower upper bound 95% confidence intervals values than the other average

emission factors, and the speed limits on these Road Types closely matched

the two road classifications used in the SEQ inventory. Urban road

emission factors related to roads with speed limits of 50 and 57 km/hr; and

for highway to roads with speed limits of 82 and 100 km/hr. The two road

classifications in the SEQ inventory related to roads with average vehicle

speeds of < 80 km/hr for urban roads and ≥ 80 km/hr for urban-major roads.

Although speed limits on the road and average vehicle speeds are not

257

directly comparable, nevertheless due to the lack of available data on

average vehicle speeds reported in the studies, this is the only comparison

able to be made.

PM10 Bus emission factors

Although the explanatory variables for the PM10 statistical model were found to be

Vehicle Type and Road Type, the decision was taken to use the average emission

factor produced by the statistical model derived from dynamometer measurements

in development of the urban SEQ inventory (presented in Chapter 6).

The statistical model for PM10 for buses was based on emission factor data derived

from measurements on boulevard (n=2) and urban Road Types (n=6) and from

dynamometer measurements (n=19) (Table 5.2.5).

The authors of the boulevard Road Type study (from which two emission factors

for buses were sourced) reported that they considered their very high values of

PM10 emission factors were influenced by contributions from resuspended road

dust and, within each vehicle category, by the effects of speed and acceleration

(Abu-Allaban et al. 2003). This same study derived one bus emission factor for

PM10 from measurements on an urban Road Type. The three PM10 bus emission

factors derived in this study were measured using DustTrak Instrumentation,

which the authors stated may have overestimated measurements (Abu-Allaban et

al. 2003).

258

The remaining five emission factors in the Urban Road Type sample related to

dynamometer measurements of buses on the Central Business District (CBD)

Drive Cycle, which very closely emulates urban driving conditions, and were

included due to the extreme lack of available on-road data for buses travelling on

urban Road Types. Hence it was decided to use the more conservative average

emission factor for PM10 dynamometer bus measurements, for all Road Types in

the urban South-East Queensland inventory, as this average emission factor was

less likely to be affected by excessively high rates of resuspended road dust, and

as this sample of emission factors included a very wide range of different urban

bus Drive Cycles and had the largest sample size.

259

5.5. REFERENCES

Abu-Allaban, M., Gillies, J.A., Gertler, A.W., 2003. Application of a multi-

lag regression approach to determine on-road PM10 and PM2.5 emission

rates. Atmospheric Environment 37(37), 5157-5164.

Morawska, L., Keogh, D. U., Thomas, S. B., Mengersen, K., 2008.

Modality in ambient particle size distributions and its potential as a basis

for developing air quality regulation. Atmospheric Environment 42(7),

1617-1628.

Scheffe, H., 1959. The Analysis of Variance, John Wiley & Sons, Inc.

260

CHAPTER 6

DEVELOPMENT OF A PARTICLE NUMBER AND

PARTICLE MASS EMISSIONS INVENTORY FOR AN

URBAN FLEET

Diane U. Keogh1, Luis Ferreira2, Lidia Morawska1


University of Technology, Gardens Point, Brisbane, Australia

2 School of Urban Development, Queensland University of

Technology, Gardens Point, Brisbane, Australia

Environmental Modelling and Software 24(11), 1323-1331.

261


Title: Development of a particle number and particle mass

emissions inventory for an urban fleet

Authors: Diane U. Keogh, Luis Ferreira, and Lidia Morawska


Developed the experimental design and scientific method. Developed the

inventory of emissions and scenario models. Carried out all the calculations and

analysis for these models. Data interpretation. Wrote the majority of the

manuscript.

Luis Ferreira

Reviewed the manuscript.

Lidia Morawska


262

ABSTRACT

Motor vehicles are major emitters of gaseous and particulate matter pollution in

urban areas, and exposure to particulate matter pollution can have serious health

effects, ranging from respiratory and cardiovascular disease to mortality. Motor

vehicle tailpipe particle emissions span a broad size range from 0.003-10µm, and

are measured as different subsets of particle mass concentrations or particle

number count. However, no comprehensive inventories currently exist in the

international published literature covering this wide size range.

This paper presents the first published comprehensive inventory of motor vehicle

tailpipe particle emissions covering the full size range of particles emitted. The

inventory was developed for urban South-East Queensland by combining two

techniques from distinctly different disciplines, from aerosol science and transport

modelling. A comprehensive set of particle emission factors were combined with

transport modelling, and tailpipe particle emissions were quantified for particle

number (ultrafine particles), PM1, PM2.5 and PM10 for light and heavy duty

vehicles and buses. A second aim of the paper involved using the data derived in

this inventory for scenario analyses, to model the particle emission implications of

different proportions of passengers travelling in light duty vehicles and buses in

the study region, and to derive an estimate of fleet particle emissions in 2026.

It was found that heavy duty vehicles (HDVs) in the study region were major

emitters of particulate matter pollution, and although they contributed only

around 6% of total regional vehicle kilometres travelled, they contributed more

than 50% of the region’s particle number (ultrafine particles) and PM1

263

emissions. With the freight task in the region predicted to double over the next

20 years, this suggests that HDVs need to be a major focus of mitigation efforts.

HDVs dominated particle number (ultrafine particles) and PM1 emissions; and

LDV PM2.5 and PM10 emissions. Buses contributed approximately 1-2% of

regional particle emissions.

Keywords: Motor vehicle inventory, emission factors, traffic modelling,

particle number, particle mass, ultrafine particles.

264

6.1. INTRODUCTION

Tailpipe particle emissions generated by motor vehicles span a very wide size

range, from around 0.003µm (the current detection limit on scientific measuring

equipment) to 10µm. Only one inventory exists which attempted to estimate all

the particle metrics and subclasses, and this was developed for the UK (Group

1999). However this inventory was restricted to estimating particle emissions for

the smaller particle size ranges by applying distribution profiles for these size

ranges to PM10 estimate data (Group 1999; Goodwin et al. 2000; AQEG 2005).

This means that emission factors for these size ranges below PM10 were not based

on individual measurements of different particle sizes, but simply on mass

fractions multiplied by PM10 data values.

Current worldwide air quality standards for controlling particulate matter

pollution are mass-based and restricted to PM2.5 and PM10 (mass concentration of

particles with aerodynamic diameters < 2.5 µm and 10 µm respectively).

However, these standards are ineffective for controlling ultrafine particles

(diameters < 0.1 µm), which are very numerous in terms of their numbers, but

have little mass (weight). Most particle emissions generated by motor vehicle

tailpipes are ultrafine size and are measured in terms of particle number; hence it

is critical that future inventories include estimates for particle number emissions.

As the majority of particles measured in terms of particle number are in the

ultrafine size range, the terms particle number and ultrafine particles

265

will be used interchangeably in this paper. From this point on in the paper

tailpipe motor vehicle particle emissions will be referred to as motor vehicle

particle emissions.

In urban areas motor vehicle particle emissions are a dominant pollution source,

where more than 80% of particle number concentrations are found in the ultrafine

size range (Morawska and Salthammer 2003). However, very little information

can be obtained about particle number from particle mass measurements (ECJRC

2002), and as current air quality standards are mass and not particle number-

based, this means that the greater proportion of motor vehicle particle emissions

are not controlled or regulated.

There are many different types of emission models that have been developed to

model particle emissions generated by motor vehicle fleets. These have employed

a wide range of different methods and related to varying geographic scales. The

application of these models include developing an understanding of air quality

and climate change issues on global, regional and local scales (Parrish 2006) and

for developing control strategies, risk assessments, air quality forecasting and

transport and economic incentive programs (Mobley and Cadle 2004).

Examples of emission models which have estimated vehicle fleet emissions

include well-known models such as EMFAC (CARB 2001) and MOBILE

(USEPA 1993); as well as BRUTAL (Oxley et al. 2009), CALPUFF (Cohen et al.

2005), MM5-ARPS-CMAQ (Cheng et al. 2007), OSCAR (Sokhi et al. 2008),

TAPM (Hurley et al. 2005), TEMMS (Namdeo et al. 2002) and VERSIT+ LD

(Smit et al. 2007), to name a few. However these emission models are generally

266

limited to providing estimates for PM10 or, in some cases, PM2.5. One model,

however, COPERT IV, has available a small suite of solid particle number

emission factors for different vehicle types derived from dynamometer

measurements (Samaras et al. 2005).

Emission models can be developed with a specific focus, eg., such as for

modelling the effects of congestion (Smit et al. 2008), estimating the spatial and

temporal resolution of emissions (Constabile et al. 2008), for modelling street

canyons (Mensink et al. 2006) or modelling the influence of road classifications

on personal exposure to emissions (Chen et al. 2008). A number of emission

models often use indirect data, such as total fuel consumption data or based on

fuel properties (Goodwin et al. 1999) or remotely sensed data (Shifter et al. 2005),

rather than using performance-related emission factors and road traffic data.

A number of epidemiological studies have linked particle exposure with increases

in hospital admissions, various respiratory and cardiovascular diseases and

mortality (Pope and Dockery 2006); and current scientific debate is focused on the

premise that particle number (ultrafine particles) is more directly related to health

effects than particle mass (ECJRC 2002). In relation to the health effects due to

exposure to ultrafine particles, of which the majority are in the nanosize range

(diameters < 0.05 µm), the World Health Organization (WHO 2005) has stated

that “While there is considerable toxicological evidence of potential detrimental

effects of ultrafine particles on human health, the existing body of epidemiological

evidence is insufficient to conclude on exposure/response relationship to ultrafine

particles. Therefore no recommendations can be provided as to guideline

concentrations of these particles at this point.”

267

Whilst recommendations as to guideline concentrations for ultrafine

particles cannot presently be determined, nevertheless the importance and

high priority given to controlling these sized particles is evidenced by the

fact that particle number limits for solid particles are being introduced by the

European Union for light duty diesel vehicles in EURO V/VI and for heavy

duty diesel vehicles in EURO VI (European Union 2007; Commission of the

European Communities 2007 a,b); and are proposed for light duty diesel

vehicles in Switzerland (AQEG 2005). The particle number inventory

presented in this study for urban South-East Queensland is the first of its

kind, and to our knowledge no such extensive inventory is available. No

detailed emission inventories are available that include particle number

concentration (Jones and Harrison 2006).

Particle emissions can be reduced in a variety of ways, ranging from fitting

particle traps or introducing new vehicle standards (eg., EUROs), to policies

such as congestion charging and incentives for scrapping older vehicles.

Higher density living and transit oriented development is causing the public

to become more concerned about particulate matter pollution, and they are

demanding greater quantification of particle emission levels. Particle

emission inventories are an important tool for understanding current levels

and controlling emissions, as well as for testing the air quality implications

of future alternative transport and land use strategies.

268

This paper presents the first published comprehensive inventory of motor

vehicle tailpipe particle emissions. It was developed for urban South-East

Queensland by combining a comprehensive set of emission factors with

transport modelling to produce road-link based estimates of particle number

(ultrafine particles), PM1, PM2.5 and PM10 for light and heavy duty vehicles

and buses for different road types. Future scenarios were tested involving

moving proportions of LDV trips to new buses to quantify the impact on

emission levels. Modelling bus trips was important as major busways and

tunnels are under construction in the study region to address urban sprawl,

increased travel demand and congestion. An estimate of fleet emissions in

2026 was also derived.

It is important to note that inventory estimates are based on government

prototype data from the Brisbane Strategic Transport Model for 2004

(Queensland Department of Main Roads 2008), and utilised vehicle

kilometres travelled (VKT) data, but excluded consideration of specific origin

and destination trip data.

6.2. METHOD

Particle number (ultrafine particles), PM1, PM2.5 and PM10 inventories for the

motor vehicle fleet in urban South-East Queensland for 2004 were developed

and different scenario analyses were modelled relating to travel behaviour and

mode choice. Details related to the study region, the emissions inventory

269

model components (both for the travel demand and emission factor models), and

variables used in scenario analyses conducted are outlined below.

The inventory was calculated by combining two techniques from distinctly

separate disciplines – from aerosol science and transport modelling. Computation

of the emissions inventory involved compilation of relevant emission factors for

different vehicle and road type combinations for different particle metrics with

individual travel demand model links. The method applied appropriate particle

emission factors for different vehicle and road type combinations, which were

derived from statistical analysis of a large body of published emission factor data,

as well as identifying a small number of relevant local bus emission factors.

Travel data for the region was sourced from a Government household travel

survey (SEQHTS 2004) and freight matrices, this data was analysed in terms of

trip mode and trip purpose and assigned to different travel demand model links

that represented roads in the study region, and the average fleet speed on each

model link was estimated (Queensland Department of Main Roads 2008). Model

links were classified into different road types based on the average speed of the

fleet on different model links. Inventories were calculated for different particle

metrics based on average weekday travel data scaled up to annual estimates. The

scaling factor used was based on the difference found between weekday and

weekend VKT in a local household travel survey conducted over an entire week

(Ministry of Transport 2007).

270

Average passenger occupancy rates for buses in different travel times were

derived from an analysis of all bus timetable and occupancy rate data relating to

the study region. Raw data for this analysis was provided by Translink (2007).

Scenario models were developed based on assumptions related to varying shifts in

travel modes in different travel times, and corresponding average vehicle

occupancy rates, and on anticipated future levels of particle emissions. The

inventory was validated based on the results of an extensive worldwide literature

review.

6.2.1. Study region

This was the Brisbane Statistical Region located in South-East Queensland

(SEQ), Australia (hereafter termed urban SEQ), which had 1.2 million motor

vehicles and a resident population in 2004 of 1.7 million (ABS 2004 a,b).

Although urban SEQ only makes up around 26% of the area in SEQ (ABS 2006;

OESR 2005), the urban SEQ vehicle fleet accounted for more than 70% of private

passenger trips in SEQ in 2004 (SEQHTS 2004). One million people are

predicted to move to SEQ in the next 20 years (Office of Urban Management

2004) and the Bureau of Transport and Regional Economics have forecast that the

freight task will double over this period (SKM 2206).

271

6.2.2. Transport model

Traffic data from the Brisbane Strategic Transport Model (BSTM) was used to

derive the inventory, and this model covers an area of around 4600 square

kilometres (Queensland Department of Main Roads 2008). The latest version of

this model is in the prototype stage and does not represent current Government

policy. It is a conventional 4-step demand model, incorporating trip generation,

distribution, modal split and assignment (Ortuzar and Willumsen 2001). It covers

urban SEQ and is populated with data from a Queensland Government household

travel survey (SEQHTS 2004). The model contains VKT data for a typical

weekday in 2004 in four travel times for light and heavy duty vehicles and buses.

Roads in urban SEQ are represented by 22,985 individual model links.

VKT on each model link: was calculated by multiplying the number of vehicles

in each vehicle class by the length of the model link in kilometres. Total weekday

VKT was 45.5 million km (93.3% from LDVs, 6.3% from HDVs and 0.4% from

buses). LDVs were classed as passenger cars and trucks with vehicle weights ≤ 5

tonnes; and HDVs vehicles had gross vehicle weights ranging from 3.5-12 tonnes

to > 25 tonnes (Keogh et al. 2009). A slight overlap in the weight ranges of LDV

and HDV vehicle classes occurred due to the nature of the data reported by the

authors of emission factor studies (Keogh et al. 2009). At the time of this study

the majority of LDV vehicles were petrol-fuelled and HDVs diesel-fuelled

(Keogh et al. 2009).

272

Total annual VKT for the 2004 BSTM was 14,514 million, which excluded

transport-related industry travel (eg., couriers and taxis) and trips by persons

staying in non-private accommodation (eg., tourists and business travellers

staying in hotels) (SEQHTS 2004). The BSTM VKT is within 20% of the VKT

estimate for the region derived in the 2004 Survey of Motor Vehicle Use of

18,331 million (ABS 2004a), which would be considered reasonable accuracy for

a strategic model. Other VKT estimates used to model the study region in 2004

ranged from 16,340-21,017 million (BTRE 2003; Apelbaum 2006).

Road types: Model links were classed as urban or urban-major roads, based on

the average vehicle speed travelled on the links in the different travel times.

Urban roads had average vehicle speeds of < 80 km/hr and urban-major ≥ 80

km/hr. Different vehicle speeds can be associated with different levels of

particulate matter emissions for different vehicle types. Hence the model links

representing roads in the study region were classified according to average

vehicle speed and the technique used to develop the inventory applied emission

factors relevant to those average vehicle speeds.

Travel times: VKT related to four travel times - 7am-9am, 9am-4pm, 4pm-6pm,

and 6pm-7am. The 24 hour average VKT was the sum of VKT in the four travel

times. Emissions were calculated for each travel time, and summed for each

model link.

273

Average vehicle speed: on each individual model link was available for the four

travel times. The 24 hour average vehicle speed was the length of time based

average of the four time periods, that is, 2*7am-9am speed + 7*9am-4pm speed +

2*4pm-6pm speed + 13*6pm-7am speed)/24 hours, weighted by kilometres

travelled for each vehicle class in the BSTM model (LDV, HDV and Bus).

Annual VKT: Model VKT represented a typical weekday and was converted to

annual VKT by multiplying by 319 days based on the average difference found

between weekday and weekend VKT in a household travel survey. Inventories

were calculated for different particle metrics based on average weekday travel

data scaled up to annual estimates. The scaling factor used was based on the

difference found between weekday and weekend VKT in a local household travel

survey conducted over a full seven-day week (Ministry of Transport 2007).

Calculation of the inventory: The calculation for the total particle inventory for

urban SEQ for each particle metric consisted of the sum of:-

Total particle emissions on each model link = (EFLDV * VKTLDV) + (EFHDV * VKTHDV) + (EFCNG buses * VKTCNG buses ) + (EF Diesel buses* VKT Diesel buses) Daily total on each model link = emissions 7-9am + emissions 9-4pm + emissions 4-

6pm, + emissions 6pm-7am Scaled up to per annum * 319 days where EF relates to the emission factor relevant to the model link in terms of its road type classification (urban or urban major road type)

274

6.2.3. Emission factors

Emission factors used to derive the inventory are shown in Table 6.1, sourced

from two studies (Keogh et al. 2009; Jayaratne et al. 2008).

Particle metrics: Emission factors related to particle number (number

concentration of particles with diameters 0.003-1µm); and PM1, PM2.5 and PM10.

Instrumentation measuring particle number does not usually measure particles

with diameters greater than 1µm, and the majority of these particles are ultrafine

particles, < 0.1 µm in diameter.

LDV and HDV emission factors: were sourced from a study that derived

emission factors based on a statistical analysis of 667 emission factors published

in the international literature derived from measurement studies. The derived

emission factors had 95% confidence interval values associated with them,

providing the range within which there is a 95% probability that the true value

will lie (Keogh et al. 2009). The statistical models developed to derive these

emission factors were found to explain 86%, 93%, 87%, 65% and 47% of the

variation in published emission factors (Keogh et al. 2009).

Bus emission factors: In 2004 the bus fleet in the study region comprised 89%

Diesel and 11% CNG buses (Translink 2007). Diesel bus emission factors for

particle number on urban roads and for PM10 for urban and urban-major roads

were sourced from Keogh et al. (2009). The remaining bus emission factors were

sourced from a chassis dynamometer study in Brisbane that tested six Scania

275

Table 6.1 Tailpipe particle emission factors for motor vehicles used to develop particle number, PM1, PM2.5 and

PM10 inventories presented in this study

Particle metric

Road Type

HDV emission

factor c h

LDV emission

factor c

Diesel Bus

emission factor

CNG Bus

emission factor


65 [60.19-69.81]

3.63 [9.85 a]

--

--

Particle number

All Road Types Urban Road e Urban-major Road e

-- --

-- --

3.08 [9.30 a] 1.80 d g

9.75 d f 0.22 d g

mg per km

PM1 All Road Types 287 [257-317] 16 [50 a] b b PM2.5 All Road Types 302 [236-367] 33 [80 a] 299 [205-394] b PM10 All Road Types 313 [753 a]

Urban Road e 538 [1145 a] 156 [635 a] 1.10 d f Urban-major Road e

840 [1947 a] 141 [924 a] 0.05 d g

a In the statistical analysis used to derive the emission factor values, the lower bound 95% confidence interval value calculated to be negative, and although physically uninterpretable, can be obtained as a consequence of the normal assumptions underlying the models (Keogh et al. 2009). b Relevant emission factors were not available for this class of vehicle. c LDVs had vehicle weights ≤ 5 tonnes and HDVs gross vehicle weights ≥ 3.5 tonnes. d Lower and upper bound 95% confidence interval values were not available for these emission factors. e Urban roads had average vehicle speeds of < 80 km/hr and urban-major ≥ 80 km/hr. f Based on DT80 transient drive cycle test (Jayaratne et all. 2008). g Based on 50% engine load test (Jayaratne et al. 2008).

276

CNG buses and 5 new generation Mercedes OC500 Diesel Buses (Jayaratne et al.

2008). Their emission factors derived from the DT-80 transient drive cycle test

were selected to represent urban road emissions and those from the 50% engine

load test to represent urban-major road emissions. No relevant bus emission

factors are currently available in the literature for PM1, or for PM2.5 for CNG

buses (Keogh et al. 2009).

6.2.4. Variables used in the scenario analyses

The scenarios modelled in this paper investigate aspects of the transport system

relating to travel mode choice and vehicle occupancy rates (private passenger car

versus public transport bus travel), as well as the effect of growing urbanisation in

a metropolitan city and predicted doubling of the on-road freight task.

They illustrate not only the particulate matter emission rate associated with each

passenger-km travelled for different travel modes, but also highlight the extent of

shift in travel behaviour needed in a passenger car-dependent environment to

effect reasonable reductions in particulate matter pollution.

These scenarios reflect current trends around the world and the aims of many

government initiatives which seek to deal with the problems of congestion and

growing urbanisation, by encouraging shifts from private passenger car travel to

public transport, such as buses, and by increasing vehicle occupancy rates.

277

The rationales for scenarios presented in this paper were based on local

government initiatives, policies and strategic plans developed for the region,

which were the obvious choice for scenario analyses. To encourage shifts from

private passenger car travel to buses, the Brisbane City Council boosted public

transport spending by 76% during 2003-2006, and since 2001 has added an

additional 330 new business to the Brisbane fleet (BCC 2007). The relevance of

modelling emissions in terms of passenger-km trips for different travel modes and

their effect on emission rates was informed by the TravelSmart™ initiative, which

had been introduced in the study region in 2003-2007 to encourage sustainable

transport choices, including car pooling, use of public transport, walking and

cycling, and reductions in single-occupancy car usage (McKay & McGaw 2005).

Mode of travel, trip purpose and average LDV occupancy data were sourced from

the same household travel survey used to populate the BSTM model (SEQHTS

2004). In order to derive the average bus occupancy rates for bus travel in the

region the authors analysed data provided by Translink (2007) relating to bus

timetable and occupancy rate information for all buses operating in the study

region, including Brisbane City Council and privately owned and operated bus

fleets. From this analysis, the average passenger occupancy rates for buses in the

region were derived for the four different travel times.

Trip Mode: categories included Vehicle Driver, Vehicle Passenger, Public

Transport, Walk/Cycle and Other (non-road based, eg., air and rail travel).

278

Trip purpose: categories included trips to and from home to work, education,

shopping, other, and trips to and from work which did not begin or end at home.

Home Based Work trips: were trips between home and work. Only Vehicle

Driver trip data was used in scenario analyses, and these related to vehicle

numbers. This avoided possible double counting of trips which had both a driver

and passenger in the same vehicle. These trips accounted for 46% of total VKT in

7am-9am, 50% in 4pm-6pm and 41% for the 24 hour average (SEQHTS 2004).

Average VKT: was calculated for LDVs and buses for different time periods and

used to compare emission scenarios.

Average vehicle occupancy: was 1.6 passengers for LDVs (7am-9am); and 1.5

passengers (9am-4pm, 4pm-6pm and for the 24 hour average) (Translink 2007).

Average bus occupancy rates calculated to be 18.3, 13.4, 16.9, 13.3 and 15.5

passengers for 7am-9am, 9am-4pm, 4pm-6pm, 6pm-7pm and 24 hour average

respectively. As specific rates were not available for the different fuel types, the

same occupancy bus rate was used for Diesel and CNG buses.

New bus VKT: was calculated for future scenarios that modelled movement of

LDV passengers to new buses. The number of new buses required was calculated

based on LDV and bus occupancy rates in the different travel times. New bus

VKT was assigned as 40% to Diesel buses and 60% to CNG buses. For the

scenario modelling involving the movement of different proportions of LDV

passengers to new buses, we made the assumption that these additional new buses

279

would change the composition of the bus fleet to 40% CNG-fuelled buses and

60% diesel-fuelled buses. This assumption was based on the purchase by

Brisbane City Council of 300 additional new CNG buses for their vehicle fleet

(Brisbane City Council, 2007).

Average particle emission factors per passenger per km: These were derived by

dividing LDV and bus emission factors used to develop the urban SEQ inventory

(Table 6.1) by the respective average vehicle occupancy rates in the 24 hour

average period.

6.3. RESULTS AND DISCUSSION

This section presents the total inventory for urban SEQ for 2004; compares its

PM10 inventory with three other model estimates; compares emission factors

from a UK inventory with the urban SEQ inventory; and discusses scenario

analyses results.

6.3.1. Particle inventory for urban SEQ for 2004

Table 6.2 presents the estimated particle inventory for urban SEQ for 2004.

Most model links were classed as urban roads; hence urban road emissions

dominated the inventory. The inventory was also influenced by very high LDV

VKT which was 93% of total regional VKT. The conclusions below are based

on data in Table 6.2.

280

On lower speed roads (urban roads) total LDV emissions were high for PM2.5 and

PM10 influenced by the fact that total LDV VKT is almost double that of total

HDV VKT on these roads. However, on higher speed roads (urban-major roads)

total HDV emissions dominated particle number and PM1 influenced by the

higher value HDV emission factor, which was almost 6 times higher than the

emission factor for LDVs, and the fact that total HDV VKT was slightly higher

than total LDV VKT on this Road Type.

On urban roads total LDV and total HDV particle number and PM1 emissions

were found to be similar; total LDV emissions for PM2.5 were more than double

total HDV emissions; and for PM10 were more than 5 times total HDV emissions.

On urban-major roads total HDV emissions for particle number and PM1 were

almost double total LDV emissions; PM2.5 total emissions for LDV and HDV

were similar; and PM10 total LDV emissions were more than 2.75 times total

HDV emissions.

Most HDVs were diesel-fuelled and their large contribution to the smaller

particle size range is of considerable concern from the health effects

perspective. In Switzerland these diesel particle emissions are classified as

carcinogenic (Dieselnet 2008).

Total particle number emissions for CNG buses on urban roads were more than 4

times total Diesel bus emissions; however on urban-major roads total Diesel bus

emissions were almost 10 times those for CNG buses; and were 3 orders of

magnitude higher than total CNG buses for PM10. Cadle et al. (2008) also found

281

CNG buses had higher particle number and lower particle mass emissions than

diesel and diesel-hybrid bus emissions. Although bus emission contributions to

the region were small (only around 1-2%) their quantification at the local scale is

important. This is due to high localised exposure to particulate matter emissions at

busways and in tunnels.

Particle number inventory: Total LDV and HDV emissions were similar; and

total bus emissions around 2-3 orders of magnitude lower. HDVs contributed

only around 6% of VKT, but contributed 54% of total particle number emissions

in the region. Freight is predicted to double over the next 20 years and therefore

HDVs are a pollution source that requires urgent attention. Total bus emissions

were less than 1%. Annual total particle number emissions were 3.40 (1.71-6.18)

x 1022 per day or 1.08 (0.54-1.97) x 1025 per annum. No studies were found which

can be compared to the urban SEQ particle number inventory. The only study

available estimated particle flux from all sources, from both natural and

anthropogenic sources (Bigg and Turvey 1978), and as data is not available on

particle flux from natural sources in urban SEQ, the studies can not be compared.

282

PM1 inventory: Relevant PM1 emission factors were not available for Diesel or

CNG buses, hence were not included in the inventory. The annual estimate for

PM1 emissions was 477 (233-964) tonne, and total HDVs contributed 55% of this

particulate matter pollution. More studies are needed to measure the PM1 mass

size range; as it has been shown that PM1 and PM10 are likely to be a more

relevant and discerning combination of air quality standards than the current

standards of PM2.5 and PM10 for combustion sources such as motor vehicles

(Morawska et al. 2008).

PM2.5 inventory: No relevant emission factors were available for CNG buses,

therefore these were not included. Total LDVs emitted 61% of PM2.5 as compared

to 37% from total HDVs and about 2% from total Diesel buses. The annual

estimate for PM2.5 was 736 (225-1436) tonne.

PM10 inventory: Total LDVs contributed 81% and total HDVs 18%. Buses

contributed less than 1%. The annual PM10 estimate was 2614 tonne, with an

upper 95% confidence interval value of 9668 tonne. This high upper confidence

interval value indicates lower certainty in the upper bound value estimate.

283

Table 6.2 Particle emission inventories for the urban South-East Queensland motor vehicle fleet for particle number, PM1, PM2.5 and PM10 on urban and urban-major roads. Lower and upper 95% confidence interval bound values are shown in parentheses.

PARTICLE METRIC & ROAD TYPE d

ESTIMATED PARTICLE EMISSIONS PER DAY

ESTIMATED PARTICLE

EMISSIONS PER ANNUM e Particle number

Light Duty Vehicles

1022

Heavy Duty Vehicles 1022

CNG Buses 1019

Diesel Buses 1020

All Vehicles 1022

All Vehicles 1025

Urban roads

1.13 [3.07 b]

1.08 [1.00-1.16]

1.76 c

0.4 [1.31 b]

2.22 [1.00-4.24]

0.70 [0.31-1.35]

Urban-major roads 0.41 [1.11b] 0.77 [0.71-0.83] 0.006 c 0.04 c 1.18 [0.71-1.94] 0.38 [0.23-0.62] Total Particle Number 1.54 [4.18 b] 1.85 [1.71-1.99] 1.76 c 0.44 [1.31 b] 3.40 [1.71-6.18] 1.08 [0.54-1.97] Particle mass Light Duty

Vehicles kg Heavy Duty Vehicles

kg CNG Buses

kg Diesel Buses

kg All Vehicles

kg All Vehicles

Tonne PM1 Urban roads 498 [1556 b] 478 [428-527] a a 976 [428-2083] 311 [136-664] Urban-major roads 181 [565 b] 340 [304-375] a a 521 [304-940] 166 [97-300] Total PM1 679 [2121 b] 818 [732-902] a a 1497 [732-3023] 477 [233-964] PM2.5 Urban roads 1027 [2490 b] 502 [393-610] a 42 [29-56] 1571 [422-3156] 501 [134-1006] Urban-major roads 373 [904 b] 358 [279-435] a 7 [5-9] 738 [284-1348] 235 [91-430] Total PM2.5 1400 [3394 b] 860 [672-1045] a 49 [34-65] 2309 [706-4504] 736 [225-1436] PM10 Urban roads 4855 [19765 b] 895 [1905 b] 0.02 c 44 [109 b] 5794 [21779 b] 1847 [6943 b] Urban-major roads 1763 [7176 b] 637 [1356 b] 0.0001 c 7 [15 b] 2407 [8547 b] 767 [2725 b] Total PM10 6618 [26941 b] 1532 [3261 b] 0.02 c 51 [124 b] 8201 [30326 b] 2614 [9668 b]

a Relevant bus emission factors were not available. b Lower bound 95% confidence interval values, although physically uninterpretable, can be obtained as a consequence of the normal assumptions underlying the models. c Lower and upper 95% confidence interval values were not available. d Urban roads had average vehicle speeds < 80 km/hr and urban-major roads ≥ 80 km/hr. e The number of days per annum were considered to be 318.8 (refer method section).

284

6.3.2. Comparing the urban SEQ particle inventory with other inventories &

models

Inventories and models which can be compared to the inventory in this study are

discussed below.

Local model estimates of PM10

Local model estimates of total annual PM10 are compared with the quantification

derived for the inventory in this study, and these are shown in Table 6.3.

The urban SEQ inventory quantification for total PM10 of 2614 tonne presented in

this study for 2004 (Table 6.2) compares well with the EPA’s estimate for SEQ of

2249 tonne prepared for 2000 (EPA 2004) shown in Table 6.3. EPA developed a

fleet emissions model using estimates of VKT, emission factors and operating

conditions, and estimated emissions for 6 vehicle classes, 4 fuel types, operating

conditions (average travel speed, road grade, engine hot and cold starts), time of

day, and day of week, summer/winter season (EPA 2004). Despite differences in

area and year of preparation of the inventories, a 10-20% difference for a strategic

model would not be considered unreasonable. Hence we believe this comparison

is valid. Therefore we can have confidence in the total particle inventory

developed in our study for urban SEQ (Table 6.2).

285

Table 6.3 Comparison of estimates of total annual PM10 for SEQ and urban SEQ a

Modellers

Region modelled

Year of inventory

Annual VKT,

millions

Estimate of total PM10 emissions,

Tonne per annum

This study

Urban SEQ a

2004

14,514 b

2614

Environmental Protection Agency (EPA 2004)

SEQ

2000

21,362

2249

Bureau of Transport and Regional Economics (BTRE 2003)

Urban SEQ a

2004

16,340

1840

Apelbaum Consulting (Apelbaum 2006)

Urban SEQ a 2003-2004

21,017 1549

a Although Urban SEQ covers only around 26% of South-East Queensland (ABS 2006;

OESR 2005), the urban SEQ vehicle fleet accounted for more than 70% of private

passenger trips in SEQ in 2004 (SEQHTS 2004).

b Excludes transport-related industry travel (eg., couriers and taxis) and trips by persons

staying in non-private accommodation (eg., tourists and business travellers staying in

hotels) (SEQHTS 2004).

286

PM10 inventories and models prepared by Apelbaum and BTRE were prepared for

2003-2004 and 2004 respectively for urban SEQ (Table 6.3). The Apelbaum

inventory used speed dependent emission factors for different road types based on

a combination of Australian and European data (Apelbaum 2006) and the BTRE

model considered growth in the economy, population, travel demand and urban

congestion, as well as deterioration due to vehicle age and rises in fuel

consumption (BTRE 2003).

Both the Apelbaum and BTRE models exhibited deficiencies. Although

Apelbaum’s annual estimate for PM10 for HDV emissions of 520 tonne was

similar to this study of 488 tonne (Table 6.2), their LDV emission factors were

lower than those used in this study. The BTRE study reported that the uncertainty

in their particulate matter estimates were high, and the part of their analysis with

the greatest levels of uncertainty (BTRE 2003).

Comparing this inventory to a UK inventory

There is only one example of a particle emissions inventory attempted for motor

vehicles that can be compared to this study, prepared for the UK in 1996, 1998

and 2001 (Group 1999; Goodwin et al. 2000; AQEG 2005). These PM10

inventories were derived by multiplying emission factors for different vehicle and

road types by annual VKT data.

Most importantly, in order to derive estimates for PM0.1, PM1 and PM2.5, they

applied distribution profiles for these particle size ranges to PM10 estimate data.

This means that emission factors for these particle size ranges below PM10 were

287

not based on individual measurements of different particle sizes, but simply on

mass fractions multiplied by PM10 data values. The authors stated that their

method depended on PM10 emission rates, which in themselves had substantial

uncertainties, and therefore believe their inventories for small size particles

contain even more uncertainty, due to additional uncertainties in the size

fractions (Group 1999).

The 1996 and 1998 inventories applied the same distribution profiles for petrol

(catalyst) and diesel exhaust, viz., mass fractions of PM10 of 85% for PM1 and

90% for PM2.5 based on 33 different particle size distributions (Group 1999;

Goodwin et al. 2000). The PM0.1 distribution profile was based on size fractions

taken from a European inventory (TNO 1997). Mass fractions of PM10 used in

the 2001 inventory for PM0.1, PM1 and PM2.5 were derived from distribution

profiles taken mostly from the USEPA compilation of emission factors (USEPA

1995) known as AP-42 (AQEG 2005).

Our urban SEQ mass fractions are influenced by the high proportion of LDV

VKT (93% of total VKT). Other possible differences between the inventories are

likely to relate to differences in fleet composition, fuel types, road types, VKT for

different vehicle types, and methods for deriving PM1 and PM2.5 emission factors.

Emission factors used in the UK inventory for petrol vehicles were several orders

of magnitude lower than their diesel values.

288

6.3.3. Results of scenario analyses

Four future scenarios were modelled to emulate likely future responses to events

such as rises in fuel costs, congestion charging, higher density living and transit

oriented development, which could see reductions in on-road VKT, and increases

in walking and cycling, car pooling or rail travel. Other events were staggering of

work and school hours, home based work or schooling, and regular, voluntary

“car free” days. PM1 was not included as relevant Diesel and CNG bus emission

factors were not available, nor PM2.5 emission factors for CNG buses. The

scenarios provide indications of the rates of change in VKT and travel mode

associated with reasonable reductions in particle emissions.

HDV regional emissions

The inventory estimated high levels of HDV emissions, which present a major

problem for the region. Although not entirely feasible or practical, the effect of

completely removing HDVs from urban SEQ would result in reductions in

particle emissions of 54%, 55%, 37% and 19% for particle number, PM1, PM2.5

and PM10 respectively. These reductions do not take into account the predicted

doubling of the on-road freight task within 20 years.

289

Scenarios modelling changes in LDV and bus VKT

Scenarios 1 and 2 modelled percentages of passengers travelling in LDVs and

buses. The results of these scenarios are shown in Tables 6.4 and 6.5, and are

discussed below.

Scenario 1 modelled reductions of 30% and 50% in LDV VKT shifting

percentages of these trips onto new buses. Given that the 24 hour average LDV

occupancy rate was 1.5 passengers, an increase in this occupancy rate to 3

passengers would lead to a 50% reduction in LDV VKT.

In Scenario 1 for each 10% reduction in LDV VKT added to new bus trips,

reductions of around 3-4% for particle number, 1-2% for PM2.5 and 1-6% for

PM10 were modelled for the 24 hour average period (Table 6.4). Table 6.4 shows

modelled reductions in particle number, PM2.5 and PM10, with PM10 modelled

reductions under these scenarios almost double those for particle number.

290

Table 6.4 Modelled reductions in total particle emissions in urban SEQ

in the 24 hour average period

Scenario 1: Reducing Light Duty Vehicle VKT 30% and 50%, and moving proportions of these passengers onto new buses in the 24 hour average period

Reduction in LDV VKT

30%

50%

Percentage of passengers moved to new buses

100% 70% b 100% 70% b

Modelled reduction in total particle emissions (%)

Particle number

9.6

11.2

15.9

18.7

PM2.5 a 1.9 6.0 3.1 10.0

PM10 18.1 19.5 30.2 32.2

a PM2.5 excluded CNG buses due to lack of relevant emission factors, hence the bus fleet was

assumed Diesel-fuelled, resulting in lower modelled reductions. b Assumed the remaining LDV

passengers chose to walk, cycle, catch a train or fill an existing bus, car pool or undertook home-

based work or schooling.

291

Scenario 2 reduced 20% of Home Based Work trip VKT and added 50% of these

trips onto new buses. For each 10% reduction in Home Based Work trip VKT,

where half these trips were added to new buses, reductions were about 2% for

particle number, 2-3% for PM2.5 and 3-4% for PM10 (Table 6.5).

Table 6.5 Modelled reductions in total particle emissions in urban SEQ in

the peak travel times and in the 24 hour average period

Scenario 2: Reducing Home Based Work trip VKT by 20%, and moving 50% of these passengers onto new buses in the peak periods and in the 24 hour average period b

Travel times

7am-9am

4pm-6pm

24 hour

average

Modelled reduction in total particle emissions (%)

Particle number

3.5

4.8

3.1

PM2.5 a

4.7 6.1 4.3

PM10

6.1 7.5 5.5

a PM2.5 excluded CNG buses due to lack of relevant emission factors, hence the bus fleet was

assumed Diesel-fuelled, resulting in lower modelled reductions. b Assumed the remaining LDV

passengers chose to walk, cycle, catch a train or fill an existing bus, car pool, or undertook home-

based work or schooling.

292

Average particle emissions per passenger per km for LDVs and Buses:

Scenario 3 is shown in Table 6.6. As relevant emission factors were not available

for buses for PM1 and for PM2.5 for CNG Buses, these were not included. Average

particle emission factors per passenger per km for particle number for LDVs were

1-2 orders of magnitude higher than for buses; on urban roads for CNG buses

were three times those for Diesel buses; and on urban-major roads Diesel buses

were ten times higher than CNG buses. PM2.5 average particle emission factors

per passenger per km were similar for LDVs and Diesel buses.

On urban roads average particle emission factors per passenger per km for PM10

for LDVs were about 5 times higher than those for Diesel buses, and several

orders of magnitude higher than CNG buses, suggesting opportunities for major

reductions in PM10 by moving proportions of LDV passengers to CNG buses.

There is also ample opportunity to comfortably double LDV and bus occupancy

rates in the region, which were 1.5 passengers and 15.5 passengers respectively

(for the 24 hour average period), leading to further reductions in regional

emissions.

293

Table 6.6 Scenario 3: Average tailpipe particle emission factors per passenger per km for LDVs and buses in urban SEQ (shown in italics) in the 24 hour average period

AVERAGE PARTICLE EMISSION FACTORS PER PASSENGER PER KM

Particle metric

Road Type a

LDV Emission

Factor b

LDV emission

factor per passenger

per km

Diesel Bus Emission Factor b

Diesel Bus

emission factor per passenger

per km

CNG

Emission Factor b

CNG Bus

emission factor per passenger

per km

1014 particles/km

Particle Number Urban road -- -- 3.10 0.20 9.80 0.60 Urban-major road -- -- 1.80 0.10 0.20 0.01 All Road Types 3.60 2.40 -- -- -- --

Particle Mass

mg/km

PM1 All Road Types 16 11 c -- c -- PM2.5 All Road Types 33 22 299 19 c -- PM10 All Road Types 313 20

Urban road 156 104 1.10 0.07 Urban-major road 141 94 0.05 0.003

a Urban roads had average vehicle speeds < 80 km/hr and urban-major roads ≥ 80 km/hr. b Emission factors used to calculate the 2004 urban SEQ inventory, this study. c Relevant PM1 CNG & Diesel and PM2.5 CNG bus emission factors were not available.

294

An estimate of particle emissions in urban SEQ in 2026

The assumptions applied to Scenario 4 are shown in Scenario 4A in Table 6.7,

and the estimated 2026 inventory is presented in Scenario 4B in Table 6.8.

Emission factors for 2026 were based on the emission factors used in the 2004

inventory (Table 6.1), reduced by different percentages (Table 6.7).

The percentage reduction applied to the 2004 emission factor values to derive

emission factors for the 2026 scenarios were generally, but not precisely, based on

vehicle regulations. Firstly, the proposed Swiss ordinance which requires

reductions in solid particle number emissions for LDV Diesel vehicles of 95%, or

greater (BUWAL 2004) and, secondly, the 40% reduction observed between the

highest emission limit values for EURO III and those for EURO IV LDV Diesel

vehicles for PM10 (SAEFL 2004). Based on these vehicle regulations, to derive

emission factors for the 2026 scenario model developed in this paper, the 2004

particle emission factor for LDV Diesel vehicles was reduced by 80% and the

2004 particle emission factor for LDV petrol vehicles by 20%.

295

Table 6.7 Scenario 4A: Model variables and assumptions used to

predict particle number and particle mass emissions in urban

SEQ in 2026

Percentage reduction in 2004 particle emission factor values

Vehicle Type % increase in 2004 VKT

2026 Fleet composition

Particle number

Particle mass,

PM1, PM2.5, PM10

LDV 26% a 50% Diesel 45% Petrol 5% Electric (zero emissions)

80% 20% n/a

40% 40% n/a

HDV 90% b Mainly Diesel 20% 40% Buses 29% a 40% Diesel

50% CNG 10% Hybrid

20% 20%

10% of 2004 Diesel Bus emission

factor

40% 40%

10% of 2004 Diesel Bus emission

factor

a These percentage increases are lower than those predicted for the study region for 2004-2020 of

32% for LDVs and 36% for buses by BTRE (BTRE 2003), as we are basing our assumption on an

assumed shift to walking, cycling, rail or car pool trips in the region in 2026 of around 6-7%.

b HDV VKT is predicted to double over the next 20 years (SKM 2006). It was assumed 10% of

this freight increase would be transported by rail, or other less polluting options that may be

available in the future, such as the use of hybrid/electric intermodal solutions.

296

Table 6.8 Scenario 4B: Estimated total annual particle emissions in

urban SEQ in 2026, compared to the 2004 inventory, this study

Particle metric

2004 inventory

2026

Prediction

Increase/Decrease over

2004 estimates

1025 particles 1027 particles Particle Number

1.08

1.47

Approx. 100-fold Increase

Particle mass

Tonne

Tonne

PM1 477 296 38% Decrease PM2.5 736 472 36% Decrease PM10 2614 1808 31% Decrease

The 20% assumed reduction for LDV petrol vehicles was based on the

assumption that higher reductions in emissions would only be likely to be

achieved if significant advances in technology in the future were realised, and

mandatory particle number emission standards for these vehicle types were

introduced.

Although reductions in PM10 emission limit values of 80% are proposed under

EURO V in 2020 for LDV diesel vehicles (the same limit value is proposed for

LDV petrol vehicles) (EurActiv 2006), a more conservative approach was

adopted in the modelling, and an across the board 40% reduction applied to all

particle mass emission factors. It was also considered that the proportion of 2020-

compliant vehicles would be unlikely to dominate the 2026 fleet.

297

Total particle number emissions for urban SEQ for 2026 were predicted to

increase by more than 2 orders of magnitude, as compared to the 2004 inventory.

This was influenced by an assumed 90% increase in HDV VKT. However,

particle mass emissions in 2026 were predicted to reduce 31-36% (Table 6.8).

6.4. CONCLUSIONS

This study presents the first published comprehensive inventory of motor

vehicle tailpipe particle emissions for particle number and particle mass.

Conclusions from its development and scenario analyses are as follows:-

• Firstly, although HDVs contributed only around 6% of regional

VKT, they contributed more than 50% of particle number and

PM1 emissions to the region, signalling the need for strategies to

reduce HDV diesel vehicle emissions. This finding relates to the

study region, however similar results would be expected in other

areas that have high HDV diesel VKT. Given that the study

region is not highly industrialised and is more service and

tourism oriented, this means that regions with higher levels of

industrialisation could have even larger HDV particle emission

levels. HDV particle emissions are a global problem which

requires reduction strategies such as mandatory fitting of particle

filters, regular emissions testing, and identification of freight

options and freight routes that produce lower emissions per

298

tonne-kilometre and result in lower exposures for populations in

close proximity to truck routes.

• Secondly, the study found that when modelling the movement of

different proportions of LDV passengers to new buses,

reasonable reductions in particulate matter emissions, particularly

for PM10, were able to be achieved. Our study demonstrates the

value of examining and modelling changes in travel mode from

LDVs to new buses, which can be useful to identify the extent to

which changes in travel mode choice between these two travel

modes may lead to reductions in particle emission rates.

The study also found that when calculating average emission

factors per passenger-km for particle number and PM10 for

LDVs, that these were substantially higher than those for buses in

the study region, emphasizing the value of initiatives that

encourage shifts from LDV passenger cars to buses, and which

focus on increasing bus vehicle occupancy rates.

• Thirdly, modelling future scenarios, such as done for 2026 for the

study region which predicted an 100-fold increase in particle

number and 31-36% reduction in particle mass, offer opportunities

to design mitigation efforts tailored to expected changes in travel

demand and vehicle technologies.

299

• Fourthly, recent research has found that PM1 and PM10 constitute a

more discerning combination of mass-based air quality standards for

combustion sources such as motor vehicles than the current

standards of PM2.5 and PM10 (Morawska et al. 2008). To adequately

control particle emissions emitted by motor vehicles, guidelines and

standards need to be introduced for both particle number and PM1 to

complement existing standards. Future development of inventories

for these particle metrics, such as presented in our study, can provide

very important data to inform development of future ambient air

quality guidelines and standards, and this research supports the

relevance and importance of modelling emission inventories which

cover the full size range of particles generated by motor vehicle

fleets.

• Fifthly, urban congestion is a problem not only in SEQ but in many

urban centres around the world. It affects travel time and also has

environmental implications in terms of particle pollution and issues

such as climate change. Regular particle inventories are needed for

urban areas that focus not only on quantifying total daily and annual

particle emissions, but on identifying ‘hot-spots’ and travel routes,

such as roads, truck routes, busways and tunnels posing a risk to

exposed populations. The problem of congestion is an issue

requiring further research, particularly the need to derive speed-

related particle emission factors to model congestion and vehicles

travelling at lower speeds. In addition, research is needed to

300

evaluate particle emission levels pre- and post construction of

transport infrastructure, high density living and transit oriented

developments.

• Sixthly, it is important to extend work such as that presented in this

inventory to estimate the spatial distribution of particle

concentrations, and to gain an understanding of the socioeconomic

characteristics of populations affected by ‘hot-spots’. Inventories

such as presented in our work provide new knowledge that can be

used in climate models to develop an understanding of the quantity

and impacts of motor vehicle particle emissions on the global

airshed, including particle concentrations reaching into the

troposphere and stratosphere, as well as their potential effects, such

as contributing to the cooling and dimming of the planet.

301

6.5. REFERENCES



Australian Bureau of Statistics (ABS), 2004a. Survey of Motor Vehicle Use

Australia. Australian Bureau of Statistics, Canberra.

Australian Bureau of Statistics (ABS), 2004b. Population by Age and Sex.

Australian Bureau of Statistics, Canberra.

Australian Bureau of Statistics, Census Data (ABS) 2006, Community Profile for

Brisbane, Australian Bureau of Statistics, Canberra.

Brisbane City Council, 2007, Brisbane City Council Transport Plan for Brisbane

2006-2026 Draft, Brisbane City Council, Brisbane.

Air Quality Expert Group (AQEG), 2005. Particulate Matter in the UK. London,

Department for Environment, Food and Rural Affairs.

Bigg, E.K., Turvey, D.E., 1978. Sources of atmospheric particles over Australia.


Bureau of Transport and Regional Economics (BTRE), 2003. Urban pollutant

emissions from motor vehicles: Australian trends to 2020, Final Draft Report for

Environment Australia. Canberra, BTRE.

BUWAL., 2004, 1 March. Ordinance on the determination of the particle number

emission level of passenger cars with compression ignition engines, Draft.

http://www.puntofocal.gov.ar/doc/che39.pdf. Date verified 1 November 2008.

302

Cadle, S. H., Ayala, A., Black, K. N., Graze, R. R., Koupal, J., Minassian, F.,

Murray, H. B., Natarajan, M., Tennan, C. J., Lawson, D. R., 2008. Journal of Air

and Waste Management Association. Real-World Vehicle Emissions: A Summary

of the Seventeenth Coordinating Research Council On-Road Vehicle Emissions

Workshop 58, 3-11.

California Air Resources Board (CARB), 2001. Heavy-Duty Emissions

Laboratory, Heavy Duty Testing and Field Support Section, California Air

Resources Board. Report No. 01-01.

Chen, H., Namdeo, A., Bell, M., 2008. Classification of road traffic and roadside

pollution concentrations for assessment of personal exposure. Environmental

Modelling & Software, 23(3), 282-287.

Cheng, S., Chen, D., Li, J., Wang, H., Guo, X., 2007. The assessment of

emission-source contributions to air quality by using a coupled MM5-ARPS-

CMAQ modeling system: A case study in the Beijing metropolitan region, China.

Environmental Modelling & Software, Volume 22, Issue 11, November 2007,

Pages 1601-1616.

Cohen, J., Cook, R., Bailey, C.R., Carr, E., 2005. Relationship between motor

vehicle emissions of hazardous pollutants, roadway proximity, and ambient

concentrations in Portland, Oregon. Environmental Modelling & Software,

Volume 20(1), 7-12.





303





Brussels.

Costabile, F., Allegrini, I., 2008. A new approach to link transport emissions and

air quality: An intelligent transport system based on the control of traffic air

pollution. Environmental Modelling & Software, 23(3, 258-267.


Date verified 1 November 2008.

Environmental Protection Agency (EPA), 2004. Air Emissions Inventory South-

east Queensland Region. Queensland Government, Brisbane.

EurActiv.com 2006. EURO 5 emissions standards for cars, EU News, Policy

Positions & EU Actors online. http://www.euractiv.com/en/transport/euro-5-

emissions-standards-cars/article-133325. Date verified 1 November 2008.

European Commission Joint Research Centre (ECJRC), 2002. Guidelines for


particulate matter for use in epidemiological studies. EUR 20238 EN 2002. L. M.

D. Schwela, D. Kotzias, European Commission, Italy.






304

Goodwin, J. W. L., Salway, A. G., Murrells, T. P., Dore, C. J., Passant, N. R.,

Eggleston, H. S., 2000. UK emissions of air pollutants 1970-1998. A Report of

the National Atmospheric Emissions Inventory. London, Department of the


Goodwin, J. W. L., Salway, A. G., Eggleston, H. S., Murrells, T. P., Berry, J.E.,

1999. National Atmospheric Emissions Inventory, UK Emissions of Air

Pollutants 1970 to 1996, National Environmental Technology Centre on behalf of

the Department of the Environment, Transport and the Regions.





Hurley, P.J., Physick, W.L., Luhar, A.K., 2005. TAPM: a practical approach to

prognostic meteorological and air pollution modelling. Environmental Modelling

& Software, Volume 20, Issue 6, June 2005, Pages 737-752.

Jayaratne, E.R., Ristovski, Z.D., Meyer, N., Morawska, L., 2008. Particle and

Gaseous Emissions from Compressed Natural Gas and Ultralow Sulphur Diesel-

Fuelled Buses at Four Steady Engine Loads. Science of the Total Environment

407 (8), 2845-2852.

Jones, A. M., Harrison, R.M., 2006. Estimation of the emission factors of particle



Keogh, D.U., Kelly, J., Mengersen, K, Jayaratne, E.R., Ferreira, L., Morawska, L.,

2009. Derivation of motor vehicle tailpipe particle emission factors suitable



0.1007/s11356-009-0210-9.

305

McKay, L., McGaw, N., 2005, TravelSmart™ - the innovative solution to

competing demands. National Conference Proceedings of the Australian Institute

of Traffic Planning & Management Inc., Brisbane.

Mensink, C., Lefebre, F., Janssen, L., Cornelis, J., 2006. A comparison of three

street canyon models with measurements at an urban station in Antwerp,

Belgium, Environmental Modelling & Software, Volume 21(4), 514-519.

Ministry of Transport, 2007. 2005 Household Travel Survey Summary Report

2007 Release, NSW Ministry of Transport, Transport Data Centre, August.

Mobley, J.D., Cadle, S. H., 2004. Innovative Methods for Emission Inventory





Wiley-VCH.

Morawska, L., Keogh, D. U., Thomas, S. B., Mengersen, K., 2008. Modality in



Namdeo, A., Mitchell, G., Dixon., R. 2002. TEMMS: an integrated package for

modelling and mapping urban traffic emissions and air quality. Environmental

Modelling & Software, Volume 17(2), 177-188.

Office of Urban Management, 2004. Draft South East Queensland Regional Plan:

For Consultation. Brisbane Department of Local Government, Planning, Sport &

Recreation, Queensland Government.

Ortuzar, J. de., Willumsen, L.G., 2001. Modelling Transport. John Wiley & Sons

Inc.

306



Pope, C. A., .Dockery, D. W., 2006. Health Effects of Fine Particulate Air



Queensland Department of Main Roads, 2008. Brisbane Strategic Transport

Model, Queensland Government, Brisbane.

Office of Economic and Statistical Research (OESR) 2005, Queensland Regional

Profiles 2004, Brisbane and Moreton Statistical Divisions, Office of Economic

and Statistical Research, Queensland Treasury, Brisbane.

Oxley, T., Valiantis, M., Elshkaki, A., ApSimon, H.M., 2009. Background, Road

and Urban Transport modelling of Air quality Limit values (The BRUTAL

model). Environmental Modelling & Software, 24(9), 1036-1050.

Samaras, Z., Ntziachristos, L., Thompson, N., Hall, D., Westerholm, R., Boulter,

P., 2005. Characterisation of Exhaust Particulate Emissions from Road Vehicles,

PARTICULATES program, European Commission. Contract No 2000-

RD.11091, source http://lat.eng.auth.gr/particulates/downloads.htm.




Sinclair Knight Merz (SKM), 2006. Twice the Task: A review of Australia's

freight transport tasks. Melbourne, Victoria, National Transport Commission.

Sokhi, R.S., Mao, H., et al., 2008. An integrated multi-model approach for air

quality assessment: Development and evaluation of the OSCAR Air Quality

Assessment System. Environmental Modelling & Software, 23(3), 268-281.

307

South-East Queensland Household Travel Survey (SEQHTS), 2004. South-East

Queensland Household Travel Survey 2003-2004 (Brisbane, Gold Coast and

Sunshine Coast Area). Queensland Transport, Brisbane.

Smit, R., Brown, A.L., Chan, Y.C., 2008. Do air pollution emissions and fuel

consumption models for roadways include the effects of congestion in the

roadway traffic flow?. Environmental Modelling & Software, 23(10-11), 1262-

1270.

Smit, R., Smokers, R., Rabe, E., 2007. A new modelling approach for road traffic



Swiss Agency for the Environment, Forests and Landscape (SAEFL), 2004. Air

Pollutant emissions from Road Transport 1980-2030 Environmental Series No.

355. Berne SAEFL.

TNO, 1997. Particulate Matter Emissions (PM10, PM2.5, PM<0.1) in Europe in



Brisbane.

US EPA., 1993. User's Guide to MOBILE5A, Mobile source emissions factor


US EPA., 1995. Compilation of Air Pollutant Emission Factors, 5th edn, AP-42,

North Carolina.

World Health Organization (WHO) 2005. Guidelines for Air Quality. World

Health Organization, Geneva.

308

CHAPTER 7

AMBIENT NANO AND ULTRAFINE PARTICLES

FROM MOTOR VEHICLE EMISSIONS:

CHARACTERISTICS, AMBIENT PROCESSING AND

IMPLICATIONS ON HUMAN EXPOSURE

Lidia Morawska1, Zoran Ristovski1, Rohan Jayaratne1,

Diane.U. Keogh1, Xuan Ling1


University of Technology, Brisbane, Queensland, Australia

(2008) Atmospheric Environment 42 (35), 8113-8138

309

STATEMENT OF JOINT AUTHORS

Title: Ambient nano and ultrafine particles from motor vehicle emissions:

characteristics, ambient processing and implications on human

exposure

Authors: Lidia Morawska, Zoran Ristovski, Rohan Jayaratne,

Diane. U. Keogh and Xuan Ling

Lidia Morawska

Review and synthesis of current knowledge on factors affecting particle

concentrations, including transport, processing, dynamics, chemical

composition, temporal, spatial and seasonal variation; and ultrafine particle

correlations with gaseous pollutants. Contributed to the manuscript.

Zoran Ristovski

Review and synthesis of current knowledge on sources of particles in natural

environments, and of the role of after-treatment devices in terms of motor

vehicle emissions. Contributed to the manuscript.

Rohan Jayaratne

Review and synthesis of current knowledge on vehicle emissions as a source of

ultrafine particles, primary and secondly particles, the role of fuels, ions, and

road-tyre interface in terms of ultrafine particles. Contributed to the manuscript.

310


Review and synthesis of current knowledge related to the location of the mode

within particle size distributions in a wide range of different worldwide

environments. Review and synthesis of current knowledge on published

emission factors for different vehicle types for particle number, and on

development of motor vehicle particle emission inventories. Contributed to the

manuscript.

Xuan Ling

Literature review, analysis and synthesis of current knowledge on the

capabilities and limitations of current measurement techniques used to measure

particle number concentration. Contributed to the manuscript.

311

ABSTRACT

The aim of this work was to review and synthesize the existing knowledge on

ultrafine particles in the air with a specific focus on those originating due to

vehicles emissions. This constitutes Part II of a literature review on ultrafine

(UF) particles, with industrial and power plant emissions covered in Part I. As

the first step, the review considered instrumental approaches used for UF

particle monitoring and the differences in the outcomes they provide. This was

followed by a discussion on the emission levels of UF particles and their

characteristics as a function of vehicle technology, fuel used and after treatment

devices applied. Specific focus was devoted to secondary particle formation in

urban environments resulting from semi volatile precursors emitted by the

vehicles. The review discussed temporal and spatial variation in UF particle

concentrations, as well as particle chemical composition and relation with

gaseous pollutants. Finally, the review attempted to quantify the differences

between UF particle concentrations in different environments. These, as well as

other aspects of UF characteristics and dynamics in the air, were discussed in

the context of human exposure and epidemiological studies as well as in relation

to management and control of the particles in vehicle affected environments.

312

7.1. INTRODUCTION

Ultrafine and nano particles present in the air due to natural sources and

processes, as well as those resulting from anthropogenic activities have attracted

an increasing level of interest in the last decade. Ultrafine particles (UF) are

defined as those with diameters smaller than 0.1 μm, and their subset,

nanoparticles as smaller than 0.05 μm. Both these terms constitute a somewhat

arbitrary classification of particles in terms of their size, indicating the

significant role of this physical characteristic on particle fate in the air. Also

health and environmental effects of particles are strongly linked to particle size,

as it is the size which is a determinant (in a probabilistic sense) of the region in

the lung where the particles would deposit or the outdoor and indoor locations,

to which the particles can penetrate or be transported. In addition, sampling of

particles and choice of an appropriate instrumentation and methodology is

primarily based on particle size. Airborne concentration of UF and nanoparticles

is most commonly measured and expressed in terms of number concentrations

of particles per unit volume of air, in contrast to larger particles, which are

measured in terms of mass concentration.

The size of particles, however, depends on the multiplicity of sources and

processes which lead to their formation, and therefore, on the material from

which the particles were formed, with the complex scientific knowledge behind

these processes still containing many significant gaps. The recent interest in UF

particles is to a large extent due to the impact of anthropogenic processes,

resulting in unprecedented increases in particle concentration, often by one or

two orders of magnitude above their natural concentrations. The most

313

significant are the various outdoor anthropogenic combustion sources, including

vehicles (and other forms of transport), as well as industrial and power plants,

all utilising fossil fuels. Another significant combustion source is biomass

burning, including controlled and uncontrolled forest and savannah fires. There

are also indoor combustion sources such as stoves and heaters utilising fossil

fuels and biomass, as well as tobacco smoking.

The interest in UF particles has resulted in a large body of literature published in

recent years, reporting on various aspects and characteristics of these particles.

Therefore, the aim of this work was to review and synthesize the existing

knowledge and to draw conclusions as to the picture emerging with regard to

these particles in atmospheric systems. Out of the two main outdoor

anthropogenic sources, this paper is focused on vehicle emissions, while the

companion paper targets industrial and power plants as sources of UF particles.

Not included in this review is the contribution of biomass burning (controlled

and uncontrolled fires), and incineration of refuse to local or global UF particle

concentrations. Both are topics for independent reviews.

7.2. CAPABILITIES AND LIMITATIONS OF PARTICLE NUMBER MEASUREMENT METHODS

A full review of the instrumental methods for measuring of UF particle

properties is outside the scope of this review paper and the reader is directed to

several recent publications addressing this topic, e.g. (McMurry 2000).

However, it is important to consider the existing methods for particle number

and size distribution measurements, since it is the very nature of the

314

instrumental method which determines the measurement outputs and in turn

their compatibility with those obtained utilising different methods. The majority

of the published studies reporting on particle number and number size

distribution applied electrostatic classifiers (EC) and condensation particle

counters (CPC) manufactured by TSI Incorporated (www.tsi.com), with a much

smaller number using other instruments, for example GRIMM

(www.dustmonitor.com), or air ion mobility spectrometers, which have enabled

measurements down to 0.4 nm (Mirme et al. 2007). The latter measures only

naturally charged particles, and have been used only in a handful of studies.

When referring to UF or nanoparticles, an unspoken assumption is made that the

instrumental methods used provide information on particles in the two specific

size ranges (<0.05 and <0.1 μm, respectively). This is possible if the

instrumental method enables measurements of particle number size distribution,

usually in a broader range, from which the sections of data encompassing UF or

nanoparticles is extracted. Such methods are most commonly based on

electrostatic classifiers operating in combination with particle counters as

differential/scanning mobility particle sizers (DMPS or SMPS, respectively)

(Baron and Willeke 2001). The lower end of the size window is determined both

by instrumental factors and operator decisions. In the first instance, the lower

size limit is determined by the capability of the CPC and ranges from 2-10 nm.

However, most commonly the DMPS/SMPS lower end of the window is set to a

value above this, in the range from 10 – 20 nm. The reason for setting it up to 10

nm higher than the achievable lower limit is that this provides a compromise as

to the overall size of the window. Losing the few nanometres at the lower end

315

enables a significant extension of the window at the upper end, which in most

cases is a preferable option, unless a study specifically focuses on the nucleation

mode.

If, rather than employing instrumentation for particle size distribution

measurement, only a particle counter is used, the outcome of the measurement is

the total particle number concentration in the detection size range of the

instrument. There are two important implications of this to the interpretation of

this value as a measure of UF particles. Firstly, this means that the outcomes of

the measurements are not specifically UF or nano particle concentrations, unless

specific inlets are used which restrict the range of particles entering the

detecting arm of the instrument. While it is true that in most typical

environments particle number concentration is dominated by UF particles,

which is, thus, usually a good approximation of the total particle number

concentration, it is important to keep in mind that these are not the same, that

there are environments where there are significant particle modes outside the

UF range (see section 7.1, below) and therefore the two concentrations (UF and

total number) differ significantly. Secondly, and even more significantly, the

condensation particle counters often detect particles in the range extending to

lower sizes than the window set by the DMPS/SMPS. This means that the

counters are capable of detecting particles in the earlier stages of nucleation, and

the presence of the nucleation mode which is below the size detection limit set

by the DMPS/SMPS. Therefore in most situations, the counters would detect

more particles than the DMPS/SMPS, and significantly more in the

environments where a nucleation mode is frequently present.

316

The above points are important when comparing particle number concentrations

reported in different papers and when specifically considering UF or

nanoparticles. Since different studies use different sets of instrumentation and

investigate a different size range window, comparison of the total particle number

concentrations reported should be conducted with caution. In order to develop at

least a broad understanding of the impact which these differences have on the

measured particle concentrations, data from 52 studies reporting total particle

number concentrations for a range of different environments was compiled and

then the results grouped according to the measurement technique used: CPC or

DMPS/SMPS.

1

10

100

CPC SMPSInstrument

103 Pa

rtic

les/

cm3

meanmedian

Figure 7.1. Comparison of reported particle number concentrations measured

by CPC or DMPS/SMPS*.

* These CPC and SMPS results were extracted from the following papers: Aalto et al. (2001), Harrison et

al. (1999), Kittelson et al. (2004) and Shi et al. (2001a) who used both the CPC and SMPS; Vakeva et al.

(1999), Zhu et al. (2004), Imhof et al. (2005a), Paatero et al. (2005) and Westerdahl et al. (2005) who used

only the CPC and McMurry and Woo (2002), Tuch et al. (1997), Morawska et al. (1999a), Hitchins et al.

317

(2000), Junker et al. (2000), Jamriska and Morawska (2001), Pitz et al. (2001), Ruuskanen et al. (2001),

Cheng and Tanner (2002), Molnar et al. (2002), Morawska et al. (2002), Thomas and Morawska (2002),

Wehner et al. (2002), Zhu et al. (2002a), Zhu and Hinds (2002b), Ketzel et al. (2003), Longley et al.

(2003), Tunved et al. (2003), Wehner and Wiedensohler (2003), Gidhagen et al. (2004), Gramotnev and

Ristovski (2004), Gramotnev et al. (2004), Hussein et al. (2004), Jamriska et al. (2004), Janhall et al.

(2004), Jeong et al. (2004), Ketzel et al. (2004), Morawska et al. (2004), Stanier et al. (2004a), Gidhagen et

al. (2005), Holmes et al. (2005), Imhof et al. (2005b), Rodriguez et al. (2005), Janhall et al. (2006),

Virtanen et al. (2006), Wahlin et al. (2001), Woo et al. (2001b), Abu-Allaban et al. (2002), Laakso et al.

(2003), Hussein et al. (2005a) and Mejia et al. (2007a) who used only the SMPS. Other studies, such as

(Hameri et al. 1996; Kaur et al. 2006), which measured particle concentration without using a CPC or

SMPS (e.g. P-trak etc.) were not included in Figure 7.1, nor were the four tunnel studies (Abu-Allaban et

al. 2002; Gouriou et al. 2004; Jamriska et al. 2004; Imhof et al. 2005b) (see comments in relation to tunnels

in section 7.4 below).

The mean concentrations measured by the CPC's and DMPS/SMPS's are

36.8×103/cm3 and 30.6×103/cm3, respectively, and the median concentrations

are 24.9×103/cm3 and13.5×103/cm3, respectively. In other words, the mean and

the median CPC measurements are 32% and 56%, higher than DMPS/SMPS's

measurements, respectively. The difference in the means was tested using a

Students t-test and found to be statistically significant at a confidence level of

over 99%.

The overall comparison of the differences between the total particle

concentration values measured by CPCs and DMPS/SMPSs has some

shortcomings. In particular, the differences for specific environments could

vary, where larger differences are expected for environments where a nucleation

mode is present and smaller where aged aerosol dominates. Moreover,

corrections for particle losses within the two instruments may play a significant

role. Nevertheless, the comparison shows what overall magnitude of differences

318

can be expected when comparing results using these different measuring

techniques. It is important to keep these differences in mind when attempting to

establish quantitative understanding of variation in particle concentrations

between different environments, which is of significance for human exposure

and epidemiological studies.

It is worth mentioning that large discrepancies have also been observed when

comparing the results of particle number concentrations measured directly from

vehicle exhaust. While particle volume/mass showed reasonable reproducibility

in between different studies, results of particle number measurements were

difficult to reproduce, even in the same study. Some artefacts and poor

reproducibility in vehicle emission measurements were due not only to the

different instruments used but also to the fact that the majority of particles (in

terms of number) belonged to the nucleation mode and were formed in the

process of dilution. The number of particles formed in the nucleation mode is

very sensitive to the dilution conditions and any slight changes (of the dilution

temperature, for example) can result in a significant change in particle number

concentration. A detailed discussion on the effects of dilution conditions on

sampling and measurements of particle numbers in vehicle emissions can be

found in Kasper (2005). In order to develop a method that could be used in a

reproducible and comparable manner in laboratories around the world, the

UNECE-GRPE Particulate Measurement Program (PMP) was formed. This

program focused on future regulation of nano-particle emissions from light duty

vehicles and heavy duty engines with the goal to amend existing approval

legislation to stipulate an extensive reduction of particle emissions from mobile

319

sources (Mohr and Lehmann 2003). Based upon the recommendation of the PMP,

the European Commission has added a particle number limit to its Euro 5/6

proposed emission standards for light-duty vehicles. Only solid particles are

counted, as volatile material is removed from the sample, according to the PMP

procedure.

7.3. SOURCES OF PARTICLES IN NATURAL ENVIRONMENT

While the main focus of this review is on the impact of vehicle emissions on

ambient characteristics of UF and nanoparticles, for completeness, and in order to

fully understand this impact, firstly natural sources and their contributions are

discussed, as they result in the natural background of the particles in ambient air.

Vehicle emissions increase particle concentrations over this background and result

in an overall change of particle characteristics.

Of particular importance in natural environments is the formation of new

particles, of which the main mechanism is nucleation of low-volatile gas-phase

compounds, followed by their growth into small particles. It is not the intention of

this review to go into great detail about the mechanisms of particle formation

(both natural and anthropogenic). For more details on this topic, the reader is

referred to several recent literature reviews (Kulmala et al. 2004; Holmes 2007).

There were a number of observations of new particle formation in natural

environments ranging from very clean environments such as the arctic (Birmili

and Wiedensohler 2000), boreal forests in the northern hemisphere (O'Dowd et al.

2002; Tunved et al. 2006), eucalypt forests of Australia (Suni et al. 2007) and a

large number of studies in the coastal areas (see for example a review by Kulmala

320

et al (2004)). In general, there are a much larger number of observations from the

northern hemisphere than from the southern hemisphere. Observations were also

made from a variety of platforms ranging from ground based to ships and

aeroplanes. In most of these observations the measurements were made such that

the platform was not moving along with the same air parcel. Therefore

observations of new particle formation may be biased by spatial variations of

constituents in different air parcels (Kulmala et al. 2004).

In remote environments particle formation events are preceded by an increase in

the atmospheric concentration of sulphuric acid, with the increase in the particle

number occurring about 1–2 h after an increase in sulphuric acid was measured

(Weber et al. 1997). This is followed by a relatively small particle growth rate

between 1 and 2nm h-1 (Weber et al. 1996; Marti and Weber 1997; Weber et al.

1997; Birmili and Wiedensohler 2000). These events showed a linear relationship

between the number of newly formed particles and the production rate of

sulphuric acid indicating the importance of sulphuric acid. The question still

remains: Is the binary nucleation solely responsible for the formation of these

particles or is a third species such as ammonia or an organic involved? Birmilli

and Wiedensohler (2000) estimated that the concentration of sulphuric acid

needed to achieve the same nucleation rates through binary nucleation was over

two orders of magnitudes higher than that measured. Napari et al. (2002) using

their observations and parameterisation of the ternary nucleation rate with an

atmospheric ammonia concentration of 20 pptv and the measured sulphuric acid

concentration obtained good agreement with the observed nucleation rates.

321

In forests the sources of new particle formation are different. The mechanisms

responsible for the formation and growth of these particles are still uncertain.

Although sulphuric acid is one of the most likely candidates thought to be

responsible for the formation of the initial nanometer-sized particles (Riipinen et

al. 2007), sulphur chemistry does not sustain enough sulphuric acid in the

atmosphere to explain more than a small fraction of the observed particle-size

growth rate. To explain the observed growth, which is up to a diameter of 50 to

100 nm, other compounds are required. O’Dowd et al (2002) showed that particle

formation can commonly occur from biogenic precursors. A recent study by

Tunved et al. (2006) showed a direct relation between emissions of monoterpenes

and gas-to-particle formation over these regions which were substantially lacking

in anthropogenic aerosol sources. Therefore, secondary organic aerosol formation

from monoterpenes is an important source in these environments. Further, the

authors show that the forest provides an aerosol population of 1-2 x 103 cm-3 of

climatically active particles during the late spring to early fall period, presenting a

substantial source of global importance.

Proposed particle production mechanisms in the marine environment include the

seawater bubble-burst process (O'Dowd et al. 2004), ternary nucleation

producing a reservoir of undetectable particles upon which vapours can

condense (Kulmala et al. 2000), free tropospheric production with mixing down

to the boundary layer (Raes 1995), and the generation of coastal iodine particles

from macroalgal iodocarbon emissions (Raes 1995; Kulmala et al. 2000;

O'Dowd et al. 2004; O'Dowd and Hoffmann 2005). While iodine-containing

particles were found in large numbers at Mace Head research station in Ireland,

322

they are not likely to play an important role globally (McFiggans 2005). Wind-

produced bubble-burst particles containing salt are ubiquitous in the marine

environment (Ayers and Gras 1991), but these represent less than 10% of

particle numbers. The majority of particles are much smaller than these salt

particles and their origins remain only partially explained.

Several conclusions can be derived from this brief review. Firstly, particles are

formed in the environment due to natural processes and therefore are always

present at some background concentration levels. Therefore, when considering

particle concentrations in urban environments it is important to compare them to

the background levels in order to assess the magnitude of the anthropogenic

impacts (see Section 7.6 below). Secondly, the mechanisms of new particle

formation exhibit similar complexities in both types of environments (natural

and vehicle affected), strongly depend on local meteorological factors, and

therefore a complete picture of the dynamics of particle formation in urban

environments must include all factors involved. These issues are further

discussed in Section 7.6 below.

323

7.4. VEHICLE EMISSIONS AS A SOURCE OF ULTRAFINE

PARTICLES

7.4.1. Introduction

As discussed previously, many studies have conclusively shown that motor vehicle

emissions constitute the major source of ultrafine particle pollution in urban

environments (Harrison et al. 1999; Shi and Harrison 1999; Shi et al. 1999; Shi et

al. 2001a; Wahlin et al. 2001). Particles emitted from diesel engines are in the size

range 20-130 nm (Kittelson 1998; Morawska et al. 1998a; Harris and Maricq

2001; Ristovski et al. 2006) and from petrol engines in the range 20-60 nm

(Harris and Maricq 2001; Ristovski et al. 2006). Therefore, it is not surprising that

a large fraction of the particle number concentration in urban air is found in the

UF size range (Morawska et al. 1998b). Overall, it has been shown that in urban

environments the smallest particles make the highest contribution to the total

particle number concentrations, while only a small contribution to particle volume

or mass. A US study by Stanier et al. (2004a) showed that 25% of the aerosol

number is less than 10 nm and 75% of the aerosol number is less than 50 nm.

Similarly Woo et al. (2001b) showed that 26% of particle number is smaller than

10 nm and 89% smaller than 100 nm. Zhang et al. (2004b) showed that the

number concentrations of particles the size ranges of 0.011-0.050 µm and 0.011-

0.1 µm accounted for approximately 71% and 90%, respectively of the total

number concentration. However, particles smaller than 0.050 µm contributed only

3% to the total volume concentration, while the largest contribution of 87% was

from particles larger than 0.1 µm. In Europe Junker et al. (2000) found that the

highest in terms of number were concentrations of particle <0.1 μm, averaging

between 82–87% of the total particle numbers <0.421 μm while the accumulation

324

mode (0.1–2.8 μm) made up for most of the particle mass (mean >82%). Shi et al.

(2001a) showed that particles smaller than 10 nm contributed more than 36-44%

of the total particle number concentration in an urban roadside location and

particles within the size range 3–7nm accounted for 37% of total measured

particles. Charron and Harrison (2003) showed that particles ranging from 11 to

100 nm represent from 71% to 95% (median 88.7%) of the particle number

between 11 and 450 nm. Pirjola et al. (2006) reported that in winter in Helsinki,

Finland, 90–95%, and in summer, 86–90% of particles were smaller than 50 nm,

while Virtanen et al. (2006) estimated for the same data set that particles smaller

than 63 nm made up ~90% of particles in the winter and ~80% of particles in

summer. Peak concentrations often exceeded 2 x 105 cm-3 and sometimes reached

1 x 106 cm-3. In Australia, Mejia et al. (2007a) showed that UF contributed to 82-

90% of the particle number and nanoparticles to around 60-70%, except at a site

mainly influenced by heavy duty diesel vehicles, where the nanoparticle

contribution dropped to 50%.

A large fraction of these ultrafine particles come from heavy-duty diesel vehicles.

Kirchstetter et al. (1999) measured particle emissions from light and heavy duty

vehicles in a roadway tunnel and showed that heavy-duty diesel trucks emitted 24,

37 and 21 times more fine particles, black carbon and sulphate mass per unit fuel

mass burned than light duty vehicles. Heavy-duty vehicles also emitted 15-20

times the number of particles per unit mass of fuel burned compared to light-duty

vehicles. In general, a heavy duty diesel truck or bus exhibits particle number

emission factors that are one to two orders of magnitude larger than a typical

petrol car (Morawska et al. 2005; Ristovski et al. 2005; Ristovski et al. 2006). The

325

only exception was observed by Graskow et al. (1998) who found that when

petrol vehicles were driven at higher velocities (~120 km/h) or with higher

loads, i.e. during acceleration, the particle number emissions from petrol

vehicles came close to that observed from diesel vehicles.

In general, particles from vehicle emissions can be divided into two broad

categories, depending on the location of their formation. They can be formed in

the engine or tailpipe (primary particles) or they can be formed in the

atmosphere after emission from the tailpipe (secondary particles).

7.4.2. Primary Particles

Primary particles are generated directly from the engine and are mostly

submicrometer agglomerates of solid phase carbonaceous material ranging in

size from 30 to 500 nm and residing mainly in the accumulation mode. They

may also contain metallic ash and adsorbed or condensed hydrocarbons and

sulphur compounds. Metallic ash is generally derived from lubricating oil

additives and from engine wear. The size distribution of particles in the

accumulation mode are very well represented by lognormal distributions, with

an almost constant standard deviation of 1.8–1.9 (Harris and Maricq 2001), and

it does not vary significantly between measurements from a given vehicle under

different operating conditions. Repeated measurements from a diesel engine, in

particular, can be very consistent (Kasper 2005; Ristovski et al. 2006). For this

reason, the primary solid particle number limit has been added to the European

Commissions proposed Euro 5/6 emission standards for light-duty vehicles.

326

7.4.3. Secondary Particles

As the hot exhaust gases are expelled from the tailpipe of a vehicle, they cool

and condense to form large numbers of very small particles in the air. They are

volatile and consist mainly of hydrocarbons and hydrated sulphuric acid. These,

so-called secondary particles, are generally in the nanoparticle size range below

30 nm and compose the nucleation mode, which have been commonly observed

near busy freeways, especially carrying a large fraction of heavy duty diesel

vehicles (Harrison et al. 1999; Kittelson et al. 2002; Charron and Harrison 2003;

Sturm et al. 2003; Gramotnev and Ristovski 2004; Zhu et al. 2004; Rosenbohm

et al. 2005; Westerdahl et al. 2005; Ntziachristos et al. 2007). They have also

been observed in on-road studies, such as when a vehicle is being followed by a

mobile laboratory (Vogt et al. 2003; Kittelson et al. 2004; Pirjola et al. 2004;

Gieshaskiel et al. 2005; Kittelson et al. 2006a; Ronkko et al. 2006; Casati et al.

2007). However, while sometimes present, they are not commonly observed in

dynamometer measurements where dilution tunnels are used to cool and dilute

the exhaust gases (Rickeard et al. 1996; Khalek et al. 1999, 2000; Kittelson et

al. 2006a; Ristovski et al. 2006).

It has been shown that the conditions necessary for the production of these

volatile nanoparticles are strongly affected by the dilution conditions such as the

dilution rate, dilution ratio, temperature and residence time (Khalek et al. 1998,

1999; Shi and Harrison 1999; Khalek et al. 2000; Kawai et al. 2004; Mathis et

al. 2004; Kasper 2005). Lyyranen et al (2004) investigated particle number

distributions obtained from a turbo-charged diesel off-road engine using several

different dilution systems and concluded that nucleation modes were observed

327

when the dilution process involved rapid cooling and mixing of the exhaust.

Casati et al. (2007) measured particle emissions from a diesel passenger car in

field and laboratory conditions and concluded that the nucleation mode was

strongly affected by dilution conditions and decreased when the exhaust was

more diluted.

Similarly, on-road dilution of the exhaust plume is very important in the

generation of secondary particles in the exhaust plume. Ronkko et al. (2006)

studied particle size distributions in emissions from an on-road heavy-duty

diesel vehicle and demonstrated that the formation of the nucleation mode 5m

behind the vehicle was favoured by low ambient temperatures and high relative

humidity. For smaller distances no nucleation modes were observed neither by

Ronkko et al (2006) or Morawska et al (2007b) for a diesel vehicle. During on-

road measurements using light duty spark ignition (SI) vehicles, Kittelson et al

(2006b) did not observe a significant particle signature above background under

highway cruise conditions. Much higher number emissions were observed

during acceleration, at high-speed cruise, and during cold starts.

In addition to the dilution and cooling effects, there is another factor that plays

an important role in determining the concentration of secondary particles. The

gaseous precursors condense or adsorb on to the surface of carbon particles in

the accumulation mode. If the concentration of carbon particles is low, the gases

will nucleate homogeneously, giving rise to large concentrations of volatile

nanoparticles. This has been clearly observed with diesel vehicles equipped with

particle filters (Burtscher 2001), where the accumulation mode has been

328

removed by the particle filter leading to large nucleation modes. The presence

of a large accumulation mode will act to suppress the formation of the nuclei

mode because the carbonaceous agglomerates scavenge volatile material

reducing the likelihood of nucleation. Older vehicles with excess soot emissions

are less likely to exhibit nucleation modes. Therefore, the number concentration

of the nucleation mode particles, unlike the accumulation mode particles, is

highly unstable and unpredictable. Further, as in some instances, as many as

90% of the total particle number may occur in the nucleation mode, total

particle number emissions from similar types of motor vehicles may vary by

over an order of magnitude (Ristovski et al. 2004).

7.5. ROLE OF FUELS

Particle emissions from motor vehicles are significantly affected by the nature

of the fuel used and thus a considerable effort is being devoted to investigations

of fuel properties and their impacts on particle emissions.

Sulphur in Diesel Fuel: The main source of chemically-bound sulphur in diesel

fuel is that which occurs naturally in crude oil and is in a volatility range which

leads to its incorporation in the diesel fuel fraction. The presence of sulphur is

useful as it increases the lubricating properties of the fuel. During combustion, a

fraction of this sulphur is oxidised to sulphur trioxide which binds with water to

form sulphuric acid that contributes to total particle emissions. However, the

presence of sulphur in diesel has several other adverse effects such as the

corrosion of the exhaust system and increased wear and tear on engine parts. For

these reasons, measures have been taken to progressively reduce the sulphur

329

content in diesel fuel worldwide. Currently, most industrialised countries use

diesel fuel with a sulphur level of 5 to 50 ppm. In addition, it should be noted

that sulphur compounds are also present in lubricating oils.

Several studies were conducted to investigate the effect of reducing the sulphur

content in diesel on particle number emissions. Bagley et al. (1996) found a

significant reduction of the number of nuclei-mode particles from a heavy duty

engine when the sulphur levels were reduced from 3200ppm to 100ppm.

Andersson et al. (2001) investigated three heavy duty vehicles using diesel fuel

of three sulphur levels: 340 ppm, 53 ppm, and less than 10 ppm, and showed

that, while changes in accumulation mode particles could be attributed to

changes in engine technology, the variation in nanoparticle number was more

likely influenced by fuel properties. The fuel containing the highest sulphur

content (340 ppm) showed the highest nanoparticle emissions for the ‘weighted

cycle’ (where each stage of the cycle was weighted according to the time it

contributed to the overall cycle), while the fuel with the lowest sulphur content

(<10ppm) shared the lowest. They also concluded that the influence of the fuel

sulphur content could not be fully decoupled from other chemical and physical

effects within the tested fuels, such as the total aromatic content which varied

between the fuels.

Ristovski et al. (2006) reported particle emissions from a fleet of twelve in-

service buses fuelled by 50 and 500 ppm sulphur diesel at four driving modes

on a chassis dynamometer and showed that particle number emission rates were

330

30-60% higher with the 500 ppm over the 50 ppm fuel. Most of the excess

particles were smaller than 50 nm and resided in the nucleation mode.

Kittelson et al. (2002) measured nanoparticle emissions from a diesel engine on

a dynamometer using fuels with three different levels of sulphur (1, 49 and 325

ppm) and two different lubricating oils (4000 ppm and 385 ppm sulphur). They

observed that for conventional lubricating oil (385 ppm) and both 1 ppm and 49

ppm sulphur fuel, there was no significant formation of a nucleation mode.

Increasing fuel sulphur to 325 ppm increased nanoparticle emissions, especially

at high engine load. Sulphate particles are formed at high temperature

conditions, such as at full engine load when more of the fuel sulphur is

converted to sulphuric acid. When present, most of the nucleation mode

particles were removed when passed through a thermodenuder, suggesting that

they were highly volatile. Other researchers have observed nucleation mode

particles even with very low sulphur levels (<10ppm) (Vaaraslahti et al. 2004)

suggesting that other components, such as unburned hydrocarbons, can have an

important role. At the low levels of sulphur in the fuel the amount of sulphur in

the lubricating oil can have a major influence. The most surprising result was

the large influence of specially formulated lubricating oil (Kittelson et al. 2002).

Contrary to expectations, low sulphur oil led to an increase in nanoparticle

formation in nearly all cases. It is possible that the increase in nanoparticle

formation when using low sulphur oil was related to the formulation of the oil

necessary to compensate for the removal of sulphur. It could also be due, in

part, to the release of volatile components from the oil, related to the lack of oil

break-in. Lubricating oil, unburned hydrocarbons from the fuel, as well as

331

PAH’s, could also play critical role in the formation of the nucleation mode

(Kittelson et al. 2002; Sakurai et al. 2003; Vaaraslahti et al. 2005; Ristovski et

al. 2006). On the other hand, Vaaraslahti et al (2005) have observed clear

correlation between the lubricating oil sulphur content and nanoparticle

formation only when the engine was equipped with a continuously regenerating

diesel particulate filter (CRDPF).

Alternative Fuels: Liquefied petroleum gas (LPG) is generally perceived to be a

cleaner fuel than unleaded petrol (Gamas et al. 1999). Ristovski et al. (2005)

found that particle number emissions from LPG cars were up to 70% less than

from similar unleaded petrol cars. Compressed natural gas (CNG) vehicles are

known to emit considerably lower particle mass than equivalent diesel vehicles.

However, there is considerable disagreement as to particle number emission

levels. This is due to the small number of measurements reported and the

difficulties in quantifying the effects of engine operating and testing conditions

and fuel and lubricating oil composition on secondary particle production. In

relation to buses, it has been shown that, in general, particle number emissions

from CNG buses are smaller than from diesel buses, but there are some

exceptions, particularly related to high engine load conditions where large

nuclei modes (<10 nm) and ultrafine particle number concentrations have been

observed (Holmen and Ayala 2002). In addition, the nuclei mode particles

observed at high loads are highly volatile (Meyer et al. 2006).

332

7.6. ROLE OF AFTER-TREATMENT DEVICES

There are two main particulate matter control technologies in use today. They

are oxidation catalysts and particle traps.

Oxidation Catalysts: While oxidation catalysts reduce the soluble organic

fraction (SOF), they have little effect on the soot or black carbon in the exhaust.

Still, some reduction in particle mass emissions is achieved through the removal

of the SOF. The maximum total particle mass reduction is dependent on the

magnitude of the SOF compared to the carbonaceous portion in the engine-out

exhaust, and is usually between 20% and 30% (Harayama 1992). The sulphate

fraction of diesel particles (SO4) is increased in the diesel oxidation catalyst, due

to the oxidation of SO2 with subsequent formation of sulphuric acid. Under

certain conditions, however, the SOF decrease can be more than off-set by an

increase of sulphate particle mass, leading to an overall increase in total particle

mass emission. In general, the effect of oxidation catalysts on particle number

emissions is often unpredictable. Total particle number concentrations,

especially in diesel emissions, are attributed primarily to nucleation mode

particles which are composed mostly of hydrocarbon and sulphuric acid

condensates. If the catalyst removes hydrocarbons (gas phase and SOF), it will

prevent their subsequent nucleation, thus reducing the particle number

concentrations. If, however, the catalyst produces sulphates, an effect more

prominent with high sulphur fuels, and more active noble metal catalysts, the

observed particle number concentrations may be higher than with no catalyst

due to sulphuric acid nucleation (Vogt et al. 2003).

333

Particle Traps: Particle traps are very effective in controlling the solid fraction

of exhaust particles, including elemental carbon (soot) and the related black

smoke emissions. The filtration efficiencies of some commercially available

diesel particle traps frequently exceed 90%. In light duty vehicles, Mohr et al

(2006) showed a high efficiency of diesel particle filters (DPFs) in curtailing

nonvolatile particle emissions over the entire size range. High emissions were

observed only during short periods of DPF regeneration and immediately

afterwards. However, they have limited effectiveness in controlling the non-

solid fractions of particle mass, such as the SOF or sulphate particles that occur

mainly in the liquid phase within the hot and humid emissions. For this reason,

trap systems designed to control the total particle mass emission are likely to

incorporate additional functional components targeting the SOF emission (e.g.,

oxidation catalysts). More recently, the introduction of ultra low sulphur diesel

has helped to improve the efficiency of abatement devices, many of which are

poisoned by sulphur. Emissions from diesel fuels containing sulphur levels of

less than about 12 ppm will not poison these devices. Very often, volatile

material pass through particle traps and nucleate to form nanoparticles that

increase the total particle number. To make matters worse, by retaining carbon

particles, the trap removes the material, which otherwise acts as a “sponge” for

condensates formed in the sampling system. Therefore there is a possibility that

in some cases particle traps can increase the formation of nanoparticles through

nucleation. In effect, particle traps reduce the numbers of solid agglomeration

mode particles by replacing them with liquid nucleation mode nanoparticles

(Burtscher 2001).

334

Vaaraslahti et al (2004) have observed that in heavy duty vehicles at high

loads nucleation mode particles form only when the engine is equipped with a

continuously regenerating diesel particulate filter (CRDPF). The tests were

conducted with two fuels of 2 and 40 ppm sulphur content and the nucleation

mode correlated with the sulphur level in the diesel fuel. In a later publication,

Vaaraslahti et al (2006) show that the formation of nucleation modes in heavy

duty engines with CRDPF is positively correlated not only with the fuel

sulphur level but also with the lubricant sulphur level, suggesting that

sulphuric compounds are the main nucleating species in this situation.

Formation of nucleation mode particles was also observed on a heavy duty

vehicle equipped with a continuously regenerating trap (CRT) during on-road

highway cruise conditions (Kittelson et al 2006). The CRT has reduced the

concentrations of accumulation mode particles to levels indistinguishable

from background while increasing the emissions of particles in the nucleation

mode. Similar to the observation by Vaaraslahti et al the increased emissions

of nanoparticles was observed at higher engine loads when the exhaust

temperature increased above about 300 °C therefore increasing the conversion

of SO2 emitted by the engine to SO3.

However, the toxicity of these particles has recently been brought into question.

Grose et al (2006) showed that nucleation mode particles emitted by a heavy

duty diesel engine equipped with a catalytic trap are composed mainly of

sulphates. This provides support for the argument that particulate emissions

from diesel vehicles equipped with advanced particulate control devices might

be less toxic than typical uncontrolled diesel emissions, which contain high

335

concentrations of organic compounds. However, due to the complexity of diesel

exhaust and the fact that sulphuric acid enhances polymerisation of organic

compounds, as well as solubilises metals, further toxicology studies are required

to evaluate the toxicity of these particles.

7.7. ROLE OF IONS

The mechanisms behind particle nucleation in the atmosphere have been

discussed in section 7.3 above, and in particular the role of binary homogeneous

nucleation of sulphuric acid and water or ternary homogeneous nucleation

involving sulphuric acid, water and ammonia. Theoretical models and

experimental observations show that binary homogeneous nucleation alone

cannot explain the observed formation and growth rates of particles in the

environment and, while ternary nucleation can explain observed nucleation rates

in urban areas, it does not assist in explaining the rates observed in other

environments that do not contain sufficiently high concentrations of sulphuric

acid (Weber et al. 1996; Clarke et al. 1998; Yu 2001; Kulmala 2003; Alam et al.

2003). However, neither mechanism is capable of explaining the observed

growth rates of ultrafine particles to cloud condensation nuclei (CCN) sizes. For

example, Weber et al. (1997) demonstrated that growth rates of nanoparticles

driven by binary and ternary nucleation is an order of magnitude too low to

explain the rapid appearance of fresh ultrafine aerosols during midday. Alam et

al (2003), while noting similar observations at urban sites showed that particle

formation by homogeneous nucleation occurred on approximately 5% of the

days studied and required condensable materials apart from sulphuric acid and

water, together with a relatively low pre-existing particle surface area.

336

An alternative mechanism of particle formation is ion-induced nucleation. It has

been shown that ion-induced nucleation occurs at a lower saturation ratio than

homogeneous nucleation (Hara et al. 1997, 1998). Homogeneous nucleation can

only occur spontaneously in highly supersaturated air. These conditions do not

occur naturally in the atmosphere. However, homogeneous nucleation is aided

by ions as gas molecules tend to condense and cluster around them. Yu and

Turco (2000; 2001) showed that charged molecular clusters can grow

significantly faster than neutral clusters and achieve stable observable sizes.

Thus, this mechanism can operate under conditions that are unfavourable for

binary or ternary nucleation. However, in a recent paper based on results from a

study conducted in Hyytiala, Southern Finland, Kulmala (2007) has argued that

ion-induced nucleation cannot explain the large number of neutral clusters that

were observed, suggesting that ternary, and not ion-induced, nucleation was

probably the dominant process taking place in this Boreal forest environment.

On the other hand, Yu and Turco (2008) used a global chemical transport model

to show that ion-induced nucleation was an important global source of

tropospheric aerosols. From these studies, it is clear that the relative importance

of ion-induced nucleation and neutral nucleation under varying atmospheric

conditions remains largely unresolved.

Environmental ions are formed naturally by cosmic rays at a rate of about 2 ion

pairs cm-3 s-1, while the main anthropogenic source of ions is combustion. The

ion concentration falls off rapidly with distance from the source due to

recombination. Ions generated from hydrocarbon flames play an important role

337

in the formation of nanoparticles (Yu 2001). Positively charged ions have been

detected in concentrations of up to 1.6 x 108 cm-3 in jet engine plumes (Arnold

et al. 2000) and these might play a key role in the formation of volatile particles

in the aircraft wake (Yu and Turco 1997). Also motor vehicle combustion is a

significant source of ions that may play an important role in the formation of

nanoparticles via ion-induced nucleation during the dilution and cooling of the

hot emissions in air, especially in urban environments. Kittelson et al. (1986)

monitored electric charges present on diesel emission particles and they showed

approximately equal numbers of positively and negatively charged particles

with 1-5 units of elementary charge per particle. The charge distribution with

respect to size followed a Boltzmann equilibrium relationship equivalent to

1500K.

Yu et al. (2004) measured the ionic emissions from a petrol car and a diesel

generator engine. They found that the total ion concentrations from the two

engines were about 3.3 x 106 cm-3 and greater than 2.7 x 108 cm-3, respectively.

Maricq (2006) studied the electric charge of particles in petrol and diesel vehicle

exhaust using single and tandem differential mobility analysis. This method

provided the means to sort the particles according to both their size and charge.

About 60-80% of the particles were charged but with nearly equal numbers of

positive and negative charge, leaving the exhaust electrically neutral. Charge

increased with particle size, up to about ±4 units of charge per particle. At a

fixed particle size, charge per particle followed a Boltzmann distribution with

temperature range 800-1100 K. Jung and Kittelson (2005) used an electrostatic

filter and an SMPS to examine the charged fraction of diesel particles as a

338

function of their size. They showed that the diesel nanoparticles carried very

little charge, while there was a large charged fraction of 60-80%, in the

accumulation mode. In summary, these results show that vehicles emit ions of

both signs, with the majority of the charges being carried on the larger particles.

7.8. ROAD-TYRE INTERFACE

So far, tyre wear on the road has been considered to contribute mainly larger

size particles (>10 µm) in the air (Pierson and Brachaczek 1974). A recent

study by Dahl et al. (2006) showed, however, that road–tire interface can also be

a source of sub-micrometer particles. The study conducted in a road simulator

showed that the mean particle number diameters were between 15–50 nm. The

emission factor increased with increasing vehicle speed, and varied between

3.7×1011 and 3.2×1012 particles vehicle−1 km−1 at speeds of 50 and 70 km h−1,

which corresponds to between 0.1–1% of tail-pipe emissions in real-world

emission studies. The authors hypothesised that the particles may originate

from three components of the tires: (i) the carbon black reinforcing filler, (ii)

small inclusions of excess ZnO or ZnS (ZnO is an activator for organic

accelerators that are used to speed up the vulcanization process), (iii) the oils

used as softening fillers. The authors suggested that since speed determines the

amount of mechanical stress in the tire material it also determines the

temperature in the tire, and increased temperature in turn leads to increased

emissions of loosely bound reinforcing filler material and evaporation of semi-

volatile softening oils. Clearly more research is needed on this topic.

339

7.9. EMISSION FACTORS AND EMISSION INVENTORIES

Emission Factors: An emission factor is the amount of pollutant emitted either

when an activity is performed, for example, a vehicle drives a kilometre; or the

amount of pollutant emitted per unit of fuel burned. Emission factor values are

used for developing inventories for gaseous or particulate motor vehicle

emissions, however in order to derive them many issues need to be considered

and resolved. In particular, they depend on motor vehicle type, fuel used,

engine load, after-treatment devices fitted, road type, travel speed, road grade

and local meteorological conditions. Current methods for deriving emission

factor values range from measurement of single vehicles, to vehicle fleets using

direct methods, such as measurements on a dynamometer, on or near roadways

or in tunnels, or indirect methods such as estimates based on remote sensing or

fuel consumption, particularly in relation to UF particles. In addition, a wide

range of different instrumentation are used that measure different size ranges (as

discussed in section 7.2). As a result, there are a lot of different values of

emission factors published, on different types of measurements, in different

parts of the world. This leads to the question as to which values should be used

in quantifying and modelling traffic emissions.

Statistical analysis of particle number emission factors around the world: A

detailed analysis by Keogh et al. (2007) that included statistical analysis of more

than 160 particle number emission factors relating to motor vehicle tailpipe

emissions revealed that emission factor values estimated from CPC

measurements produced the highest mean values for Fleet, Heavy Duty Vehicle

(HDV) and Light Duty Vehicles (LDV) of 7.26, 65 and 3.63 x 1014 particles per

340

vehicle per kilometre respectively. The mean Diesel Bus emission factor value

relating to measurements using the SMPS was found to be 3.08 x 1014 particles

per vehicle per kilometre.

The review found that there is a significant difference between the means for

measured particle number for CPC and SMPS instrumentation (23 and 2 x 1014

particles per vehicle per kilometre respectively); but no significant difference

between the means of particle number emission factors for studies conducted in

different countries (Australia, Austria, Germany, Sweden, Switzerland, United

Kingdom, USA), nor between those for studies conducted on a dynamometer, or

near roadways. Particle number emission factors for HDV were found to be

significantly higher than the corresponding values from Fleet and LDV.

The range of particle number emission factors reported in four studies, that

measured nanoparticle and ultrafine subclasses using the SMPS and DMPS

(Gidhagen et al. 2003; Imhof et al. 2005a; Imhof et al. 2005b; Jones and

Harrison 2006), are summarised in Table 7.1 below, where the range given

represents the sum of the ranges taken from the four studies. Examination of this

table highlights the importance of measuring size ranges < 18nm, where particle

numbers tend to be more prolific, and in which the emission factor values are

generally larger than those estimated for 18-50nm and 18-100nm.

341

Table 7.1. The range of particle number emission factors reported for nano and ultrafine size ranges

Particle size classification

Size range measured, nm

Fleet 1014 particles

per vehicle per km

Heavy Duty Vehicles

1014 particles per vehicle per km

Light Duty Vehicles

1014 particles per vehicle per km

Nanoparticles

< 10

--

14.5

0.63-4.14

10-30 -- 2.14-37.8 0.067-4.87

18-50 0.59-1.31 1.55-8.2 0.13-0.56

Ultrafine 18-100 0.84-1.55 1.7-10.5 0.37-0.81

30-100 -- 3.19 0.284

Inventories of motor vehicle particle emissions: Estimates of emission

inventories for particle number concentration are not available (Jones and

Harrison 2006); nor does there exist a comprehensive inventory of vehicle

particle emissions covering the full size range emitted by motor vehicles. The

one reported inventory is the assessment conducted by Airborne Particles Expert

Group (1999). For this assessment emission trends for the years 1970 to 1996

including inventories for PM2.5, PM1 and PM0.1 were estimated based on PM10

UK monitoring data, using mass fractions in this size range available for

different emission sources and fuel types and 33 particle number size

distribution spectra. It was shown that in all size fractions, vehicle emissions are

the major contributor, compared with all other combustion and non-combustion

sources in urban areas. With decreasing particle size, the contribution of road

transport to the total emissions increases and for PM0.1 reaches 60%.

Contributions from other combustion sources tend to decrease with decreasing

particle size. One of the conclusions from the data presented in the report is that

342

there was a significant decrease in emissions in the PM10 and PM2.5 ranges

during the investigated period of time, less in the PM1 range and very little in

the PM0.1 range. This could be related to the lack of strategies for decreasing

emissions of the UF particles. More effort is needed towards compilation of

vehicle emission inventories for UF particles.

7.10. TRANSPORT OF PARTICLES WITHIN URBAN SCALE AND

AMBIENT PROCESSING

Vehicle emissions are highly dynamic and consist of reactive mixtures of hot

gases and particles. The main factor determining the speed and direction of the

pollutant plume away from the emission site is generally the prevalent wind.

Other contributing factors include the initial speed of the pollutants emitted

from vehicles, turbulence caused by vehicle motion, location of the exhaust,

precipitation and the topography of the area. The pollutant plume undergoes

dilution with ambient air and is subject to a range of physical and chemical

processes, which change its chemical composition, physical characteristics and

concentration in the air during the transport process. Also, soon after emission,

when the pollutant plume is still concentrated, is the most likely period for

secondary particle formation by nucleation involving precursors present in the

emission plume, as discussed in section 7.4. Therefore particles measured away

from the emission site, and some time after emission, have different

characteristics to those measured immediately after formation.

343

7.10.1. Role of meteorological factors on particle concentration

The most important factors to consider include wind speed, precipitation,

relative humidity and temperature.

Wind speed. Wind speed affects dispersion and dilution and thus atmospheric

mixing, but also resuspention of particles. Several studies shown that UF

particle concentration decreases, while concentrations of larger particles

displays a “U shape” relationship with wind speed (Harrison et al. 2001;

Ruuskanen et al. 2001; Molnar et al. 2002; Charron and Harrison 2003).

Hussein et al. (2005a) showed that UF particle number concentrations are best

represented by a decreasing exponential function, with the minimum observed

during wind speeds >5ms−1, as a result of a higher coagulation rate, better air

mixing, and more particle losses due to deposition and scavenging at these wind

speeds. Particles larger than 100 nm showed a “U-shape” relationship, best

represented by a second-order polynomial, with the minimum during wind

speeds between 5–10 m s−1. Similar results were found by Charron and Harrison

(2003) as to the trends for particles ranging from 30 to 100 nm and from 100 to

450 nm, as well as a twofold decrease from the weaker to the stronger winds. A

decrease of about 10,000 normalised counts per cm3 was seen for particles in the

range from 30 to 100 nm and modal shift toward smaller values with increasing

wind speed. However, no obvious relationship with the wind speed was seen for

the particles ranging from 11 to 30 nm, and thus no dilution effect was evident

for this particle range, which could be explained as characteristic of new particle

formation. A study by Hussein et al. (2005a) showed another trend, namely that

at some sites, particle number concentrations displayed a linear decrease with

344

wind speed. This was explained by higher summer temperatures, which are

accompanied by a high boundary layer and thus mixing of aerosol particles

within a bigger volume. Under such conditions, the relative changes of particle

concentration with wind speed are smaller compared to a shallow boundary

layer and thus smaller volume of air mixing.

Precipitation. In general, precipitation has been shown to have a washout effect,

which means removing particles from the atmosphere e.g. (Garcia-Nieto et al.

1994; Morawska et al. 2004) (Jamriska et al. 2007). However, a study by

Charron and Harrison (2003) showed an opposite effect in relation to particles

below 150 nm, namely an increase of particle numbers during rain, with larger

rain drops (more than 0.4 mm) leading to higher particle numbers than smaller

ones (0.2 mm). In addition, the highest particle numbers were measured just

after a rain event (1 h after). The possible explanation for this phenomenon is an

effect of reduced temperatures during precipitation events and thus higher

saturation ratio of semi volatile species combined with low pre-existing surface

area of particles, both favouring new particle formation, and thus a significant

increase of particle number concentration.

Relative humidity and temperature. These two parameters commonly display

diurnal anti-correlation, with the increased temperature during the day

accompanied by a decreased relative humidity. In general, both temperature and

relative humidity play a role in UF particle number concentration (Charron and

Harrison, 2003; Ruuskanen et al. 2001; Jamriska et al. 2007). Kim et al. (2002)

showed that during the warmer months, there was some increase in particles

345

smaller than 100 nm, in the afternoon, linked to an increase in temperature.

Charron and Harrison (2003) showed that particles in the size range 11-30 nm in

a roadside environment peaked during the early morning showing an inverse

association with air temperature. Olivares et al (2007) found a distinctive

dependence of particle number concentration with ambient temperature in a

street canyon in Sweden. They found that the total particle number more than

doubled when the temperature decreased from 15°C to -15°C. The variation was

most pronounced for particles smaller than 40 nm. Modelling results predicted

that the changes in the particle sizes observed were consistent with the

condensation of volatile compounds onto pre-existing aerosols. They also

showed that nucleation mode particles were largely influenced by relative

humidity with high concentrations during high humidity periods. Hussein et al.

(2005a) found that the high number concentration of particles larger than

100 nm during the higher summer temperatures was partly due to the growth of

aerosol particles in the presence of condensable vapours emitted from the

surrounding boreal forest in southern Finland. In general, higher atmospheric

water content is expected to favour homogeneous binary nucleation of sulphuric

acid and water (Easter and Peters 1994), while ternary nucleation involving

ammonia (Korhonen et al. 1999), similarly to nucleation from organic

compounds, is expected to be independent of relative humidity Therefore, in the

study by Charron and Harrison (2003) for example, the lack of a dependence on

the relative humidity during the daytime was considered indicative that the

binary nucleation from sulphuric acid and water was not a major factor in

particle production.

346

Temperature inversion. Under such conditions, there is little vertical mixing,

wind speeds are lower and thus pollution concentration in general is expected to

increase. For example, Janhall et al. (2006) showed that morning temperature

inversion in Göteborg, Sweden, resulted in significantly elevated concentrations

of traffic-related pollutants, including UF particles. Mean particle number

concentrations on the days with and without morning inversions were 6500 ±

4800 and 2800 ± 1900 particles cm−3, respectively. However, there was no

impact of inversion on PM10 concentrations.

7.10.2. Relative role of various processes

The processes that dominate particle dynamics shortly after emissions by

vehicles (i.e. when the concentration of the emission plume, while decreasing

due to dilution and mixing, is still high) include: condensation/evaporation,

coagulation and new particle formation. Several studies investigated the relative

importance of these processes under various meteorological and pollution

concentration conditions. During the dilution, when the initially hot mixture of

pollutants is cooled down, the saturation ratio of gaseous compounds of low

volatility reaches a maximum. This is when two of the above processes are

possible: new particles formation by nucleation of vapours, and vapour

condensation onto existing particles. The availability of pre-existing particle

surface area for the condensation of the semi-volatile vapours along with the

dilution rate, since it governs the cooling rate, will determine which of the two

processes dominates. Small aerosol concentrations favour new particles

formation and their growth to larger sizes (Kulmala et al. 2000), while high

concentrations promote the condensation of the vapours on the existing particles

347

and disfavour new particle formation (Kerminen et al. 2001). This is the reason

why cleaner air resulting from stronger winds and rain, as discussed above,

favour the occurrence of high numbers of UF particles. For example, Charron

and Harrison (2003) showed that large amounts of semi-volatile vapours from

vehicle exhausts in the early morning, associated with low pre-existing particle

surface area at that time (from about 300 to 500 μm2/cm3) favour production of

new particles and their growth to detectable sizes (>11 nm). However, during

daytime, when the particle surface area ranges from 800 to 1100 μm2/cm3,

condensation of the condensable gases onto existing particles is likely to

dominate. In comparison to the processes discussed above, coagulation appears

to be overall a less important process, due to short time available for it to be

effective before the high initial concentrations at the road are diluted. Analysis

conducted by Shi et al (1999) to estimate the expected effect of coagulation on

the decrease of particle number concentration showed that the concentrations

between the road and a site 100m away from the road would decrease by less

than 11% due to coagulation, compared to a 72% decrease in measured

concentrations, implying that dilution with background air is the main

mechanism for the rapid decrease in particle number concentration. However,

Zhu et al. (2002a) concluded that both atmospheric dilution and coagulation

play important roles in the rapid decrease of particle number concentration and

the change in particle size distribution with distance away from a freeway. The

study by Zhu et al. (2002b) suggested that coagulation is more important than

atmospheric dilution for ultrafine particles and the reverse is true for large

particles. This contradicts some earlier studies which concluded that the rapid

dilution of the exhaust plume made coagulation insignificant (Vignati et al.

348

1999; Shi et al. 2001a), with the possible reason for this being that the earlier

studies assumed a much lower particle number concentration for particles

smaller than 15 nm, while this study accurately measured freshly emitted

particles down to 6 nm.

Zhang and Wexler (2004a) identified two distinct dilution stages after emission.

The first stage, termed tailpipe-to-road, was induced by traffic-generated

turbulence and occurred soon after emission, lasting about 1-3 s, when the

dilution ratio reached up to 1000. The second stage was mainly dependent on

atmospheric turbulence and lasted 3-10 min, with an additional dilution ratio of

about 10. In the first stage, aerosol dynamic processes such as nucleation,

condensation and coagulation played major roles. In the second stage,

condensation was the dominant mechanism in altering the aerosol size

distribution, with coagulation and deposition playing minor roles. Exhaust

plumes emitted by different types of engines maintained their characteristics in

the first stage but generally mixed with each other in the second stage. A similar

conclusion was reached by Pohjola et al. (2003) based on the application of a

aerosol process model MONO32. The effect of coagulation was substantial

only if the dilution of the exhaust plume was neglected, which is not realistic

under most conditions (unless during temperature inversions or a very stable

atmosphere). Condensation of an insoluble organic vapor was shown to be

important if its concentration exceeds a threshold value of 1010 or 1011 cm-3 for

the Aitken (~50nm) and accumulation (>100nm) mode particles, respectively.

The importance of condensation or evaporation of water was shown to be

strongly dependent on the hygroscopicity of particles. The modeling showed

349

that after a time of 25s, most of the particulate matter transformation processes

have already taken place.

It has been suggested that sulphuric acid induced nucleation was the dominant

secondary particle production mechanism in the first stage, followed by the

condensation of organic compounds (Kittelson 1998; Maricq et al. 2002; Zhang

and Wexler 2004a). Schneider et al. (2005) showed that nucleation was mainly

due to sulphuric acid and water and low volatile organic species condensed only

on pre-existing sulphuric acid/water clusters. However, other studies have

suggested that the volatile component of total diesel emission particles was

comprised mainly of unburned lubricating oil (Tobias et al. 2001; Sakurai et al.

2003). Recently, Meyer and Ristovski (2007) have shown that ternary nucleation

involving ammonia as the third species, is the main nucleating mechanism,

followed by the condensation of volatile organic components. Zhang and Wexler

(2004a) showed that the first stage of dilution was crucial for the activation of

nuclei mode particles due to the high concentration of condensable material

during this time.

A deeper insight into the role of the key process was provided by Zhang and

Wexler (2004b) whose modelling showed that for particles larger than 0.05 μm,

coagulation is too slow to influence number distributions and condensation is

the leading process. However, for particles smaller than this, under typical urban

conditions, condensation and evaporation, coagulation, nucleation and

emissions interact with each other and the Kelvin effect must be considered in

modelling. Gravitational settling was shown to significantly affect particle dry

350

deposition, but negligible for vertical turbulent transport; chemical reactions are

negligible. It was noted that the relative importance of different mechanisms

remains about the same during the day and night time. During the night time,

with photochemistry cut off and significantly decreased emissions,

concentrations of both volatile or condensable gases and particles is lower

leading to the increased time scales for the process, but thus maintaining their

relative importance.

In addition to particle growth and dilution, Jacobson et al (2005) showed that

that small (< 15 nm) liquid nanoparticles emit semi volatile organics (< C-24)

almost immediately upon emission and that the shrinking of these particles

enhances their rates of coagulation by over an order of magnitude. Enhanced

coagulation in isolated emission plumes may also affect evolution of particle

size distribution. Importantly, they concluded that neither condensation,

complete evaporation, coagulation alone, nor preferential small-particle dilution

appears to explain the evolution of particle sizes in the vicinity of busy roads.

351

7.11. PARTICLE SIZE DISTRIBUTIONS AND MODAL LOCATION

IN URBAN ENVIRONMENTS

Particle number size distributions in vehicle emissions often show a

characteristic bimodal distribution (Kittelson 1998; Kwon et al. 2003; Vogt et

al. 2003; Ristovski et al. 2006). Most of the particulate mass is generally found

in the accumulation mode (> 50nm) while a considerable number fraction of the

particles may occur in the nucleation mode (< 30nm). Bimodal size distributions

have also been observed near busy roads (Pirjola et al. 2006). Distinct ultrafine

modes in the particle number distributions have been found in the size range 10-

20 nm on the downwind side but not on the upwind side of busy freeways (Zhu

et al. 2002a; Rosenbohm et al. 2005).

A recent literature review of modal locations identified in ambient particle size

distributions in a range of worldwide environments (34 studies) found that for

particle number modal location values spanned 0.006 to 30 µm, with

approximately 98% of these values being ≤ 1 µm (Morawska et al. 2007a).

Anthropogenic-influenced environments included those in suburban

environments in Australia and Finland (Morawska et al. 1999c; Hussein et al.

2005b), urban in Australia, Finland, Germany, Hungary, India and the USA

(Morawska et al. 1999c; Salma et al. 2002; Wehner et al. 2002; Wiedensohler et

al. 2002; Fine et al. 2004; Hussein et al. 2004; Hussein et al. 2005b; Monkkonen

et al. 2005); and traffic-influenced environments in Australia, Finland and

Germany and the US (Morawska et al. 1999c; Zhu et al. 2002a; Zhu et al.

352

2002b; Pirjola et al. 2004; Zhu et al. 2004; Rosenbohm et al. 2005; Zhu et al.

2006).

When considering modes identified in these studies in the ≤ 50nm size range it

was found that in anthropogenic-influenced environments they ranged from 8.2-

50nm for suburban and urban, and included modes at 7,10, 13, 15, 16, 19, 20,

27 and 30nm in traffic influenced, likely reflecting the influence of motor

vehicle emissions. Modes identified in the > 50 ≤ 100nm size range in the

anthropogenic-influenced environments were in the range 50.2-65nm in

suburban, traffic and urban; and at around 80nm in urban and traffic. Additional

modes were found in suburban in the range 92-100nm.

7.12. CHEMICAL COMPOSITION OF ULTRAFINE PARTICLES IN DIFFERENT ENVIRONMENTS

There have been relatively few studies reported, which investigated any aspects

of chemical composition of UF particles in ambient air. Moreover, each study

was conducted in a different way, sampled particles in a different size range,

and focused on different aspects of particle chemical composition. Therefore, at

present a comprehensive knowledge on UF particle composition in different

environments is not available. This review summarises the results of the few

studies reporting UF particle chemical composition.

As discussed earlier, engine emissions include SO2 or SO3 and NO (later

converted to NO2). Transformations in the air of SO2, NO2 and NH3 from

vehicle emissions into SO42−, NO3

−, and NH4+ is important for increasing

353

secondary aerosol formation near traffic sites. Three studies investigated these

aspects of UF particle chemistry in urban environments. (Kuhn et al. 2005)

investigated particle chemical composition in a park 5 km downwind of

downtown Pittsburgh. The study showed that UF mass was about 0.6 μg m−3

and its summer composition was 45% organic matter and 40% salts of

ammonium and sulphate, compared to 55% organic matter and 35% of the sum

of ammonium and sulphate during winter. This shift was explained as being

likely due to higher summertime levels of photochemical activity for oxidation

of SO2, and increased wood burning and vehicular organic contributions in

wintertime. UF chemical composition was also studied by Sardar et al. (2005)

and showed a significant seasonal change from summer to winter. Lin et al.

(2007) investigated water-soluble ions in nano (PM0.01–0.056) and ultrafine

(PM0.01–0.1) size ranges in samples collected near a busy road and at a rural site.

Several conclusions were derived from this study. It was shown that the

primary, secondary and tertiary peaks of SO42− and NH4

+ were in the fine (0.56–

1.0 μm), coarse (3.2–5.6 μm), and nano (0.032–0.056 μm) size ranges,

respectively. The second and third peaks of NO3− were in the same sizes ranges

as those for SO42− and NH4

+; however, the primary peak for NO3− was in the

size range of 1–1.8 μm and slightly larger than that of SO42− and NH4

+. Nano

(PM0.01-0.056) and coarse (PM2.5-10) particles exhibited the highest (16.3%) and

lowest (8.37%) ratio of nitrate mass to total particle mass, respectively. The

mass ratio of NO3− was higher than that of SO4

2− (contrary to the trend

commonly observed for urban atmospheric particles) and also more variation

existed in NO3− concentrations for different size range particles at the roadside

site than at the rural site. For both sites, NO3− concentrations in nano and UF

354

particles were about two times higher than that for coarse particles and the nano

NO3− concentration at the roadside site was 1.34 times that at the rural site. The

peak of NO3− in the ultra-fine size range in this study was attributed mainly to

vehicle engines, while the peaks of SO42− and NH4

+ in the nano size range to

(NH4)2SO4 aerosols formed via the interaction of NH4+ and SO4

2−. At the two

sites, no significant difference in NH4+ concentration was found in each of the

different size ranges investigated. However, at the rural site, NH4+ was highly

correlated with SO42− in each of the three particle size ranges - nano, ultrafine,

and coarse, with R values of 0.89, 0.60, and 0.86, respectively.

Somewhat different aspects of UF particle chemistry were studies by Woo et al.

(2001a) who investigated evaluation of outdoor and indoor particle volatility

near a freeway by heating particles and detecting changes in their diameters and

number concentrations. The study showed that aerosol volatility decreases with

increasing distance from the source (in this case the differences between

outdoor and indoor particles). Monodisperse distribution was observed for 18

and 27 nm particles with the mode diameter decreasing with increasing heater

temperature (thus suggesting that all of these particles are composed mostly of

volatile material), and also broadening because not all particles shrink to the

same degree due to differences in their chemical composition. Bi-modal

distributions were observed for outdoor 45 and 90 nm particles heated to 1100C.

In particular the 90 nm mode split into two at heater temperatures of 900C, with

one mode remaining close to the monodisperse mode of the unheated aerosol,

while the other shifting to a lower diameter. This bi-modal distribution indicates

that a fraction of the 90 nm aerosol consists of particles that are composed of

355

almost entirely non-volatile material and therefore does not change size upon

heating, and the remaining fraction contains mostly volatile material and

continue to shrink in size with increasing temperature. For temperatures of up to

1300C the diameter of the volatile particles in the whole range investigating

(18–90 nm) was still shrinking (at this temperature decreased to about half the

diameter compared to the original size at ambient temperature) without a clear

evidence of a plateau which would indicating the presence of a non-volatile

core; thus if a non-volatile core exists, its diameter will be smaller than that

reached by particles at 1300C.

Elemental Carbon (EC) is usually considered a marker of the combustion

process; with diesel engines as predominant sources of EC to the urban

atmosphere. Sardar et al. (2005) investigated chemical composition of UF

particles at two sites (urban and inland). The study showed that organic carbon

(OC) ranged from 32 to 69%, EC from 1 to 34%, sulphate from 0 to 24% and

nitrate from 0 to 4%. This was somewhat different to the findings from the

study by Kim et al. (2002) who showed that OC and EC contribute 35%,

sulphate 33%, and nitrate and ammonium 6% and 14%, respectively, with other

unknown substances contributing 12%. Kim et al. (2002) also showed that in all

cases, a greater fraction of the total mass consists of OC in the UF mode than in

the corresponding accumulation mode. Similar conclusions can also be drawn

for EC. A distinct OC mode was observed between 18 and 56 nm in the summer

and not present in the other seasons, indicating photochemical secondary

organic aerosol formation. The EC levels were higher in winter at the source

sites and in summer at the receptor sites due to lower inversion heights and

356

increased long-range transport from upwind source areas, respectively. Nitrate

was more prevalent in the accumulation mode and almost not measurable in the

UF particles. This was related to the process of formation of ammonium nitrate

when the nitric acid in photochemical smog in air parcels originating near

downtown Los Angeles Basin encounters ammonia. Higher summertime

temperatures and lower relative humidity favour the dissociation of particulate

ammonium nitrate, which is more pronounced for UF particles due to the Kelvin

effect. Sulphate, which similar to nitrate, was only detectable in the size ranges

above 56 nm made up greater percentage of the total mass in the accumulation

than the UF mode in fall and winter, however a higher proportion in both modes

during the summer, likely due to its photochemical origin. It was suggested that

the absence of sulphate in the smaller particles indicted that the majority of the

particulate sulphate is formed on pre-existing particle surfaces or by liquid-

phase reactions of sulphuric acid. The study by Kim et al. (2002) also supported

the general findings from an earlier study conducted in the same area by Turpin

and Hutzincker (1995), who also investigated the significance of secondary

organic aerosol formation by plotting OC against the EC concentrations. The

average ratio of OC to EC at a site affected directly by traffic and at a site

downwind from that one estimated from slope of the linear regression, were 3.5

and 8.6, respectively implying the existence of secondary organic aerosols in

UF particles at second site.

UF chemistry, including elemental composition was investigated by Pakkanen

et al. (2001), including over 40 chemical components of samples from an urban

and rural site in the Helsinki area. While the average UF particle mass

357

concentration was higher at the rural (520 ng/m3) than at the urban site (490

ng/m3), the average chemical composition of UF particles was similar at the two

sites. The most abundant of the measured components were sulphate (32 and 40

ng/m3 for the urban and rural sites, respectively), ammonium (22 and 25

ng/m3), nitrate (4 and 11 ng/m3) and the Ca2+ ion (5 and 7ng/m3). The most

important metals at both sites were Ca, Na, Fe, K and Zn with concentrations

between 0.7 and 5 ng/m3. Of the heavy metals, Ni, V, Cu, and Pb were

important with average ultrafine concentrations between about 0.1 and 0.2

ng/m3. Also the organic anions oxalate (urban 2.1 ng/m3 and rural 1.9 ng/m3)

and methanesulphonate (1.3 and 1.7 ng/m3) contributed similarly at both sites.

The measured species accounted for only about 15-20% of the total UF mass.

While not measured, it was estimated that the amount of water was about 10%

(50ng/m3) and that of carbonaceous material about 70% (350 ng/m3) at both

sites. At both sites the contribution of UF to fine was especially high for Se, Ag,

B, and Ni (10-20%) and at the rural site also for Co (20%), Ca2+ (16%) and Mo

(11%). Enrichment in the UF particles suggests that local sources may exist for

these elements. Aitken modes turned out to be useful indicators of local sources

for several components. The Aitken modes of Ba, Ca, Mg and Sr were similar in

several samples, suggesting a common local combustion source for these

elements, possibly traffic exhaust. Co, Fe, Mo and Ni formed another group of

elements often having similar Aitken modes, the likely source being combustion

of heavy fuel oil.

In summary, it can be concluded from the review of the studies which analyse

UF particle composition that almost all the studies focused on different aspects

358

of particle chemistry, with some of them targeting ion or elemental composition,

other studies particle volatility, and yet others investigating elemental and

organic carbon fractions. It is therefore important that more studies are

conducted on particle chemical composition, which would investigate in parallel

a range of different aspects to provide a more complete understanding on the

chemistry of UF particles and its local variation.

7.13. TEMPORAL VARIATION OF PARTICLE CHARACTERISTICS

7.13.1. Diurnal variation

In urban environments, strong diurnal variation of particle concentration has

been reported by many studies and shown to closely follow the temporal

variation in traffic density, with the highest levels observed on weekdays during

rush hours (Ruuskanen et al. 2001; McMurry and Woo 2002; Charron and

Harrison 2003; Paatero et al. 2005; Morawska et al. 2007c). It has been reported

that there are differences in the pattern for different weekdays, with for example

Friday having higher concentrations, and also between Saturday and Sunday,

reflecting different traffic flowrate patterns on different days (Hussein et al.

2004). As shown by many studies (e.g. (Morawska et al. 2002; Hussein et al.

2004)), the daily pattern of aerosols on weekdays is characterized by two peaks

coinciding with the traffic rush hours, while on weekends, by a wide peak in the

middle of the day. It is, however, the concentration of the nucleation and Aitken

modes particles which follows this trend, with far fewer particles in the

accumulation mode, indicating exhaust as a major source of UF, but not of

larger particles in urban air (Hussein et al. 2004; Hussein et al. 2005a). Stanier

359

et al. (2004a) showed that strong diurnal trends were most apparent in the

aerosol number, in contrast to aerosol volume distribution, for which seasonal

trends were the strongest. In addition to the impact of traffic emissions, the

process of secondary particle formation contributes to the variation in daily

pattern of aerosol concentration, with the impact varying during different

seasons.

7.13.2. Seasonal variation

There are several factors contributing to seasonal variation in particle

concentration, including those leading to an increase, such as: lower mixing

layer height and greater atmospheric stability in winter (due to less dispersion),

lower winter temperature (increased nucleation events of combustion exhaust

emitted from motor vehicles particularly during morning rush), and increased

photochemical particle formation during summer; as well as those leading to a

decrease, such as: lower traffic flow rate during summer holiday periods. All the

studies investigating seasonal variation in particle concentration in the Northern

hemisphere showed that there are clear seasonal trends (Zhang et al. 2004a;

Zhang et al. 2004b; Paatero et al. 2005; Pirjola et al. 2006; Virtanen et al. 2006),

contrary to a study conducted in the Southern Chemosphere, in Brisbane,

Australia, which did not show a trend (Mejia et al. 2007b). While the studies

conducted in the Northern Chemosphere were in areas where there are

significant meteorological differences between the seasons, and thus seasonal

variations in human activities, in subtropical Brisbane there are much smaller

differences between the seasons.

360

With the number of factors affecting particle concentrations it is not surprising

that there are differences in the magnitude and time of occurrence of peaks and

troughs in the concentrations between different geographical locations. Most of

the studies reported the lowest total number concentrations in summer with the

highest in winter but sometimes also in other seasons. Pirjola et al (2006) and

Virtanen et al. (2006) showed that the average concentrations in winter in

Finland were 2–3 times higher, with the highest of 183 000 cm−3 observed in

February. Similar results were reported by Wehner and Wiedensohler (2003) in

Leipzig, Germany. Zhang et al. (2004b) found the highest monthly mean for

N11-50 particles for Pittsburgh, USA, in December, with a mean of 7630 ± 3710

cm-3 while lowest of 4280 ± 2250 cm-3 in July. In winter most of the factors

leading to the increase in particle concentration tend to occur at the same time:

morning traffic rush, when the mixing height is the lowest, coinciding with the

lowest wind speed and temperature. However, Hussein et al. (2004) and Laakso

et al. (2003) showed that while the lowest total and nucleation mode particle

number concentrations were observed in Helsinki and in northern Finland

during summer, the highest were during spring and autumn, and the cleaner the

air, the stronger the cycle was. Summer minima were associated with higher

temperatures (which limit the nucleation due to temperature dependence of

saturation vapour pressures) and better mixing, and in Helsinki, with less traffic

in July. Springtime maximum was associated with nucleation of exhaust gases

(favoured by low temperatures combined with the low boundary layer height

and high radiation) and transport of new particles from cleaner areas.

361

In general, seasonal daily variation in the accumulation mode particles is less or

not significant, while the variation was larger in the UF particles; especially

nucleation mode particles (Hussein et al. 2004). Wehner and Wiedensohler

(2003) showed that the maximum concentration for both summer and winter

occurred for a particle size between 10 and 20 nm. McMurry and Woo (2002)

studied ambient aerosols in urban Atlanta, Georgia, and found that for particles

between 10 and 100 nm, average concentrations tended to be highest during

winter, while concentrations of particles in the 3-10 nm range increased in the

summer due to photochemical nucleation, which depends strongly on the

intensity of solar radiation. Evidence of summer nucleation was also found by

Sardar et al. (2005) and Geller et al. (2002), who showed elevated levels in the

mass concentrations in the 32-56 nm size range. Summer nucleation events lead

to rapid changes in particle concentrations and as Zhang et al. (2004b) showed,

the highest variation was found in July.

7.13.3. Long term variation

The number of studies investigating long term trends in particle concentrations

is limited. Some insight into particle number concentration trends was provided

by the studies conducted at different locations in former East Germany,

including the city of Erfurt and the counties of Bitterfeld, Hettstedt and Zerbst,

during two different campaigns, the first one in the early and the second one in

the late 1990’s, which found that UF particle number increased between 38.1%

(Pitz et al. 2001) and 115% (Ebelt et al. 2001). Wahlin et al (2001) measured

particle number concentration and size distributions in a street canyon in

Copenhagen over two 2-month campaigns during Jan-Mar 1999 and Jan-Mar

362

2000 and found a significant decrease in ultrafine particle concentration. This

observation was attributed to a 56% fall in the average particle emission from

diesel vehicles due to the reduction of the sulphur content in the fuel from

approximately 0.05% to less than 0.005% implemented in Denmark in July

1999. A handful of investigations have continuously monitored submicrometer

particles for periods of at least one year but no evidence on the long-term trends

was reported (Morawska et al. 1998b; Woo et al. 2001a; Wehner and

Wiedensohler 2003; Cabada et al. 2004; Paatero et al. 2005; Watson et al.

2005). Exceptions to this are two studies conducted in Finland and in Australia.

Hussein et al. (2004), who measured particle number for over a six-year period

in Helsinki, found that annual geometric mean particle number concentration

increased by 3.2% in 1999, followed by a decrease of 6.7% in 2000, and 17.6%

in 2002. Although the monitoring site was moved after the first three years 3km

from its original location, thereby influencing the results, their main conclusion

was that the annual variation in total particle number was associated with traffic

density and the predominance of new vehicles in the Helsinki area. Long trends

were investigated over a five-year period in Brisbane, Australia. Particle size

distribution was summarized by total number concentration and number median

diameter (NMD) as well as the number concentration of the 0.015-0.030 (N15-

30), 0.030-0.050 (N30-50), 0.050-0.100 (N50-100), 0.100-0.300 (N100-300) and 0.300-

0.630 (N300-630) µm size classes. Morning and afternoon measurements, the

former representing fresh traffic emissions from a nearby freeway (based on the

local meteorological conditions) and the latter well-mixed emissions from the

central business district, during weekdays were extracted for time series

analysis. Only the morning measurements exhibited significant trends. During

363

this time of the day, total particle number increased by 105.7% over the period

of five years and the increase was greater for larger particles, resulting in a shift

in NMD by 7.9%. There was no evidence to suggest that traffic flow in the

freeway has increased to cause the increase in particle concentration, and it was

suggested that the increase was likely due to changes in the composition of the

freeway traffic, however, this hypothesis could not be verified. More studies on

long term trends are critically needed.

7.14. SPATIAL DISTRIBUTION OF PARTICLE CONCENTRATIONS

WITHIN URBAN ENVIRONMENT

The three most common approaches to experimental studies concerned with

small-scale (urban scale) spatial variation in particle concentration, have

generally included: (i) measurements as a function of the distance from a major

road; (ii) measurements at a major road or in its immediate vicinity, as well as at

side streets; and (iii) measurements at several locations within the city. This

section discusses separately the outcomes of each of these types of studies, and

concludes with an overall comparison of ultrafine particle number concentration

in different environments. Not discussed here are vertical profiles of particle

concentrations, of particular importance in relation to urban canyon effects,

which are discussed elsewhere (Hitchins et al. 2002; Longley et al. 2003;

Vardoulakis et al. 2003; Imhof et al. 2005b).

364

7.14.1. Particle concentration as a function of the distance from the road

These studies are ideally conducted in areas where airflow between the road and

the monitoring site is undisturbed by buildings or other barriers. As an outcome,

such studies provide information about small-scale particle dispersion, where

the shape of the dispersion function is similar between the studies, and thus of

general applicability to other sites of similar topography. The studies

investigated changes to the total particle number concentration as well as to the

size distribution, and some of them also compared changes to particle

concentrations with the concentrations of gaseous pollutants emitted by

vehicles. In general, all such studies showed, as expected, a decrease in particle

concentration with distance from the road, up to about 300 m, beyond which

particle concentration levels and size distributions approach the local urban

background (Morawska et al. 1999b; Shi et al. 1999; Hitchins et al. 2000; Zhu et

al. 2002a).

At the road, particle concentrations range between 104 and 106 particles cm−3,

and show association with vehicle flow characteristics (higher the speed, the

greater the particle concentration, and the smaller the particle size), with less

variation observed in particle volume compared to particle number size

distributions (Kittelson et al. 2004). Virtanen et al.(2006) showed that the total

concentrations at roadside were dominated by nucleation mode particles, were

increasing with increasing traffic rate and the effect of traffic rate were stronger

on particles smaller than 63 nm than on the larger particles. Harrison et

al.(1999) reported that at the road significant numbers of particles are in the 3-7

nm size range, with a mode below 10nm, attributable to homogeneous

365

nucleation processes. The roadside measurements show rapid variation, with

two modes at 10 and 30 nm, and other modes present at a number of different

particle sizes (between 20 and 50 nm and 100 and 200 nm), and changing very

quickly between measurements (Harrison et al. 1999; Morawska et al. 2004).

Studies investigating total particle concentration levels showed the decrease

with the distance from the road to be exponential or according to the power law.

This is similar to the studies investigating gradients of NO2 concentrations in

the vicinity of roads (e.g. (Nitta et al. 1993; Kuhler et al. 1994; Roorda-Knape et

al. 1998)), which found them to be curvilinear. Shi et al (1999) showed that total

particle number concentration in the size range from 9.6 to 352 nm at the busy

road site was 3.6 and 3 times higher, compared to two sites at a distances of 30

and 100m from the road, respectively. A study by Hitchins et al. (2000)

conducted for total particle number concentration in the size range from 15 to

697 nm showed that for conditions where the wind was blowing directly from

the road, at a distance of approximately 100 - 150 m from the road, the

concentration decayed to about half that of the maximum occurring at 15 m

from the road (the nearest measuring point to the road), which reduced to 50 –

100 m for wind blowing parallel to the road. Zhu et al. (2002a) showed that

particle number concentration in the size range from 6 to 220 nm as a function

of distance from a road ranging from 17 to 300 m displayed an exponential

decreasing trend, similarly to the concentrations of CO and black carbon. Pirjola

et al. (2006) showed that at a distance of 65 m from the roadside, the average

concentration reduced to 39% in winter and to 35% in summer with the wind

366

perpendicular to the road whilst for wind blowing along the road in it decreased

to 19% in summer.

Gramotnev et al (2003) modelled particle dispersion using a scaling procedure

for CALINE4 developed for this purpose, and compared the results with

measured total particle concentration in the size range of 0.015 to 0.700 μm at

distances between 15 and 265 m from the road. The authors concluded that

particle concentration reduces as a power law with the distance from the road,

with the typical difference between the theory and average measured

concentration being of the order of 10%, which was likely due to processes not

included in the modelling, in particular coagulation.

As discussed in Section 7.2, it has been shown that the dynamic pollutant mix

evolves during transport from the road: nucleation leads to formation of new

particles very soon after emissions, followed by their growth by condensation,

diffusion to surfaces and coagulation. Therefore, at the road, particle

concentrations are dominated by the smallest particles, with the peak in

distribution shifting to the larger sizes at greater distances. Initially these

smallest particles are below < 10 nm, and therefore not measured if the

instrument window is above this. As they grow with time and thus distance

from the road, they become “visible” to these instruments. As this occurs over

the distance of about 90 m, there have been particle number increases reported

at such distances. In particular Zhang et al. (2004a) showed that a large number

of sub-6nm particles emitted from freeways may grow above 10 nm around 30–

90 m downwind. Afterwards, some shrink back to sub-10 nm and some

367

continue growing to above the 100nm range. Zhu et al. (2002a) measured

particle size distributions in the size range of 6–220nm at a distance from 17 to

300m downwind of freeway. At 17m downwind from the freeway, the dominant

mode was around 10nm with a modal concentration above 3.2×105/cm3 and at

20m from the freeway its concentration decreased to 2.4×105/cm3. At 30 m its

concentration decreased by about 60% with a slight shift in its location. It then

kept shifting to larger size ranges with its concentration decreasing for farther

sampling locations and to disappear at distance >150m from the freeway. The

second mode at 17m downwind from the freeway was around 20nm with a

concentration of 1.5×105/cm3 which remained more or less unchanged at 20m,

but the mode shifted to 30 nm at 30 m and continued to shift to larger sizes with

the distance from the freeway. Number concentrations of particles <50nm,

dropped significantly with increasing distances from the freeway, but for those

>100nm, number concentrations decreased only slightly. Particles in the size

range of 6 to 25nm accounted for about 70% of total UF particle number

concentration, which decayed exponentially to about 80% of the roadside value,

at 100m, levelling off after 150m. The concentrations in the size ranges 25–50

and 50–100nm, experienced a shoulder between 17 and 150 m. Very similar

results were obtained by Zhu et al. (2002b) who also showed that at 30 m

downwind from the freeway, three distinct modes were observed with geometric

mean diameters of 13, 27 and 65 nm and the smallest mode, with a peak

concentration of 1.6 x 105 cm-3, disappeared at distances greater than 90 m.

368

7.14.2. Relationship between on-road and urban background particle

concentration

A number of studies monitored concentration of particle characteristics in urban

sites located at various orientations in relation to the urban traffic, or from mobile

laboratories moving around the city. Most commonly the aims of such studies

were to compare the differences between local hot spots and urban background

locations, rather than to provide a comprehensive characterization of the

relationship between the concentrations and the distance from a particular street or

traffic flow. In addition, each of the studies was concerned with investigating

some other relationship, for example between the parameters measured in relation

to the site location. While direct comparison of the results from the studies is not

possible, as the design of each of the studies was different in terms of site

locations in relation to traffic areas, and also in terms of parameters measured,

some more general conclusions can be drawn from them. In particular the studies

showed that:

• Concentration of particle number decreases as the distance of the monitoring

site from the street increases e.g. (Buzorius et al. 1999; Kittelson et al. 2004;

Westerdahl et al. 2005) and that the difference between close to traffic and

away from the traffic concentrations are much larger for particle number

than PM10 concentrations e.g. (Harrison et al. 1999; Holmes et al. 2005).

• Particle size distribution is much more stable at background urban sites,

where it its likely to be unimodal, than close to traffic where it is multimodal

and rapidly changing e.g. (Harrison et al. 1999; Morawska et al. 2004).

369

• Near traffic, the nanometre fraction of UF dominates the total particle

number concentrations, and their contribution decreases with distance from

traffic (Ketzel et al. 2003; Kittelson et al. 2004).

• Characteristics of UF particles are more closely related to the number of

heavy-duty vehicles than to the number of light-duty vehicles (Junker et al.

2000; Holmes et al. 2005; Westerdahl et al. 2005).

All these findings support the findings of studies which investigated particle

characteristics as a function of distance from the road (see Section 7.1), which, by

their design, controlled better for all the influencing factors. The main

significance of studies comparing particle characteristics across different urban

areas is in provision of information on the magnitude of local variation in particle

concentrations levels as well as size distribution, with less significance for global

comparisons.

7.15. NUCLEATION MODE AND ITS IMPACT ON URBAN PARTICLE

CONCENTRATIONS

The mechanisms and conditions that favour formation of secondary particles by

nucleation of condensable species in the air were discussed in general in chapter

4.4. This section will focus on the frequency, temporal variation and the

magnitude of the contribution of nucleation events to particle number

concentrations. Since during some of these events particle concentrations

increase sometimes by one to two orders of magnitude, their occurrence may

have a profound impact on any future approaches to routine monitoring of

particle number concentration, interpretation of monitoring results and setting

370

particle number concentration guidelines or standards. Most commonly,

nucleation events were observed during: (i) morning rush hours, and (ii) around

midday. Therefore, mainly these two types of events and their impact on particle

characteristics are discussed here. However, Woo et al. (2001b) reported a third

type of event, with particles in the range 35–45 nm, occurring during the late

evening on several days in April and September, at an average temperature of

23°C, which showed extraordinarily high concentrations, with an increase by

factors ranging from 26 to 350. The source of these particles remains unknown,

although elevated levels of SO2, NOx and NOy were also observed, indicating

that a plume from a larger source may have passed the measurement site during

this time. The presence of elevated SO2 suggests that an industrial source may

have played a role.

Morning rush hours nucleation events, when increased emissions of

condensable species from vehicles combined with lower temperatures

(particularly during winter months), result in conditions enabling particle

nucleation when exhaust mixes with cool ambient air. Occurrence of such

events were reported by Wehner et al. (2002), Zhang et al. (2004b) and Zhang

and Wexler (2004a). Particles during these events are formed following direct

emissions from motor vehicles and thus their concentrations are correlated with

CO increases (Zhang et al. 2004b). Woo et al. (2001b) also reported elevated

concentrations of SO2, NO, NOx, as well as CO (for 12 out of 18 events).

Increased concentrations of particles in the size range of 3–10 nm, but mainly

10–100-nm, were reported during these events in urban Atlanta, Georgia

(Kittelson et al. 2004), with the average values of particle concentrations in the

371

latter size range being a factor of two to three higher than the average values

reported for three European cities (Aalto et al. 2005). Woo et al. (2001b)

reported early morning, winter concentrations elevated by a factor of two in the

10–35 nm range, with the average particle size being 51% of the diameter

before the event. This was mostly associated with temperatures exceeding 10°C,

while for temperatures below 10°C, however, the data showed a pronounced

enhancement of particles in the 4–10 nm range. The latter was explained by

nucleation that occurs as hot exhaust gases mix with the cool ambient air,

however, the study was unable to identify which source or combination of

sources was predominant. It is important to mention, that while such events

have been reported to occur predominantly in the morning, they have also been

observed in the afternoon, during rush hour, particularly in winter when the

mixing heights remain low (Woo et al. 2001b; Zhang et al. 2004b).

Midday nucleation events, when increased solar radiation (>750 W/m2 (Woo et

al. 2001a)), presence of SO2 and water vapour in the air (Pirjola et al. 1998), and

relatively low concentration of pre-existing particles (often occurring under

conditions of good atmospheric mixing: higher wind speed and high boundary

layer) lead to photo oxidation of SO2 to sulphuric acid and its subsequent binary

nucleation with water and thus formation of sulphuric acid nanoparticles. Shi et

al. (2001a) concluded that nucleation events cannot occur without the input of

solar radiation, but high solar radiation alone will not necessarily lead to a

nucleation event. Wehner and Wiedensohler (2003) found that midday

nucleation occurred on weekdays as well as on weekends, concluding that the

immediate emission of anthropogenic trace gases is not critical for particle

372

formation in urban areas, or that concentrations of these gases are always

sufficiently high to trigger new particle formation. Contrary to other studies,

Laakso et al. (2003) showed that nucleation mode particle concentration had a

minimum in summer, which was explained as related to high temperatures

which limit the nucleation (due to temperature dependence of saturation vapour

pressures). Alam et al (2003) reported occurrences of nucleation events on 8 out

of 232 days distributed throughout the year at two urban sites in Birmingham,

UK, and inferred that both the nucleation and particle growth processes

involved condensable molecules other than, or in addition to, sulphuric acid,

together with the requirement of a low total particle surface area.

Particle formation may be followed by growth of the newly formed particles,

sometimes rapid (Monkkonen et al. 2005), sometimes over longer periods of

time of up to 18 h (Zhang et al. 2004b). Ternary nucleation mechanisms

involving (H2SO4/H2O/NH3) have been proposed to account for discrepancies

found between calculated binary H2SO4/H2O nucleation rates and experimental

results (Stanier et al. 2004b). Woo et al. (2001b) observed that NOx was

typically depleted during these events (while higher before and after the event).

The same study also showed that O3 was elevated during the events. Such

events have been reported to occur predominantly in spring and summer (Zhang

et al. 2004b), but also in autumn (Laakso et al. 2003; Kittelson et al. 2004).

Particles formed through such event have been reported to be very small, mainly

in the 3-10 nm range (Harrison et al. 1999; Woo et al. 2001b; Kittelson et al.

2004; Monkkonen et al. 2005). Sardar et al.(2005) showed elevated mass

concentrations in the 32-56 nm size range during summer, which may be linked

373

to additional summer nucleation events. Interestingly, the study also showed a

mode in organic carbon (OC) between 18 and 56 nm in the summer that was not

present in the other seasons, linking it to summertime photochemical secondary

organic aerosol formation through hydroxyl radicals, which occurs by the

photolysis of ozone, and reacts with organic gases to reduce their volatility. The

frequency, intensity, meteorology, and possible chemistry of such new particle

formation has been analysed by Stanier et al. (2004b), based on sampling

conducted in a park 5 km downwind of downtown Pittsburgh, showing that the

nucleation events are associated with photochemical sulphuric acid production

and occur on approximately 30% of the study days. The same type of behaviour

has also been reported to occur in St. Louis (Shi 2003). Keeler (2004) detected

the presence of the noon nucleation peak on 5 out of 11 sampling days in

southwest Detroit during the summer. Woo et al. (2001a) observed that 19 of

the 23 pronounced noon peaks over 12 months occurred during August and

April. During such nucleation events, peaks of SO2 are strongly associated with

the number concentrations of UF (sources are vehicle emissions but also

industry). Particle number concentrations during such events can be as high as

3x104 cm-3 (Zhang et al. 2004b) and 7.3x104 cm-3 (Keeler, 2004). The latter

value was 3.5 times higher than the morning peak concentration. Similarly, Woo

et al. (2001a) showed that, during one such mid-day nucleation event, the

particle concentration increased by a factor of 50 over the morning peak.

McMurry and Woo (2002) concluded that the concentrations produced by the

photochemical nucleation event were about an order of magnitude higher than

concentration peaks that occurred during both morning and afternoon rush

hours.

374

It is expected that under high concentrations of pre-existing particles new

particle formation will not be favoured but rather condensation of condensable

species on these particles. For example Vakeva et al. (1999) conducted model

calculations, which showed that within a street canyon pre-existing particles

prevent new particle formation via the H2SO4ÐH2O-route, while above roof

height (25 m) nucleation potential is much higher. However, studies conducted

in highly polluted environments, such as Dunn et al. (2004), also reported

occurrence of nucleation events. Monkkonen et al. (2005) also reported some

around noon nucleation events in New Delhi (linked to the presence of SO2),

with the formation rate of the particles varying from 3.3 to 13.9 cm-3 s-1, despite

the fact that most commonly, the formation and growth of nucleation mode

particles were disturbed by high aerosol background concentration.

In general, as concluded by Kittelson et al.(2004), nucleation processes have a

greater effect on mean rather than median concentrations, with different trends

for particles in different size ranges (thus originating from different nucleation

events). For 10–100nm particles, both the median and mean are elevated during

the cold months, while mean values for 3–10nm particles are highly variable

and can be high both in winter and summer, with the highest mean occurring

during summer and the highest median observed in the winter. For example

Woo et al. (2001a) observed that annual average concentrations of particles in

the 3–10 nm range were elevated due to the appearance of very high

concentrations of these particles between 11 am and 2 pm on 23 days of the

year.

375

7.16. COMPARISON OF PARTICLE CONCENTRATION LEVELS

BETWEEN DIFFERENT ENVIRONMENTS

Comparison of particle concentration levels reported for different environment

was conducted for the purpose of this review by grouping the results from the

studies into eight categories according to measurement location including: road

tunnel, on-road, road-side (which indicates varying degree of distance from the

road), street canyon, urban, urban background, rural, and clean background, and

calculating mean and median values for each category. This review only

considered those papers that presented concentrations numerically; papers that

showed concentrations only graphically have not been included in this

comparison since values derived from graphs are of limited accuracy. The

majority of the studies reported their results in terms of mean values, and thus

for the purpose of this comparison only the reported mean particle number

concentrations were considered. Some studies used both CPC and SMPS

measurements for the same location; hence there were several overlaps between

the number of CPC and SMPS measuring sites. Most of the studies reported

multiple measurements at each study site, from which the average was

calculated. The overall average for each site was then calculated using the

averages for each study. Also, many studies included more than one site.

Overall, there were 3 tunnel studies (with 4 sites using the SMPS), 2 on-road

studies (with 7 sites using the CPC), 18 road-side studies (with 5 sites using the

CPC and 19 using the SMPS), 7 street canyon studies (with 1 site using the CPC

and 7 using the SMPS), 24 urban studies (with 1 site using the CPC and 24 sites

using the SMPS), 4 urban background studies (with 3 sites using the SMPS), 8

376

rural studies (with 2 sites using the CPC and 11 sites using the SMPS) and 5

clean background studies (with 9 sites using the SMPS).

Figure 7.2 presents a comparison of mean and median concentrations for

the different environments.

4.8310.76

42.0771.45

48.18

167.64

7.29

2.612.91

34.58 39.1399.09

47.00

8.83 8.10

3.20

1

10

100

1000

Tunnel (3) On Road (2) Road Side(18)

StreetCanyon(7)

Urban (24) UrbanBackground

(4)

Rural (8) CleanBackground

(5)Measurement Location

103 Pa

rtic

les/

cm3

meanmedian

Figure 7.2. Mean and median particle number concentrations for different

environments. In brackets are the numbers of sites for each environment*.

* 3 tunnel studies (Abu-Allaban et al. 2002; Jamriska et al. 2004; Imhof et al. 2005b), 2 on-road studies

(Shi et al. 2001b; Westerdahl et al. 2005), 18 road-side studies (Harrison et al. 1999; Morawska et al.

1999b; Hitchins et al. 2000; Shi et al. 2001a; Molnar et al. 2002; Thomas and Morawska 2002; Zhu et al.

2002a; Zhu et al. 2002b; Gramotnev et al. 2003; Ketzel et al. 2003; Gramotnev et al. 2004; Janhall et al.

2004; Ketzel et al. 2004; Kittelson et al. 2004; Morawska et al. 2004; Zhu et al. 2004; Gidhagen et al.

2005; Imhof et al. 2005a), 7 street canyon studies (Vakeva et al. 1999; Jamriska and Morawska 2001;

Wåhlin et al. 2001; Wehner et al. 2002; Longley et al. 2003; Gidhagen et al. 2004; Gidhagen et al. 2005),

24 urban studies (Tuch et al. 1997; Harrison et al. 1999; Hitchins et al. 2000; Junker et al. 2000; Pakkanen

et al. 2001; Ruuskanen et al. 2001; Woo et al. 2001a; McMurry and Woo 2002; Morawska et al. 2002;

Ketzel et al. 2003; Laakso et al. 2003; Wehner and Wiedensohler 2003; Hussein et al. 2004; Jamriska et al.

2004; Jeong et al. 2004; Ketzel et al. 2004; Morawska et al. 2004; Stanier et al. 2004a; Young and Keeler

2004; Gidhagen et al. 2005; Holmes et al. 2005; Hussein et al. 2005a; Janhall et al. 2006; Mejia et al.

2007a), 8 rural studies 4 urban background studies ( Hussein et al. 2004; Ketzel et al. 2004; Virtanen et al.

377

2006; Hameri et al, 1996), 5 clean background studies (Pitz et al. 2001; Laakso et al. 2003; Tunved et al.

2003; Morawska et al. 2004; Gidhagen et al. 2005;)

It can be seen from Figure 7.2 that both tunnel and roadside categories have

high standard deviations. In relation to the tunnel category, it could be due to the

small number of studies conducted. In relation to the roadside category, one

reason could be that the studies report concentrations for different distances

from the road kerb and many studies had different reference distances, and/or

more than one distance, which were difficult to normalize.

The mean concentration measured at both urban and urban background sites

were statistically different to that measured at rural and clean background. As

such, they were considered as a combined category, the mean of which was

compared with the mean of the other site categories (on-road, road side etc),

using a Students t-test, and in each case, the differences were found to be

statistically significant. The mean concentration measured at the tunnel sites

(1.67×105/cm3) was statistically higher than the means of each of the other

categories.

7.17. EXPOSURE TO ULTRAFINE PARTICLES

There have been very few studies investigating human exposure to UF particles.

In general exposure means concentrations experienced over a periods of time

spent in different microenvironments. A study by Kaur et al. (2006)

investigated exposures of volunteers walking or travelling by bus, car or taxi,

along two busy roads with approximate average daily traffic of 83,000 and

18,000 vehicles per day, respectively. The volunteers carried P-TRAK®

378

Ultrafine Particle Counters (TSI Model 8525) and the study showed that

different modes of transport resulted in different exposures, with average

personal UF particle count exposure (104 particles/cm3) of 4.61 (walking), 8.40

(cycling), 9.50 (bus), 3.68 (car) and 10.81 (taxi). (Note: the values presented in

the paper were rounded for the purpose of this review). Considerable variability

was seen in UF particle exposure, of up to an order of magnitude above

background, within a few seconds and over a few metres as people moved

through the polluted microenvironments, which implies that the influence of

time-activity and movement can be easily missed by using averaged results, thus

ultimately leading to underestimation of the exposures. Similar conclusions

were derived by Gouriou et al. (2004) who showed that particle concentration

encountered by car passengers may present high peaks, up to 106 particles cm−3.

Hourly average exposure is strongly influenced by the frequency with which an

individual encounters such exposure events, as well as their severity, and thus

mean and median concentrations over a time-averaged period may not reflect all

aspects of population exposure patterns.

7.18. RELATIONSHIP BETWEEN DIFFERENT PARTICLE

METRICS AND WITH GASEOUS POLLUTANTS

Many studies in addition to particle number, measured concentrations of particle

mass (or mass surrogate) and of gaseous pollutants. The relationship between

the measured pollutants was analysed with an aim to gain a better insight into

pollution sources or pollution dynamics. In some cases existence of quantifiable

relationship between some of the pollutants would provide justification for

379

using some of the pollutants as a surrogate of others and thus lowering the

overall costs of monitoring.

Relationship between particle number, surface area, volume or mass

concentrations. Correlations have been considered between total or size

classified particle number concentrations and other particle metrics. Harrison

and Jones (2005) measured particle number concentrations at eight different

urban sites and showed that hourly and daily averaged number concentrations

were only weakly correlated to PM10. Harrison et al. (1999) showed that

correlation between particle number concentrations and PM10 was higher at

traffic, than at a background sites. Zhang et al. (2004b) reported that there was

no correlation between UF particle number and PM2.5 mass concentrations, and

modest and good correlation between N50-100 and N100-470 and PM2.5,

respectively. Laakso et al. (2003) showed that there was no correlation between

particle mass and nucleation mode, total concentration or Aitken mode. Woo et

al. (2001b) did not find correlation between the number and surface area or

volume concentrations. Ruuskanen et al. (2001) showed that the total and UF

number concentrations were poorly correlated with PM2.5 levels while the

number concentration of accumulation particles showed better correlation. Lin

et al. (2007) showed that the ratios of PM0.056/PM10 and PM0.1/PM10 mass

concentrations at the roadside were 0.13 and 0.17, respectively, and were

significantly lower at the rural site (0.05 and 0.06, respectively), suggesting that

the roadside is exposed to more nano and UF particles than the rural area. Thus

in general, while there is some level of correlation between some particle

metrics reported by some studies, other studies did not find any correlation. This

380

can be explained by different sources of bigger and smaller particles in different

environments, and therefore, without local measurements, the degree, if any, of

local correlations cannot be predicted from studies conducted elsewhere.

Relationship between particle number, gaseous pollutant and black carbon

concentrations. Most of the studies conducted in the proximity to traffic showed

existence of correlations between these pollutants. Zhu et al. (2002a) and Zhu et

al. (2002b) reported that concentration of CO, BC and particle number tracked

each other with the increasing distance from the freeway. Ketzel et al. (2003)

established that NOx and total particle number were well correlated at the urban

and near-city level indicating a common traffic source. Sardar et al. (2005)

showed that during fall and winter (when vehicular emissions become the

dominant pollution source) particles greater than 56 nm are correlated well with

CO and NOx, while very low correlations are observed between O3 and particle

number of any size range. Westerdahl et al. (2005), who measured pollutant

concentrations from an all-electric mobile platform, found that particles <1µm

were highly correlated with BC and NO, while moderately correlated with CO2

and poorly with CO. The latter was explained by the fact that CO emissions are

primarily from gasoline-powered vehicles, and relatively unrelated to those

pollutants dominated by diesel vehicles (including particle number). Similar

conclusion was derived by Bukowiecki et al. (2002).

Thus, while in general there is a reasonably good correlation between UF

particles and traffic emitted gaseous pollutants as well as BC, the existence and

the degree of correlation varies. As concluded by Paatero et al. (2005), who

381

estimated levels of particle number concentrations by retrospective modelling

using measured air pollution and weather variables, models must be city-

specific: associations of particle number concentrations with other pollutants

differ between different cities.

7.19. CONCLUSIONS AND IMPLICATIONS FOR THE EXPOSURE

AND EPIDEMIOLOGICAL STUDIES

This review compiled and synthesized the existing knowledge on UF particles

in the air with a specific focus on those originating due to vehicles emissions.

As it has been shown in this review, vehicles are a significant sources of UF

particles, and it is the vehicle emissions that are commonly the most significant

source of air pollution in general in populated urban areas. It is therefore of

particular significance to understand the magnitude and characteristics of the

vehicle-affected UF particles in urban air, as it is this type of environment which

is the most likely to be considered as a target for future air quality regulations in

relation to particle number. Industrial and power plant emissions (covered in

Part I) have a significant impact on the environment and climate, but as they

often (but not always) occur outside the most populated urban settings, their

direct impact on human exposure is lower than the impact of vehicle emissions.

UF particles are most commonly measured in terms of their number

concentrations, and unlike particle mass concentration (PM2.5, PM10), there is no

standard methods for conducting size classified particle number measurement.

The review showed that the term “UF particles” is often used imprecisely,

meaning various ranges of particle number concentration in a subset of the

382

submicrometer range. In addition, the number concentrations reported depend

on the instrument used and its setting. It has been shown by this review that the

mean and the median CPC's measurements are 32% and 56%, respectively,

higher than DMPS/SMPS's ones. While the differences for specific

environments could vary (larger differences expected for the environments

where nucleation mode is present and smaller where aged aerosol dominates), it

nevertheless shows what overall magnitude of differences can be expected when

comparing results using these different measuring techniques. It is important to

keep these differences in mind when attempting to establish quantitative

understanding of variation in particle concentrations reported by different

studies. This also points out the need to develop and utilize standardised

measurement procedures, enabling meaningful comparison between the results

from different studies, which is of particular significance for human exposure

and epidemiological studies.

Despite these differences in reporting measured concentration levels, this review

showed that it is possible to quantify the differences between background

concentrations of UF particles in clean environments, with the levels in the

environments affect by vehicle emissions. It has been shown that the clean

background levels are on average of the order of 2.67 ± 1.79 x 103 cm-3 , while

levels at urban sites are 4 times higher and levels at street canyons, roadside,

road and tunnel sites [R3] are 27, 18, 16 and 64 times higher, respectively. Thus

the range of concentrations between clean and vehicle effected environments

spans over two orders of magnitude. This is very different from particle mass;

for example a review by Morawska (2003) showed the decrease in mass

383

concentration between a busy road and urban background ranges only from 0 to

about 25–30%. This large variation in particle number concentration across

different environments has profound significance in relation to human exposure

assessment and epidemiological studies. This means that unless exposure

assessment is conducted where the exposures occur and at time scales that

elucidates the temporal nature of the exposure, it is unlikely that

epidemiological studies would provide answers based only on monitoring in

central locations. In other words, central monitoring alone underestimates

exposures and may lead to inappropriate management of public health risks.

Lack of answers from epidemiological studies in relation to UF particles

(exposure-response relationships) means, that it is not possible to develop health

guidelines, a basis for national regulations. In relation to airborne particle mass

it has been shown that within the current range of concentrations studied in

epidemiological studies there are no threshold levels and that there is a linear

exposure-health response relationship. Based on this, in the recent review the

World Health Organization Air Quality Guidelines, a new set of guidelines for

particulate matter was introduced, with annual mean values for PM2.5 and PM10

of 10 and 20 µg m-3, respectively. This was based on an American Cancer

Society

study (Pope et al. 2002) and represents the lowest end of the range over which

significant effects on survival have been observed (WHO 2005). It is important

to note that these values are not much higher than the concentration levels

encountered most commonly in natural environments (while it should be

acknowledged that, in some locations and under some circumstances,

384

concentrations in natural environments may be well below or above those cited).

If future epidemiological studies report response at lower concentration levels

PM2.5 and PM10, it is likely that the guideline values will be lowered even

further. While lack of exposure response relationship makes it impossible to

propose health guideline for UF particles, it is important to point out that as

discussed above, the current levels in environments affected by vehicle

emissions are up to an order of magnitude higher than in the natural

environments. Thus, if there is also no threshold level in response to exposure to

UF particles (or if it is very low), future control and management strategies

should target a decrease of these particles in urban environments by more than

one order of magnitude. At present there is a long way to go to achieve this.

When considering future management strategies for UF particles, as discussed

in this review, there are a few challenges, which include:

1. Currently there are large uncertainties in relation to vehicle emission factors

for different particle size ranges and for particle numbers, there are no

emission inventories of UF particles from motor vehicles, and there is only

very limited data on long term trends in UF particle concentrations in urban

environments. All these aspects should attract significant research efforts, as

this knowledge is critical for management and control of UF particles.

2. Estimations of pollution concentration in the air are commonly derived

based on source emission inventories, which in turn are derived utilizing the

source emission factors. However, due to the process of secondary particle

385

formation, estimation of UF particle concentration cannot be derived solely

based on vehicle emission factors (which are more likely to reflect

emissions of primary particles), but have to include also predictions for

secondary particle formation in exhaust plumes and particle formation by

nucleation processes in the wider atmosphere.

3. Secondary particle formation results in rapid increase of particle number

concentrations by one to two orders of magnitude to the concentration levels

of the order of 105 particles cm-3. The majority of the new particles are

formed by ion-induced or binary nucleation of sulphuric acid and water or

by ternary nucleation involving a third molecule followed by condensation

of semi-volatile organics, with photochemistry playing an important role in

some of these processes. The mechanisms of new particle formation

strongly depend on local meteorological factors, and therefore models of the

dynamics of particle formation in urban environments have to include all

factors involved and thus must be local area specific.

4. These significant peaks in particle number concentration due to secondary

particle formation are a challenge, if there were future regulations

considered based on particle number. Issues to resolve include: (i) whether

the regulations should be set around the base line concentrations without the

peak concentrations, or whether they should include the peaks; and (ii) how

to define the peaks. Developing a much better picture of particle formation

dynamics in different environments, including those which are influenced by

traffic, would greatly assist such regulation formulation.

386

5. It is not only the particle number concentration, but also particle

composition which should be considered when characterising UF particles.

The review showed that there have been only a relatively small number of

studies focused on UF particle chemistry. There are large differences in

particle chemical composition including particle solubility, volatility,

elemental composition, etc reported by the studies. The differences depend

on many factors, including vehicle technology, fuel used and after treatment

devices, but also on the post formation processes occurring atmospheric

transport. Since particle composition may be a factor determining particle

toxicity there is a need for developing a much better knowledge on UF

particle chemistry in different environments.

In summary, the magnitude of the impact of UF particles on human health and

the environment has still not been fully quantified (while the picture starts

emerging) nor is it fully understood, and the first step in this direction is to

develop an in depth understanding of particle concentrations, characteristics,

time trends and spatial distribution in clean and anthropogenic modified

environments. This knowledge would, in turn, lead to an understanding of the

potential impacts of the particles on the environment and would provide

scientific foundation for future studies in the area of human epidemiology. It

would also be used as a basis for setting any future emission and air quality

standards based on particle number.

387

7.20. REFERENCES

Aalto, P., Hameri, K., Becker, E., Weber, R., Salm, J., Makela, J., Hoell, C.,

O'Dowd, C., Karlsson, H., Hansson, H., Vakeva, M., Koponen, I., Buzorious,

G., Kulmala, M., 2001. Physical Characterization of Aerosol Particles During

Nucleation Events, Tellus B53, 344-358.

Aalto, P., Hameri, K., Paatero, P., Kulmala, M., Bellander, T., Berglind, N.,

Bouso, L., Castano-Vinyals, G., Cattani, G., Cyrys, J., Von Klot, S., Lanki, T.,

Marconi, A., Nyberg, F., Pekkanen, J., Peters, A., Sjovall, B., Sunyer, J.,

Zetzsche, K., Forastiere, F., 2005. Aerosol particle number concentration

measurements in five European citied using TSI-3022 condensation particle

counter over a three year period during HEAPSS (Health Effects of Air

Pollution on susceptible Subpopulations), Journal of the Air and Waste


Abu-Allaban, M., Coulomb, W., Gertler, A., Gillies, J., Pierson, W., Rogers,

C., Sagebiel, J., Tarnay, L., 2002. Exhaust Particle Size Distribution

Measurements at the Tuscarora Mountain Tunnel, Aerosol Science and

Technology 36, 771-789.

Airborne Particles Expert Group (1999). Source apportionment of airborne

particulate matter in the United Kingdom. Report for the Department of the

Environment, Transport and the Regions, the Welsh Office, the Scottish

Office and the Department of the Environment (Northern Ireland).

388

Andersson, J., Wedekind, B., Hall, D., Stradling, R., Wilson, G., 2001.

DETR/SMMT/CONCAWE Particulate Research Programme: Light Duty

Results. SAE Technical Paper Series, Society of Automotive Engineers, No.

2001-01-3577.

Arnold, F., Kiendler, A., Wiedemer, V., Aberle, S., Stilp, T., Busen, R., 2000.

Chemiionconcentration measurements in jet engine exhaust at the ground:

implications for ion chemistry and aerosol formation in the wake of a jet

aircraft, Geophysical Research Letters 27, 1723-1726.

Ayers, G., Gras, J., 1991. Seasonal Relationship between cloud condensation

nuclei and aerosol methanesulphonate in marine air, Nature 353, 834-835.

Bagley, S.T., Baumgard, K.J., Gratz, L.D., Johnson, J.J., Leddy, D.G., 1996.

Characterization of Fuel and After-Treatment Device Effects on Diesel

Emissions, Health Effects Institute: Research Report No. 76.

Baron, P.A., Willeke, K., Eds. 2001. Aerosol Measurement: Principles,

Techniques and Applications. New York, van Nostrand Reinhold.

Birmili, W., Wiedensohler, A., 2000. Evolution of newly formed aerosol

particles in the continental boundary layer: a case study including OH and

H2SO4 measurements, Geophysical Research Letters 27, 2205-2208.

389

Bukowiecki, N., Dommen, J., Prévôt, A. S. H., Richter, R., Weingartner, E.,

Baltensperger, U., 2002. A mobile pollutant measurement laboratory-

measuring gas phase and aerosol ambient concentrations with high spatial and

temporal resolution , Atmospheric Environment 36, 5569-5579.

Burtscher, H., 2001. Literature study on tailpipe particulate emission

measurement for diesel engines. Particulate Measurement Program

BUWAL/GRPE.

Buzorius, G., Hameri, K., Pekkanen, J. and Kulmala, M., 1999. Spatial

variation of aerosol number concentration in Helsinki city, Atmospheric


Cabada, J.C., Takahama, S., Khlystov, A.Y., Wittig, B., Pandis, S., Rees, S.,

Davidson, C.I., Robinson, A.L., 2004. Mass size distributions and size

resolved chemical composition of fine particulate matter at the Pittsburgh

Supersite, Atmospheric Environment 38(10), 3127-3141.

Casati, R., Scheer, V., Vogt, R., Benter, T., 2007. Measurement of nucleation

and soot mode particle emission from a diesel passenger car in real world and

laboratory in situ dilution, Atmospheric Environment 41, 2125-2135.

Charron, A., Harrison, M., 2003. Primary Particle Formation from Vehicle

Emission During Exhaust Dilution in the Roadside Atmosphere, Atmospheric


390

Cheng, M., Tanner, R., 2002. Characterization of Ultrafine and Fine Particles

at a Site Near the Great Smoky Mountains National Park, Atmospheric


Clarke, A., Davis, D., Kapustin, V. N., Eisele, F. L., Chen, G., Paluch, I.,

Lenschow, D., Bandy, A.R., Thornton, D., Moore, K., Mauldin, L., Tanner, D.

J., Litchy, M., Carroll, M.A., Collins, J., Albercook, G., 1998. Particle

nucleation in the tropical boundary layer and its coupling to marine sulphur

sources, Science 282, 89-92.

Dahl, A., Gharibi, A., Swietlicki, E., Gudmundsson, A., Bohgard, M.,

Ljungman, A., Blomqvist, G., Gustafsson, M., 2006. Traffic Generated

Emissions of Ultrafine Particles from Pavement-Tire Interface, Atmospheric


Dunn, M. J., Jimnez, J. L., Baumgardner, D., Castro, T., Mc-Murry, P. H.,

Smith, J. N., 2004. Measurements of Mexico City nanoparticles size

distributions: Observations of new particle formation and growth,

Geophysical Research Letters 31 31, L10102.

Easter, R.C., Peters, L. K., 1994. Binary homogeneous nucleation:

temperature and relative humidity fluctuations, nonlinearity and aspects of

new particle production in the atmosphere, Journal of Applied Meteorology

33, 775-784.

391

Ebelt, S., Brauer, M., Cyrys, J., Thomas, T., Kreyling, W. G. , Wichmann,

H.E., 2001. Air quality in postunification erfurt, East Germany: Associating

changes in pollutant concentrations with changes in emissions, Environmental

Health Perspectives 109(4), 325-333.

Fine, P., Shen, S., Sioutas, C., 2004. Inferring the sources of fine and ultrafine

particulate matter at downwind receptor sites in the Los Angeles basin using

multiple continuous measurements, Aerosol Science and Technology 38(S1),

182-195.

Gamas, E.D., Diaz, L., Rodriguez, R., Lopez-Salinas, E., Schifter, I., 1999.

Exhaust emissions from gasoline and LPG-powered vehicles operating at the

altitude of Mexico City, Journal of the Air & Waste Management Assoc., 49,

1179-1189.

Garcia-Nieto, P.J., Garcia, B.A., Diaz, J.M.F., Brana, M.A R., 1994.

Parametric study of selective removal of atmospheric aerosol by below-cloud

scavenging, Atmospheric Environment 28, 2335-2342.

Geller, M.D., Kim, S., Misra, C., Sioutas, C., Olson, B. A., Marple, V. A.,

2002. A methodology for measuring size-dependent chemical composition of

ultrafine particles, Aerosol Science and Technology 36, 748-762.

392

Gidhagen, L., Johansson, C., Langner, J., Foltescu, V., 2005. Urban Scale

Modelling of Particle Number Concentration in Stockholm, Atmospheric


Gidhagen, L., Johansson, C., Langner, J., Olivares, G., 2004. Simulation of

NOx and Ultrafine Particles in a Street Canyon in Stockholm, Sweden,


Gidhagen, L., Johansson, C., Strom, J., Kristensson, A., Swietlicki, E., Pirjola,

L., Hansson, H., 2003. Model simulation of ultrafine particles inside a road

tunnel, Atmospheric Environment 37, 2023-2036.

Gieshaskiel, B., Ntziachristos, L., Samaras, Z., Scheer, V., Casati, R., Vogt,

R., 2005. Formation Potential of Vehicle Exhaust Nucleation Mode Particles

On-Road and in the Laboratory, Atmospheric Environment 39, 2191-2198.

Gouriou, F., Morin, J., Weill, M., 2004. On Road Measurements of Particle

Number Concentrations and Size Distributions in Urban and Tunnel

Environments, Atmospheric Environment 38, 2831-2840.


Determination of average emission factors for vehicles on a busy road,


393

Gramotnev, G., Ristovski, Z., 2004. Experimental Investigation of Ultrafine

Particle Size Distribution Near a Busy Road, Atmospheric Environment 38,

1767-1776.

Gramotnev, G., Ristovski, Z. D., Brown, R. J., Madl, P., 2004. New methods

of determination of average particle emission factors for two groups of

vehicles on a busy road, Atmospheric Environment 38(16), 2607-2610.

Graskow, B., Kittelson, D. B., Abdul-Khalek, I., Ahmadi, M., Moris, J., 1998.

Characterisation of Exhaust Particulate Emissions from a Spark Ignition

Engine, SAE Paper 980528, 155-165.

Grose, M., Sakurai, H., Savstrom, J., Stolzenburg, M. R., Watts, W.F.,

Morgan, C. G., Murray, I.P., Twigg, M.V., Kittelson, D.B., McMurry, P.H.,

2006. Chemical and physical properties of ultrafine diesel exhaust particles

sampled downstream of a catalytic trap, Environmental Science &


Hameri, K., Kulmala, M., Aalto, P., Leszczynski, K., Visuri, R., Hamekoski,

K., 1996. The Investigations of Aerosol Particle Formation in Urban

Background of Helsinki, Atmospheric Research 41, 281-298.

Hara, K., Nakae, S., Miura, K., 1997. Properties of ion nucleation in the

atmosphere, Atmospheric Electricity 17, 53-58.

394

Hara, K., Nakae, S., Miura, K., 1998. Properties of ion-induced nucleation

obtained from mobility measurements, Journal of Aerosol Science 29, S139-

140.

Harayama, N., 1992. Effects of sulfate adsorption on performance of diesel

oxidation catalysts. SAE Technical Paper Series, Society of Automotive

Engineers, No. 920852.

Harris, S.J., Maricq, M.M., 2001. Signature size distributions for diesel and

gasoline engine exhaust particulate matter, Journal of Aerosol Science 32,

749-764.

Harrison, R., Jones, M., Collins, G., 1999. Measurements of the Physical

Properties of Particles in the Urban Atmosphere, Atmospheric Environment

33, 309-321.

Harrison, R.M., Yin, J., Mark, D., Stedman, J., Appleby, R.S., Booker, J.,

Moorcroft, S., 2001. Studies of the coarse particle (2.5-10 μm) component in

UK urban atmospheres, Atmospheric Environment 35, 3667-3679.

Hitchins, J., Morawska, L., Gilbert, D., Jamriska, M., 2002. Dispersion of

particles from vehicle emissions around high- and low-rise buildings, Indoor

Air 12(1), 64-71.

395

Hitchins, J., Morawska, L., Wolff, R., Gilbert, D., 2000. Concentrations of

submicrometre particles from vehicle emissions near a major road,


Holmen, B.A., Ayala, A., 2002. Ultrafine PM emissions from natural gas,

oxidation catalyst diesel and particle trap diesel heavy-duty transit buses,


Holmes, N., 2007. A Review of Particle Formation Events and Growth in the

Atmosphere in the Various Environments and Discussion of Mechanistic

Implications, Atmospheric Environment 41, 2183-2201.

Holmes, N.S., Morawska, L., Mengersen, K., Jayaratne, R., 2005. Spatial

distribution of submicrometre particles and CO in an urban microscale

environment, Atmospheric Environment 39(22), 3977-3988.

Hussein, T., Hameri, K., Aalto, P., Paatero, P., Kulmala, M., (2005a). Modal

structure and spatial-temporal variations of urban and suburban aerosols in

Helsinki-Finland, Atmospheric Environment 39, 1655-1668.

Hussein, T., Hameri, K., Heikkinen, M., Kulmala, M., 2005b. Indoor and

outdoor particle size characterisation at a family house in Espoo, Finland,


396

Hussein, T., Puustinen, A., Aalto, P., Makela, J., Hameri, K., Kulmala, M.,

2004. Urban Aerosol Number Size Distributions, Atmospheric Chemistry and


Imhof, D., Weingartner, E., Ordonez, C., Gerhig, R., Hill, M., Buchmann, B.,

Baltersperger, U., 2005a. Real World Emission Factors of Fine and Ultrafine

Aerosol Particles for Different Traffic Situations in Switzerland,

Environmental Science and Technology 39, 8341-8350.

Imhof, D., Weingartner, E., Prevot, A., Ordonez, C., Kurtenbach, R., Wiesen,

P., Rodler, J., Sturm, P., McCrae, I., Sjodin, A., Baltersperger, U., 2005b.

Aerosol and NOx Emission Factors and Submicron Particle Number Size

Distributions in Two Road Tunnels with Different Traffic Regimes,

Atmospheric Chemistry and Physics Discussions 5, 5127-5166.

Jacobson, M., Kittelson, D., Watts, W., 2005. Enhanced coagulation due to

evaporation and its effect on nanoparticle evolution, Environmental Science &


Jamriska, M., Morawska, L., 2001. A model for determination of motor

vehicle emission factors from on-road measurements with a focus on

submicrometer particles, The Science of The Total Environment 264(3), 241-

255.

397

Jamriska, M., Morawska, L., Mengersen, K., 2007. The Effect of Temperature

and Relative Humidity on size Generated Traffic Exhaust Particle Emissions,

Atmospheric Environment, Submitted.

Jamriska, M., Morawska, L., Thomas, S., He, C., 2004. Diesel bus emissions

measured in a tunnel study, Environmental Science & Technology 38, 6701-

6709.

Janhall, S., Jonsson, A., Molnar, P., Svensson, E., Hallquist, M., 2004. Size

Resolved Traffic Emissions Factors of Submicrometer Particles, Atmospheric


Janhall, S., Olofson, F., Andersson, P., Pettersson, J., Hallquist, M., 2006.

Evolution of the Urban Aerosol During Winter Temperature Inversion

Episodes, Atmospheric Environment 40: 5355-5366.

Jeong, C., Hopke, P., Chalupa, D., Utell, M., 2004. Characteristics of

Nucleation and Growth Events of Ultrafine Particles Measured in Rochester

NY, Environmental Science and Technology 38, 1933-1940.

Jones, A., Harrison, R., 2006. Estimation of the emission factors of particle

number and mass fractions from traffic at a site where mean vehicle speeds

vary over short distances, Atmospheric Environment 40, 7125-7137.

398

Jung, H., Kittleson, D., 2005. Measurement of electrical charge on diesel

particles, Aerosol Science and Technology 39, 1129-1135.

Junker, M., Kasper, M., Roosli, M., Camenzind, M., Kunzli, N., Monn, C.,

Theis, G., Braun, C., 2000. Airborne particle number profiles, particle mass

distribution and particle bound PAH concentrations within the city

environment of Basle: An assessment of the BRISKA project, Atmospheric

Environment 43(19), 3171-3181.

Kasper, M., 2005. Sampling and measurement of nanoparticle emissions for

type approval and field control. SAE Technical Paper Series, Society of

Automobile Engineers, No. 2005-26-013.

Kaur, S., Clark, R., Walsh, P., Arnold, S., Colvile, R., Nieuwenhuijsen, M.,

2006. Exposure visualisation of ultrafine particle counts in a transport

microenvironment, Atmospheric Environment 40, 386-398.

Kawai, T., Goto, Y., Odaka, M., 2004. Influence of dilution process on engine

exhaust nanoparticles. SAE Technical Paper Series, Society of Automobile

Engineers, No. 2004-01-0963.

Keeler, G., 2004. Characterization of ultrafine particle number concentration

and size distribution during a summer campaign in southwest Detroit, Journal

of the Air and Waste Management Association 54, 1079-1090.

399

Keogh, D., Kelly, J., Mengersen, K., Morawska, L., Jayaratne, E.R., 2007.

Emission factors for estimating motor vehicle particle emissions in urban

areas, Environmental Science & Technology, Submitted. This paper is now

published in ESPR - see:- Keogh, D.U., Kelly, J., Mengersen, K., Jayaratne,

R., Ferreira, L., Morawska, L., 2009. Derivation of motor vehicle tailpipe

particle emission factors suitable for modelling urban fleet emissions and air

quality assessments. Environmental Science and Pollution Research –

International. Published online, doi 0.1007/s11356-009-0210-9.

Kerminen, V.-M., L.,P., Kulmala, M., 2001. How significantly does

coagulational scavenging limit atmospheric particle production?, Journal of

Geophysical Research 106, 24119-24125.

Ketzel, M., Wahlin, P., Berkowicz, R., Palmgren, F., 2003. Particle and trace

gas emission factors under urban driving conditions in Copenhagen based on

street and roof-level observations, Atmospheric Environment 37(20), 2735-

2749.

Ketzel, M., Whalin, P., Kristensson, A., Swietlicki, E., Berkowicz, R.,

Nielsen, O., Palmgren, F., 2004. Particle Size Distribution and Particle Mass

Measurement at Urban, Near City and Rural Level in the Copenhagen Area

and Southern Sweden, Atmospheric Chemistry and Physics 4, 281-292.

400

Khalek, I.A., Kittleson, D., Brear, F., 1998. Diesel exhaust particle size:

Measurement issues and trends. SAE Technical Paper Series, Society of

Automobile Engineers, No. 980525, 133-145.

Khalek, I.A., Kittleson, D., Brear, F., 1999. The influence of dilution

conditions on diesel exhaust particle size distribution measurements. SAE

Technical Paper Series, Society of Automobile Engineers, No. 1999-01-1142.

Khalek, I. A., Kittleson, D., Brear, F., 2000. Nanoparticle growth during

dilution and cooling of diesel exhaust: Experimental investigation and

theoretical assessment. SAE Technical Paper Series 2000, Society of

Automobile Engineers, No. 2000-01-0515.

Kim, S., Shen, S., Sioutas, C., Zhu, Y., Hinds, W. C., 2002. Size Distribution

and Diurnal and Seasonal Trends of Ultrafine Particles in Source and Receptor

Sites of the Los Angeles Basin, Air & Waste Manage Association 52, 297-

307.

Kirchstetter, T.W., Harley, R.A., Kreisberg, N.M., Stolzenberg, M R. and

Hering, S.V., 1999. On-road measurement of fine particle and nitrogen oxide

emissions from light- and heavy-duty motor vehicles, Atmospheric


Kittelson, B.D., 1998. Engines and Nanoparticles: a Review, Journal of

Aerosol Science 29(5), 575-588.

401

Kittelson, D., Watts, W., Johnson, J. , 2006a. On-road and laboratory

evaluation of combustion aerosols - Part1: Summary of diesel engine results,

Journal of Aerosol Science 37, 913-930.

Kittelson, D., Watts, W., Johnson, J., 2006b. On-road and laboratory

evaluation of combustion aerosols - Part 2: Summary of spark ignition engine

results, Journal of Aerosol Science 37, 931-949.

Kittelson, D.B., Watts, W.F. , Johnson, J., 2002. Diesel Aerosol Sampling

Methodology - CRC E-43, Final Report, University of Minnesota, Report for

the Coordinating Research Council.


Minnesota highways, Atmospheric Environment 38, 9-19.

Kittleson, D., Pui, D.Y H., Moon, K.C., 1986. Electrostatic collection of

diesel particles. SAE Technical Paper Series, Society of Automotive

engineers, No. 860009.

Korhonen, P., Kulmala, M., Laaksonen, A., Viisanen, Y., McGraw, R.,

Seinfeld, J.H., 1999. Ternary nucleation of H2SO4, NH3, and H2O in the

atmosphere, Journal of Geophysical Research 104, 26349-26353.

Kuhler, M., Kraft, J., Bess, H.U.H., Schurmann, D., 1994. Comparison

between measured and calculated concentrations of nitrogen oxides and ozone

in the vicinity of a motorway, Science of the Total Environment 147, 387-394.

402

Kuhn, T., Krudysz, M., Zhu, Y., Fine, P., Hinds, W., Froines, J., Sioutas, C.,

2005. Volatility of indoor and outdoor ultrafine particulate matter near a

freeway, Journal of Aerosol Science 36, 291-302.

Kulmala, M., Pirjola, L., Makela, J., 2000. Stable sulphate clusters as a source

of new atmospheric particles, Nature 404, 66-69.

Kulmala, M., Vehkamaki, H., Petaja, T., Dal Maso, M., Lauri, A., Kerminen,

V., Birmilli, W., McMurry, P., 2004. Formation and Growth Rates of

Ultrafine Atmospheric Particles: A Review of Observations, Journal of

Aerosol Science 35, 143-176.

Kulmama, M., 2003. How Particles Nucleate and Grow, Science 302, 1000-

1001.

Kwon, S., Lee, K. W., Saito, K., Shinozaki, O., Seto, T., 2003. Size-dependent

volatility of diesel nanoparticles: Chassis dynamometer experiments,


Laakso, L., Hussein, T., Aarino, P., Komppula, M., Hiltunen, V., Viisanen,

Y., Kulmala, M., 2003. Diurnal and Annual Characteristics of Particle Mass

and Number Concentrations in Urban, Rural and Arctic Envrionments in

Finland, Atmospheric Environment 37, 2629-2641.

403

Lin, C., Chen, S., Huang, K., Lee, W., Lin, W., Liao, C., Chaung, H., Chiu,

C., 2007. Water Soluble Ions in Nano/Ultrafine/Fine/Coarse Particles

Collected Near a Busy Road and at a Rural Site, Environmental Pollution 145,

562-570.

Longley, I., Gallagher, M., Dorsey, J., Flynn, M., Allan, J., Alfarra, M., Inglis,

D., 2003. A Case Study of Aerosol (4.6 nm < Dp < 10 um) Number and Mass

Size Distribution Measurements in a Busy Street Canyon in Manchester, UK,


Lyyranen, J., Jokiniemi, J., Kauppinen, E. I., Backman, U., Vesala, H., 2004.

Comparison of different dilution methods for measuring diesel particle

emissions, Aerosol Science and Technology 38, 12-23.

Maricq, M.M., 2006. On the electrical charge of motor vehicle exhaust

particles, Journal of Aerosol Science 37, 858-874.

Maricq, M.M., Chase, R. E., Xu, N., Laing, P.M., 2002. The effects of the

catalytic converter and fuel sulphur level on motor vehicle particulate matter

emissions, Environmental Science & Technology 36, 283-289.

Marti, J., Weber, R., 1997. New particle formation at a remote continental

site: assessing the contributions of SO2 and organic precursors, Journal of

Geophysical Research - Atmospheres 102, 6331-6339.

404

Mathis, U., Ristimaki, J., Mohr, M., Keskinen, J., Ntziachristos, L., Samaras,

Z., Mikkanen, P., 2004. Sampling conditions for the measurement of

nucleation mode particles in the exhaust of a diesel vehicle, Aerosol Science

and Technology 38, 1149-1160.

McFiggans, G., 2005. Atmospheric science: Marine aerosols and iodine

emissions, Nature 433(E13).

McMurry, P., 2000. A Review of Atmospheric Aerosol Measurements,


McMurry, P., Woo, K., 2002. Size Distributions of 3-10nm Urban Atlanta

Aerosols: Measurements and Observations, Journal of Aerosol Medicine 15,

169-178.

Mejia, J., Morawska, L., Mengersen, K., 2007a. Spatial variation in particle

number size distribution in a large metropolitan area, Atmospheric Chemistry

and Physics, Submitted.

Mejia, J., Wraith, D., Mengersen, K., Morawska, L., 2007b. Trends in size

classified particle number concentration in subtropical Brisbane, Australia,

based on a five year study, Atmospheric Environment 41, 1064 - 1079.

Meyer, N., Ristovski, Z., Jayaratne, R., 2006. Volatile Properties of CNG and

Diesel Bus Emissions Produced During Steady State and Transient Driving

Modes. 10th ETH-Conference on Combustion Generated Nanoparticles.

405

Meyer, N.K., Ristovski, Z.D., 2007. Ternary Nucleation as a Mechanism for

the Production of Diesel Nanoparticles: Experimental Analysis of the Volatile

and Hygroscopic Properties of Diesel Exhaust Using the Volatilization and

Humidification Tandem Differential Mobility Analyzer, Environmental

Science & Technology 41, 7309 - 7314.

Mirme, A., Tamm, E., Mordas, G., Vana, M., Uin, J., Mirme, S., Bernotas, T.,

Laakso, L., Hirsikko, A., Kulmala, M., 2007. A wide-range multi-channel Air

Ion Spectrometer, Boreal Environmental Research 12, 247-264.

Mohr, M., Lehmann, U., 2003. Comparison Study of Particle Measurement

Systems for Future Type Approval Application, Swiss Contribution to GRPE

Particulate Measurement Program (GRPE-PMP CH5). EMPA Report No.

202779

Molnar, P., Janhall, S., Hallquist, M., 2002. Roadside Measurements of Fine

and Ultrafine Particles at a Major Road North of Gothenburg, Atmospheric


Monkkonen, P., Koponen, I., Lehtinen, K., Hameri, K., Uma, R., Kulmala,

M., 2005. Measurements in a highly polluted Asian mega city: Observations

of aerosol number size distribution, modal parameters and nucleation events,


406

Morawska, L., 2003. Motor vehicle emissions as a source of indoor particles.

Indoor Environment, Airborne Particles and Settled Dust. L. Morawska and T.

Salthammer. Weinheim, Germany, WILEY-VCH.

Morawska, L., Bofinger, N., Kocis, L., Nwankowala, A., 1998a. Submicron

and supermicron particulates from diesel vehicle emissions, Environmental

Science & Technology 32, 2033-2042.

Morawska, L., Hofmann, W., Thomas, S., Ristovski, Z.D., Jamriska, M.,

Rettenmoser, T., Kagerer, S., 2004. Exploratory cross sectional investigations

on ambient submicrometer particles in the alpine region of Salzburg, Austria,


Morawska, L., Jamriska, M., Thomas, S., Ferreira, L., Mengersen, K., Wraith,

D., McGregor, F., 2005. Quantification of particle number emission factors

for motor vehicles from on-road measurements, Environmental Science &


Morawska, L., Jayaratne, E. R., Mengersen, K., 2002. Differences in airborne

particle and gaseous concentrations in urban air between weekdays and

weekends, Atmospheric Environment 36, 4375-4383.

Morawska, L., Johnson, G., Ristovski, Z.D., Agranovski, V., 1999a. Relation

between particle mass and number for submicrometer airborne particles,


407

Morawska, L., Keogh, D., Thomas, S., Mengersen, K., Wilson, W., 2007a.

Modality in ambient particle size distributions and its potential as a basis for

developing air quality regulation, Atmospheric Environment, 42 (7), 1617-

1628.

Morawska, L., Ristovski, Z., Johnson, G., Jayaratne, R., Mengersen, K.,

2007b. A Novel Method for On Road Emission Factor Measurements Using a

Plume Capture Trailer, Environmental Science & Technology 41, 574-579.

Morawska, L., Thomas, S., Bofinger, N. D., Wainwright, D., Neale, D.,

1998b. Comprehensive characterisation of aerosols in a subtropical urban

atmosphere: particle size distribution and correlation with gaseous pollutants,

Atmospheric Environment 32(14/15), 2461-2478.

Morawska, L., Thomas, S., Gilbert, D., Greenaway, C., Rijnders, E., 1999b. A

study of the horizontal and vertical profile of submicrometer particles in

relation to a busy road, Atmospheric Environment 33(8), 1261-1274.

Morawska, L., Thomas, S., Jamriska, M., Johnson, G., 1999c. The modality of

particle size distributions of environmental aerosols, Atmospheric

Environment 33(27), 4401-4411.

Morawska, L., Vishvakarman, D., Swanson, C., 2007c. Diurnal variation of

PM10 concentrations and its spatial distribution in the South East Queensland

airshed, Clean Air, Submitted.

408

Napari, I., Noppel, M., 2002. An improved model for ternary nucleation of

sulfuric acid-ammonia-water, Journal of Chemical Physics 116, 4221-4227.

Nitta, H., Sato, T., Nakai, S., Maeda, K., Aoki, S., Ono, M., 1993. Respiratory

health associated with exposure to automobile exhaust. I. Results of cross-

sectional studies in 1979, 1982, and 1983, Archives of Environmental Health

48, 53-58.

Ntziachristos, L., Ning, Z., Geller, M.D., Sioutas, C., 2007. Particle

concentration and characteristics near a major freeway with heavy duty diesel

traffic, Environmental Science & Technology 41, 2223-2230.

O'Dowd, C., Aalto, P., Hameri, K., Kulmala, M. , Hoffmann, T., 2002.

Aerosol formation - Atmospheric particles from organic vapours, Nature 416.

O'Dowd, C., Facchini, M. C., Cavalli, F., Ceburnis, D., Mircea, M., Decesari,

S., Fuzzi, S., Yoon, Y.J., Putaud, J.P., 2004. Biogenically driven organic

contribution to marine aerosol, Nature 431, 676-680.

O'Dowd, C., Hoffmann, T., 2005. Coastal New Particle Formation: A Review

of the Current State-Of-The-Art, Environmental Chemistry 2, 245-255.

Paatero, P., Aalto, P., Picciotto, S., Bellander, T., Castano, G., Cattani, G.,

Cyrys, J., Kulmala, M., Lanki, T., Nyberg, F., Pekkanen, J., Peters, A.,

Sunyer, J., Forastiere, F., 2005. Estimating Time Series of Aerosol Particle

409

Numner Concentrations in Five HEAPSS Cities on the Basis of Measured Air

Pollution and Meterological Variables, Atmospheric Environment 39, 2261-

2273.

Pakkanen, T., Kerminen, V., Korhonen, C., Hillamo, R., Aarino, P.,

Koskentalo, T., Maenhaut, W., 2001. Urban and Rural Ultrafine (PM0.1)

Particles in the Helsinki Area, Atmospheric Environment 35, 4593-4607.

Pierson, W., Brachaczek, W.W., 1974. Airborne particulate debris from

rubber tyres, Rubber Chemistry and Technology 47, 1275-1299.

Pirjola, I., Paasonen, P., Pfeiffer, D., Hussein, T., Hameri, K., Koskentalo, T.,

Virtanen, A., Ronkko, T., Keskinen, J., Pakkanen, T., Hillamo, R., 2006.

Dispersion of Particles and Trace Gases Nearby a City Highway: Mobile

Laboratory Measurements in Finalnd, Atmospheric Environment 40, 867-879.

Pirjola, L., Laaksonen, A., Aalto, P., Kulmala, M., 1998. Sulfate aerosol

formation in the Arctic boundary layer, Journal of Geophysical Research 10,

8309-8322.

Pirjola, L., Parviainen, H., Hussein, T., Valli, A., Hameri, K., Aalto, P.,

Virtanen, A., Keskinen, J., Pakkanen, T., Makela, J., Hillamo, R., 2004.

"Sniffer" - A novel tool for chasing vehicles and measuring traffic pollutants,


410

Pitz, M., Kreyling, W., Holscher, B., Cyrys, J., Wichmann, H., Heinrich, J.,

2001. Change of the Ambient Particle Size Distribution in East Germany

between 1993 and 1999, Atmospheric Environment 35, 4357-4366.

Pohjola, M., Pirjola, L., Kukkonen, J., Kulmala, M., 2003. Modelling the

influence of aerosol processes for the dispersion of vehicular exhaust plumes

in a street environment, Atmospheric Environment 37, 339-351.

Pope, C.A., Burnett, R.T., Thun, M.J., Calle, E.E., Krewski, D., Ito, K.,

Thurston, G. D., 2002. Lung cancer, cardiopulmonary mortality, and long-

term exposure to fine particulate air pollution, Journal of the American

Medical Association 287(9), 1132-1141.

Raes, F., 1995. Entrainment of free tropospheric aerosols as a regulating

mechanism for cloud condensation nuclei in the remote marine boundary

layer, Journal of Geophysical Research 100, 2893-2904.

Rickeard, D J., Bateman, J.R., Kwon, Y.K., McAughey, J.J., Dickens, C.J.,

1996. Exhaust Particulate Size Distribution: Vehicle and Fuel Influence in

Light Duty Vehicles, SAE Papers, 961980 97-111.

Riipinen, S., Kulmala, M., Arnold, F., Dal Maso, M., Birmili, W., Saarnio, K.,

Teinil¨a, K., Kerminen, V., Laaksonen, A., Lehtinen, K., 2007. Connections

between atmospheric sulphuric acid and new particle formation during

411

QUEST III-IV campaigns in Heidelberg and Hyytiala, Atmospheric

Chemistry and Physics 7, 1899-1914.

Ristovski, Z., Jayaratne, E.R., Lim, M., Ayoko, G.A., Morawska, L., 2006.

Influence of diesel fuel sulphur on the nanoparticle emissions from city buses,


Ristovski, Z.D., Jayaratne, E.R., Morawska, L., Ayoko, G.A., Lim, M., 2005.

Particle and carbon dioxide emissions from passenger vehicles operating on

unleaded petrol and LPG fuel, Science of the Total Environment 345(1-3), 93-

98.

Ristovski, Z.D., Morawska, L., Ayoko, G. A., Johnson, G., Gilbert, D.,

Greenaway, C., 2004. Emissions from a vehicle fitted to operate on either

petrol or compressed natural gas, Science of the Total Environment 323(1-3),

179-194.

Rodriguez, S., van Dingenen, R., Putaud, J., Martins-Dos Santos, S., Roselli,

D., 2005. Nucleation and Growth of New Particles in the Rural Atmosphere

of Northern Italy - Relationship to Air Quality Monitoring, Atmospheric


Ronkko, T., Virtanen, A., Vaaraslahti, K., Koskinen, J., Pirjola, L., Lappi, M.,

2006. Effect of dilution conditions and driving parameters on nucleation mode

particles in diesel exhaust: Laboratory and on-road study, Atmospheric


412

Roorda-Knape, M. C., Janssen, N.A.H., de Hartog, J.J., van Vliet, P.H.N.,

Harssema, H., Brunekreef, B., 1998. Air pollution from traffic in city districts

near major motorways, Atmospheric Environment 32(11), 1921-1930.

Rosenbohm, E., Vogt, R., Scheer, V., Nielsen, O., Drieseidler, A., Baumbach,

G., Imhof, D., Baltensperger, U., Fuchs, J., Jaeschke, W., 2005. Particulate

size distributions and mass measured at a motorway during the BAB II

campaign, Atmospheric Environment 39, 5696-5709.

Ruuskanen, J., Tuch, T., Brink, H., Peters, A., Khlystov, A., Mirme, A., Kos,

G., Brunekreef, B., Wickmann, H., Buzorious, Z., Vallius, M., Kreyling, W.,

Pekkanen, J., 2001. Concentrations of Ultrafine, Fine and PM2.5 Particles in

Three European Cities, Atmospheric Environment 35, 3729-3738.

Sakurai, H., Tobias, H., Park, K., Zarling, D., Docherty, K.S., Kittleson, D.,

McMurray, P., Ziemann, P.J., 2003. On-line measurements of diesel

nanoparticle composition and volatility, Atmospheric Environment 37, 1199-

1210.

Salma, I., Dal Maso, M., Kulmala, M., Zaray, G., 2002. Modal characteristics

of particulate matter in urban atmospheric aerosols, Microchemical Journal

73, 19-26.

413

Sardar, S.B., Fine, P.M., Mayo, P.R., Sioutas, C., 2005. Size-Fractionated

Measurements of Ambient Ultrafine Particle Chemical Composition in Los

Angeles Using the NanoMOUDI, Environmental Science and Technology 39,

932-944.

Schneider, J., Hock, N., Weimer, S., Borrmann, S., 2005. Nucleation particles

in diesel exhaust: Composition inferred from in situ mass spectrometric

analysis, Environmental Science & Technology 39, 6153-6161.

Shi, J., Evans, D., Khan, A., Harrison, R., 2001a. Sources and Concentration

of Nanoparticles (<10nm Diameter) in the Urban Atmosphere, Atmospheric


Shi, J., Harrison, R., Evans, D., 2001b. Comparison of Ambient Particle

Surface Area Measurement by Epiphaniometer and SMPS/APS, Atmospheric


Shi, J., Harrison, R.M., 1999. Investigation of ultrafine particle formation

during diesel exhaust dilution, Environmental Science & Technology 33,

3730-3736.

Shi, J.P., Khan, A.A., Harrison, R.M., 1999. Measurements of ultrafine

particle concentrations and size distribution in the urban atmosphere, The

Science of the Total Environment 235, 51-64.

414

Shi, Q., 2003. Aerosol size distributions (3 nm to 3 μm) measured at the St.

Louis Supersite (4/1/01-4/30/02), Department of Mechanical Engineering.

Minneapolis, University of Minnesota.

Stanier, C., Khlystov, A., Pandis, S., 2004a. Ambient Aerosol Size

Distributions and Number Concentrations Measured During the Pittsburgh Air

Quality Study (PAQS), Atmospheric Environment 38, 3275-3284.

Stanier, C., Khlystov, A., Pandis, S., 2004b. Nucleation events during the

Pittsburgh air quality study: Description and relation to key meteorological,

gas phase, and aerosol parameters, Aerosol Science and Tecnhology 38 (S1),

253-264.

Sturm, P., Baltensperger, U., Bacher, M., Lechner, B., Hausberger, S., Heiden,

B., Imhof, D., Weingartner, E., Prevot, A., Kurtenbach, R., Wiesen, P., 2003.

Roadside measurements of particulate matter size distribution, Atmospheric


Suni, T., Kulmala, M., Hirsikko, A., Bergman, T., Laakso, L., Aalto, P.,

Leuning, R., Cleugh, H., Zegelin, S., Hughes, D., van Gorsel, E., Kitchen, M.,

Vana, M., Hõrrak, U., Mirme, S., Mirme, A., Twining, J., Tadros, C., 2007.

Formation and characteristics of ions and charged aerosol particles in a native

Australian Eucalypt forest, Atmospheric Chemistry and Physics Discussions

7, 10343-10369.

415

Thomas, S., Morawska, L., 2002. Size selected particles in an urban

atmosphere in Brisbane, Australia, Atmospheric Environment 36(26), 4277-

4288.

Tobias, H.J., Beving, D.E., Ziemann, P.J., Sakurai, H., Zuk, M., McMurry, P.,

Zarling, D., Waytulonis, R., Kittelson, D.B., 2001. Chemical Analysis of

Diesel Engine Nanoparticles Using a Nano-DMA / Thermal Desorption

Particle Beam Mass Spectrometer, Environmental Science & Technology 35,

2233-2243.

Tuch, T., Brand, P., Wichmann, H., Heyder, J., 1997. Variation of Particle

Number and Mass Concentration in Various Size Ranges of Ambient Aerosols

in Eastern Germany, Atmospheric Environment 31, 4193-4197.

Tunved, P., Hansson, H., Kerminen, V., Strom, J., Dal Maso, M., Lihavainen,

H., Viisanen, Y., Aalto, P., Komppula, M., Kulmala, M., 2006. High natural

aerosol loading over boreal forests, Science 312, 261-263.

Tunved, P., Hansson, H., Kulmala, M., Aalto, P., Viisanen, Y., Karlsson, H.,

Kristensson, A., Swietlicki, E., Maso, M., Strom, J., Komppula, M., 2003.

One Year Boundary Layer Aerosol Size Distribution Data from Five Nordic

Background Stations, Atmospheric Chemistry and Physics Discussions 3,

2183-2205.

416

Turpin, B.J., Hutzincker, J.J., 1995. Identification of Secondary Organic

Aerosol Episodes and Quantitation of Primary and Secondary Organic

Aerosol Concentrations during SCAQS, Atmospheric Environment 29(23),

3527-3544.

Vaaraslahti, K., Keskinen, J., Gieshaskiel, B., Solla, A., Murtonen, T. ,

Vesala, H., 2005. Effect of lubricant on the formation of heavy duty diesel

exhaust nanoparticles, Environmental Science & Technology 39, 8497-8504.

Vaaraslahti, K., Virtanen, A., Ristimaki, J., Keskinen, J., 2004. Nucleation

mode formation in heavy-duty diesel exhaust with and without a particulate

filter, Environmental Science & Technology 38, 4484-4890.

Vakeva, M., Hameri, K., Kulmala, M., Lahdes, R., Ruuskanen, J., Laitinen,

T., 1999. Street level versus rooftop concentrations of submicron aerosol

particles and gaseous pollutants in an urban street canyon, Atmospheric


Vardoulakis, S., Fisher, B. E. A., Pericleous, K., Gonzalez-Flesca, N., 2003.

Modeling air quality in street canyons: a review, Atmospheric Environment

37, 155 -182.

Vignati, E., Berkowicz, R., Palmgren, F., Lyck, E., Hummelshoj, P., 1999.

Transformation of size distributions of emitted particles in streets. The

Science of The Total Environment 235, 37-49.

417

Virtanen, A., Ronkko, T., Kannosto, J., Ristimaki, J., Makela, J., Keskinen, J.,

Pakkanen, T., Hillamo, R., Pirjola, L., Hameri, K., 2006. Winter and Summer

Time Distributions and Densities of Traffic Related Aerosol Particles at a

Busy Highway in Helsinki, Atmospheric Chemistry and Physics 6, 2411-

2421.

Vogt, R., Scheer, V., Casati, R., Benter, T., 2003. On-road measurement of

particle emissions in the exhaust plume of a diesel passenger car,


Wahlin, P., Palmgren, F., Dingenen, R., Raes, F., 2001. Pronounced Decrease

of Ambient Particle Number Emissions from Diesel Traffic in Denmark After

Reduction of the Sulphur Content in Diesel Fuel, Atmospheric Environment

35, 3549-3552.

Wåhlin, P., Palmgren, F., Van Dingenen, R., 2001. Experimental studies of

ultrafine particles in streets and the relationship to traffic, Atmospheric


Watson, J.G., Chow, J.C., Lowenthal, D.H., Kreisberg, N.M., Hering, S.V.,

Stolzenburg, M.R., 2005. Variations of nanoparticle concentrations at the

Fresno Supersite, Science of The Total Environment 358, 178-187.

418

Weber, R., Marti, J.J., McMurray, P., Eisele, F.L., Tanner, D.J. and Jefferson,

A. , 1996. Measured atmospheric new particle formation rates: Implications

for nucleation mechanisms, Chemical Engineering Communications 151, 53-

64.

Weber, R., Marti, J.J., McMurray, P., Eisele, F.L., Tanner, D.J., Jefferson, A.,

1997. Measurement of new particle formation and ultrafine particle growth

rates at a clean continental site, Journal of Geophysical Research 102, 4375-

4386.

Wehner, B., Birmili, W., Gnauk, T., Wiedensohler, A., 2002. Particle number

size distributions in a street canyon and their transformation into the urban-air

background: measurements and a simple model study, Atmospheric

Environment 36(13), 2215-2223.

Wehner, B., Wiedensohler, A., 2003. Long Term Measurements of

Submicrometer Urban Aerosols: Statistical Analysis for Correlations with

Meteorological Conditions and Trace Gases, Atmospheric Chemistry and


Westerdahl, D., Fruin, S., Sax, T., Fine, P., Sioutas, C., 2005. Mobile platform

measurements of ultrafine particles and associated pollutant concentrations on

freeways and residential streets in Los Angeles, Atmospheric Environment 39,

3597-3610.

419

WHO (2005). Guidelines for Air Quality. World Health Organization.

Wiedensohler, A., Wehner, B., Birmili, W., 2002. Aerosol number

concentrations and size distributions at mountain rural, urban-influenced rural

and urban-background sites in Germany, Journal of Aerosol Medicine 15(2),

237-243.

Woo, K.S., Chen, D.R., Pui, D.Y.H., McMurry, P.H., 2001a. Measurement of

Atlanta Aerosol Size Distributions: Observation of ultrafine particle events,

Aerosol Science and Technology 34(1), 75-87.

Woo, K.S., Chen, D.R., Pui, D.Y.H., Wilson, W.E., 2001b. Use of continuous

measurements of integral aerosol parameters to estimate particle surface area,

Aerosol Science and Technology 34(1), 57-65.

Young, L., Keeler, G., 2004. Characterization of Ultrafine Particle Number

Concentration and Size Distribution During a Summer Campaign in

Southwest Detroit, Journal of the Air & Waste Management Assoc. 54, 1079-

1090.

Yu, F., 2001. Chemiions and nanoparticle formation in diesel engine exhaust,

Geophysical Research Letters 28, 4191-4194.

Yu, F., Lanni, T., Frank, B.P., 2004. Measurements of ion concentration in

gasoline and diesel engine exhaust, Atmospheric Environment 38, 1417-1423.

420

Yu, F., Turco, R.P., 1997. The role of ions in the formation and evolution of

particles in aircraft plumes, Geophysical Research Letters 24, 1927-1930.

Yu, F., Turco, R.P., 2000. Ultrafine aerosol formation via ion-mediated

nucleation, Geophysical Research Letters 27, 883-886.

Yu, F., Turco, R.P., 2001. From molecular clusters to nanoparticles: Role of

ambient ionisation in tropospheric aerosol formation, Journal of Geophysical

Research 106(D5), 4797-4814.

Zhang, K., Wexler, A., 2004a. Evolution of Particle Number Distribution Near

Roadways Part I: Analysis of Aerosol Dynamics and its Implications for

Engine Emission Measurement, Atmospheric Environment 38, 6643-6653.

Zhang, K., Wexler, A., 2004b. Modeling the Number Distributions of Urban

and Regional Aerosols: Theoretical Foundations, Atmospheric Environment

(38), 1863-1874.

Zhang, K., Wexler, A., Niemeier, D., Zhu, Y., Hinds, W., Sioutas, C., 2005.

Evolution of Particle Number Distribution Near Roadways Part III: Traffic

Analysis and On-Road Size Resolved Particulate Emission Factors,


421

Zhang, K., Wexler, A., Zhu, Y., Hinds, W., Sioutas, C., 2004a. Evolution of

Particle Number Distribution Near Roadways Part II: The Road-to-Ambient

Process, Atmospheric Environment 38, 6655-6665.

Zhang, Q., Stanier, C., Canagaratna, M., Jayne, J., Worsnop, D., Pandis, S.,

Jiminez, J., 2004b. Insights into the Chemistry of New Particle Formation and

Growth Events in Pittsburgh Based on Aerosol Mass Spectrometry,

Environmental Science and Technology 38, 4797-4809.

Zhu, Y., Hinds, W., Kim, S., Shen, S., Sioutas, C., 2002a. Study of Ultrafine

Particles Near a Major Highway with Heavy Duty Diesel Traffic,


Zhu, Y., Hinds, W., Shen, S., Sioutas, C., 2004. Seasonal Trends of

Concentration and Size Distribution of Ultrafine Particles Near Major

Highways in Los Angeles, Aerosol Science and Technology 38, 5-13.

Zhu, Y., Hinds, W. C., Kim, S., Sioutas, C., 2002b. Concentration and Size

Distribution of Ultrafine Particles Near a Major Highway, Journal of the Air

& Waste Management Assoc. 52(9), 1032-1042.

Zhu, Y., Kuhn, T., Mayo, P., Hinds, W., 2006. Comparison of Daytime and

Nighttime Concentration Profiles and Size Distributions of Ultrafine Particles

Near a Major Highway, Environmental Science and Technology 40, 2531-

2536.

422

CHAPTER 8. CONCLUSIONS

8.1. INTRODUCTION

A myriad of diverse transport problems are occurring in urban areas, ranging from

traffic congestion to the challenges of urban sprawl. Some new initiatives that

have been introduced into urban living include higher density living and transit

oriented developments, however these developments are causing the public to

become more concerned about higher exposure rates to particulate matter,

including ultrafine particles (diameters < 0.1 µm), and the likelihood of increased

health risks.

Recent advances in vehicle technologies have seen dramatic reductions in

particulate matter emissions from motor vehicles in terms of particle mass

emissions, but these newer technologies can often be associated with increases in

the smaller particle size ranges (Morawska et al. 2004). These smaller sized

particles, due to their size, are particularly adept at traversing and lodging deep in

the human respiratory system, which can lead to serious health effects in the

human body.

The currently very high particle emission rates of HDVs are a global problem, and

one for which solutions are urgently needed. Solutions can range from

technological improvements, such as installing after-treatment devices; to

adopting less polluting options such as moving freight using electric and hybrid

HDVs in the same unit or vehicle for multiple modes (eg., road, road and water)

423

(Macharis et al. 2007). HDV fleets emit more than an order of magnitude more

particulate matter than LDVs (Morawska et al. 2004) ; and their exhaust is a

declared cancer-causing substance (Swiss Clean Air Act 2000;

www.dieselnet.com/standards/ch).

Most motor vehicle particle emissions are ultrafine size and are not currently

regulated by air quality standards. Current ambient air quality standards in terms

of concentrations are based on the findings of epidemiological studies, which

have shown that airborne particle mass has a linear exposure-health response

relationship. Based on an American Cancer Society study (Pope et al. 2002) the

World Health Organization has set new particulate matter guidelines with annual

mean values for PM2.5 and PM10 of 10 and 20 µg m-3 respectively, and these

guidelines relate to the lowest end of the range across which significant effects on

survival have been observed (WHO 2005).

Inventories and air quality standards are needed to manage and control fleet

particle emissions. Inventory data provides vital information for land use and

transport planning and decision-making. It informs the development of relevant

air quality guidelines and standards and air quality assessments, and is useful for

modelling the air quality implications of changes in fleet composition and travel

demand, and changes such as the introduction of new vehicle technologies and

fuels.

424

The public and policymakers need accurate information on the health

consequences of transport and land use policies, and health professionals have an

important role to play in providing this information and in making assessments of

the health impacts of transport policies (Dora 1999). One of the current challenges

facing us today is that no one has ever quantified how much particulate matter is

emitted from motor vehicle fleets in terms of both particle number and particle

mass emissions. A comprehensive inventory covering the full particle size range

emitted by motor vehicle fleets has not been published.

8.2. PRINCIPAL SIGNIFICANCE OF THE FINDINGS

The research reported in this thesis significantly advances our

understanding of the extent of particulate matter pollution emitted from an

urban fleet. The original and significant contribution of this research

included:-

(i) Developing the first published inventory which comprehensively

quantified the total particulate matter emitted by a motor vehicle

fleet in terms of both particle number emissions and emissions

for different particle mass size fractions.

(ii) The novelty of this study relates to its successful integration of

expertise from two distinctly separate disciplines, and

development of an approach for quantifying vehicle fleet particle

emissions which takes into account all the elements. The method

included derivation of a comprehensive set of particle emission

factors for particle number and particle mass for different vehicle

425

and road type combinations. These emission factors can be used

to develop inventories and quantify the spatial distribution of

particle concentrations in urban areas of developed countries.

(iii) Investigating the location of the mode in a range of different

worldwide environments and for different particle metrics,

including traffic-influenced environments; and demonstrating the

suitability of examining modes as an effective basis for

developing air quality regulation.

(iv) Providing evidence that a PM1 mass ambient air quality standard

would suit the majority of worldwide environments and that, in

combination with PM10, is likely to be a more useful set of

standards than the current standards of PM2.5 and PM10 for

discriminating mechanical and combustion-generated particles,

such as emitted by motor vehicles.

(v) Reviewing and synthesizing existing knowledge on ultrafine

particles as it relates to motor vehicles, and the implications of

the findings of the review for exposure and epidemiological

studies, and for identifying the future directions of research

needed on ultrafine particles.

Figure 8.1 depicts the research activities undertaken in this study.

426

INVENTORY

OF EMISSIONS (MODEL)

Validation (by comparing with EPA

and other inventories) PAPER 3

Emission factors for different particle

size ranges

Travel demand model

PAPER 2

Model future scenarios (passenger & freight vehicles) to examine emission implications PAPER 3

PAPER 3

Modality: Investigation of the fractional contribution of motor vehicle emissions to different particle mode distributions

Paper 1: Modality and the fractional contribution of vehicle emissions to different particle modes; and modality within particle size distributions in a wide range of different worldwide environments.

Paper 2: Derivation of suitable emission factors to use in transport modelling and inventory development.

Paper 3: Development of a motor vehicle particle emissions inventory and validation of the inventory.

Paper 4: Review and synthesis of current knowledge on ultrafine particles, with a specific focus on motor vehicles.

Notes: 1. The dashed rectangle depicts future activities recommended based on the study results. 2. The dotted rectangle depicts a Government prototype travel demand model.

Figure 8.1 Diagram of Research Activities. 426

Modality and air

quality and emission standards for vehicles

PAPER 1

Review & synthesis of current knowledge on ultrafine particles, with specific focus on motor vehicles

PAPER 4

427

The first step of this study involved investigating modality within particle size

distributions as a potential basis for developing air quality standards, based

on gaining an understanding of particle mechanisms and source apportionment in

terms of particle size distributions. The location of the mode in a wide range of

different worldwide environments was examined for different particle metrics,

including traffic-influenced environments, and the fractional contribution of

motor vehicle particle emissions to different modes.

The second step involved deriving a comprehensive set of particle emission

factors for different particle sizes for motor vehicles which can be used in

transport modelling and air quality assessments to quantify fleet particle

emissions in urban areas. Average particle emission factors, and their 95%

confidence intervals, were derived from statistical models developed for different

vehicle and road type combinations and particle metrics.

The third step involved combining the most suitable average emission factors

produced by the statistical models with transport demand model data to

develop a road link-based inventory of tailpipe particle emissions emitted from

the urban South-East Queensland motor vehicle fleet in 2004. Where available,

the inventory quantification was validated with a relevant local model.

The fourth step involved modelling future scenarios, including different

proportions of passengers travelling in light duty vehicles and buses, to examine

the air quality implications of these scenarios, and deriving an estimate of fleet

particle emissions in 2026.

428

The fifth step involved reviewing and synthesizing current knowledge on

ultrafine particles as it relates to motor vehicles, and identifying the

implications of its findings in terms of important future directions for research

needed on ultrafine particles, and on air quality regulation, epidemiological

studies and human impact assessments.

The rectangle in Figure 8.1 depicted with dotted lines labelled ‘Travel demand

model’ relates to a Government prototype travel demand model; and the rectangle

shown by dashed lines and labelled ‘Modality and air quality and emission

standards for vehicles’, represents future research that is recommended. This

research could relate to determining the suitability of examining modes within

particle size distributions for different vehicle types as a basis for developing

vehicle standards; examine relevant vehicle characteristics such as fuel type,

vehicle technologies (eg., aftertreatment devices) etc which could be used to

develop vehicle standards and source signatures for source apportionment related

to different vehicle classes; and examine factors specifically related to ultrafine

particles emitted by different motor vehicle types.

This PhD research has provided evidence that a combination of PM1 and PM10

mass ambient air quality standards have the potential to be a more suitable and

discerning combination of air quality standards to control combustion and

mechanically-generated particle mass emissions than the present standards of

PM2.5 and PM10. The review of modality within particle size distributions

examined in this study showed that ultrafine particles, measured in terms of

429

particle number, were the dominant source of motor vehicle particle emissions

and hence the urban South-East Queensland (SEQ) inventory was developed to

include particle number.

The urban SEQ inventory provides valuable knowledge not only to understand

total emissions in the region, but also constitutes the first published

comprehensive inventory of motor vehicle particle emissions prepared for a

vehicle fleet, and includes the first published comprehensive particle number

inventory. The inventory provides regulators and planners with benchmark values

of the levels of particulate pollution emitted from the motor vehicle fleet in 2004,

which can be used to test future alternative transport and land use strategies, and

to design of health impact assessments. This inventory data is also useful

information as a basis for developing air quality standards and standards for

motor vehicles.

The comprehensive set of particle emission factors derived in this study has

application for other regions, particularly where there may be a lack of data or

insufficient measurement data to use in estimating inventories and for air quality

assessments. Estimating motor vehicle particle emissions inventories is important,

in order to understand and control human exposure, and to inform the

development of land use and transport planning, as well as to provide data for the

development of relevant ambient air quality standards and standards.

430

The outcomes of this research provide a multi-pronged approach to managing

fleet particulate matter pollution, firstly by developing a method to quantify the

pollution, secondly by devising methods to develop standards to control the

pollution, and thirdly by identifying key areas of research needed in relation to

understanding and quantifying ultrafine particles generated by motor vehicles.

8.3. THE PRINCIPAL FINDINGS AND SIGNIFICANCE OF THIS STUDY

These are summarized below grouped into nine subsections:-

(i) modality within particle size distributions;

(ii) a new method for developing comprehensive particle inventories;

(iii) derivation of suitable particle emission factors;

(iv) statistically significant differences found between mean values of

published emission factors;

(v) other particle emission inventories;

(vi) gaps in our knowledge related to motor vehicle emission factor data;

(vii) development of a comprehensive particle emissions inventory;

(viii) validation of the developed inventory;

(ix) scenario modelling findings and likely future particle emissions.

(x) review and synthesis of current knowledge on ultrafine

particles, with a specific focus on motor vehicles.

431

Table 8.1 appears at the end of this Section and provides a précis of principal

findings and the significance of their application.

432

MODALITY WITH PARTICLE SIZE DISTRIBUTIONS

1. The research found that in marine-influenced, modified background,

suburban background, traffic-influenced, urban-influenced and vegetation

burning environments examined in urban South-East Queensland, a clear

separation existed between the accumulation and coarse modes for particle

volume and particle number size distributions at around 1 µm. A similar

clear and distinct separation was also found between the modes at around

1 µm in an examination of 600 modal location values reported in the

international literature for a wide range of worldwide environments for

particle number, surface area, volume and mass size distributions. These

findings demonstrate the relevance of developing a PM1 mass ambient air

quality standard and its suitability for the majority of environments

worldwide, including traffic-influenced environments. The research also

demonstrated the usefulness of examining modes within particle size

distributions as a basis for developing air quality regulations, and its value

in providing information about contributions from different pollution

sources and for understanding particle mechanisms.

433

2. Another important finding of the research was that PM1 and PM10 mass

measurements enabled a clearer distinction to be made between

mechanically-generated and combustion-generated particles than the

current standards of PM2.5 and PM10. An investigation of the fractional

contribution of particle mass from marine-influenced, modified

background, suburban background, traffic-influenced, urban-influenced

and vegetation burning environments in South-East Queensland, and

different particle modes in particle size distributions to PM1, PM2.5 and

PM10 revealed that PM2.5 measurements may not be an adequate parameter

as a basis for a standard to control particle emissions and concentrations.

It was found that in all South-East Queensland environments examined the

division at 2.5 µm cut across the coarse particle mode close to its peak.

PM2.5 coarse mode measurements provided information mainly on

mechanically-generated processes, but for some environments

contributions from the combustion process modes (nucleation and

accumulation modes) were significant. These results make evident the

substantial complexity which may be associated with interpreting PM2.5

data in order to distinguish contributions from different sources.

434

3. The study found that PM1 mass contributions from marine-influenced,

modified background, traffic-influenced and vegetation burning

environments were from particles in the nucleation and accumulation

modes, and not from particles in the coarse mode. The PM1 measurements

provided very good information about nucleation and accumulation mode

particles (such as from combustion sources, eg., motor vehicles) and

enabled a clearer distinction to be made between combustion and

mechanically-generated aerosols.

4. Another finding of the study was that contributions to PM10 mass from all

urban South-East Queensland environments examined (with the exception

of vegetation burning) were mainly from coarse mode particles generated

from mechanical processes (such as particle resuspension by motor

vehicle traffic or production from mechanical wear and tear of tyres), with

negligible contribution from combustion processes. Therefore PM10

measurements were found to be suitable for discriminating mechanically-

generated particles in the coarse mode.

5. The findings of the modality review and examination of the urban South-

East Queensland study and other studies conducted around the world (1-4

above) make a unique and significant contribution to knowledge by

identifying that PM1 and PM10 mass standards may be a more suitable and

discerning combination of air quality standards than the current standards

of PM2.5 and PM10 for controlling particle emissions and concentrations,

435

including in traffic-influenced environments, and that a PM1 mass standard

would suit the majority of environments worldwide. Although few data are

available on PM1 concentrations, measurement technologies are currently

available to undertake measurements in this size range.

6. Examination of the location of the modes in traffic-influenced environments

revealed that most modes were found in the submicrometre size range, and

were dominant in the ultrafine size range. As most motor vehicle particle

emissions are < 1 µm and concentrated in the ultrafine size range, further

scientific investigation is needed to identify the best possible combination of

particle number and particle mass standards to control motor vehicle

emissions. Future legislation needs to consider the inclusion of particle

number standards, in addition to particle mass standards, to control motor

vehicle particle emissions, as ultrafine particles have negligible mass, but

are prolific in terms of their numbers. Therefore particle number-based

standards are very relevant for controlling motor vehicle emissions.

436

A NEW METHOD FOR DEVELOPING COMPREHENSIVE

INVENTORIES

7. A new method for developing comprehensive inventories of motor vehicle

particle emissions was devised in this study, which enables quantification

of emissions in terms of the full size range of particles emitted for both

particle number and particle mass. The method included derivation of a

comprehensive set of average particle emission factors for different

vehicle and road type combinations for particle number, particle volume,

PM1, PM2.5 and PM10 to use in transport modelling and air quality

assessments.

The derivation of average particle emission factors contribute significant

knowledge as these emission factors have application not only for urban

regions in the developed world, but for regions which do not have specific

location or application data, or have measurement data of insufficient

scope. They can be combined with traffic data to produce road link-based

inventories, and used as input values to motor vehicle mobile emission

inventories to quantify the spatial distribution of particle concentrations in

urban areas. Without the development of these inventories land use and

transport planners will have a very limited capacity to consider the

important and severe health effects of exposure to particulate matter in

land use and transport planning decisions.

437

DERIVATION OF SUITABLE PARTICLE EMISSION FACTORS

8. The data reported in the international literature showed that there were

more than 900 particle emission factors found for different vehicle types

cited in 59 papers. From this review, sixteen model variables were

developed, based on variables measured in several of the papers, and data

relating to 667 particle emission factors were classified and examined in a

statistical analysis. Five separate statistical models were developed, which

produced average particle emission factors, and these statistical models

were found to explain 86%, 93%, 87%, 65% and 47% of the variation in

published emission factor values for particle number, particle volume,

PM1, PM2.5 and PM10 respectively. The sixth model for total particle

mass was found to be a null model, which is likely to relate to the fact that

the emission factors were derived for total particle mass, and not for more

discrete sub-sets of particle mass fractions.

The explanatory variables for the five statistical models were Vehicle Type

and Instrumentation for particle number and PM2.5; Vehicle Type and Fuel

Type for PM1; Vehicle Type, Size Ranged Measured and Speed Limit on

the Road for particle volume; and Vehicle Type and Road Type for PM10.

The results of the statistical analysis provide important information

on key factors which may have a major influence on the values of derived

emission factors. Such key factors require special attention in the design,

conduct and reporting of results of emission factor studies.

438

9. The five statistical models produced average particle emission factors, and

their corresponding standard error and 95% confidence intervals, for

different vehicle and road type combinations for particle number, particle

volume, PM1, PM2.5 and PM10, for estimating particle emissions emitted

per vehicle per kilometre travelled. The most suitable emission factors to

use in transport modelling and air quality assessments were selected based

on examination of the statistical characteristics of the average emission

factor values, and for some particle metrics Size Range Measured and

Road Type were also considered. These emission factors have universal

application, and are suitable to be used for urban regions of developed

countries, and in particular for regions which have little or no

measurement data to develop particle inventories.

10. The most suitable average particle emission factors to use in transport

modelling and air quality assessments are outlined below, those relating to

HDVs and buses related principally to diesel-fuelled vehicles, and for

LDVs principally to petrol-fuelled vehicles.

(i) particle number derived from Condensation Particle Counter

measurements for Fleet, HDV and LDV and from Scanning Mobility

Particle Sizer measurements for Diesel buses;

(ii) particle volume related to Fleet, HDV and LDV emission factors

for roads with Speed Limits on the Road of ≤ 60 km/hr where the Size

Range Measured was 18-300nm, and > 60 km/hr related to Size Range

Measured of 18-700nm;

(iii) PM1 for Fleet, LDV and HDV based on Fuel Type.

439

(iv) PM2.5 for Fleet were measured using Tapered Element Oscillating

Microbalances and the Differential Mobility Analyzer; for LDV measured

by DustTrak and for HDV related to the overall average emission factor

value derived from all Instrumentation used in PM2.5 measurements; and

(v) PM10 for Fleet, HDV and LDV related to different Road Types,

including freeway, highway, motorway, rural road, tunnel and urban road;

and for buses to boulevard, and urban Road Types and dynamometer

measurements. The authors of the boulevard and urban Road Type study

for buses reported that they considered their very high values of PM10

emission factors were influenced by contributions from resuspended road

dust and, within each vehicle category, by the effects of speed and

acceleration. Hence as the average emission factor for buses measured on

dynamometers was more conservative than those derived for the two Road

Types, and less likely to be affected by resuspended road dust, the

dynamometer emission factor was considered a more suitable emission

factor for buses for PM10 for all Road Types.

440

STATISTICALLY SIGNIFICANT DIFFERENCES IN MEAN

VALUES OF PUBLISHED EMISSION FACTORS

11. The research found that for the majority of particle metrics no statistically

significant differences were found between the mean values of published

emission factors for different categories of Country of Study and Study

Location (eg., measured on a dynamometer, in a tunnel or in the vicinity

of the road). The exceptions related to statistically significant differences

found between the mean values for PM2.5 for Country of Study between

Australian and Other Countries (Austria, Belgium, Denmark, Germany,

Sweden, Switzerland and the United Kingdom) studies and between

Australian and USA studies. However, it is crucial to note these

differences are likely to be influenced by the fact that the majority of the

Australian PM2.5 emission factors were derived for diesel vehicles.

Statistically significant differences were also found for PM1 between

dynamometer and vicinity of the road and between dynamometer and

tunnel mean values. These differences are likely to be influenced by the

fact that the PM1 dynamometer measurements related exclusively to LDV

and HDV diesel vehicles tested in Australia. Higher values of emission

factors are likely to be associated with diesel-fuelled vehicles as compared

to petrol and other fuelled-vehicles. More data and studies are needed for

PM1 as there was insufficient data to test Country of Study, and no bus

emission factors available for this metric.

441

An important finding of this research is that relatively few statistically

significant differences were found between the mean values of published

emission factors for the different particle metrics for different Countries of

Study and Study Locations. This finding reinforces the value and universal

application of the average particle emission factors derived in this research

work, and their application for other urban areas, including for areas that

have limited measurement data and emission factors for developing motor

vehicle particle emission inventories.

12. No statistically significant differences were found between the means of

published emission factors relating to vehicle emissions measured on a

dynamometer and those measured on different Road Types for all particle

metrics (particle number, particle volume, total particle mass, and PM1,

PM2.5, PM10), with only two exceptions. Firstly, statistically significant

differences were found between the mean values for published emission

factors for PM1 between motorway and the four Road Types - rural area,

highway, tunnel and urban; and between motorway and dynamometer.

Secondly, statistically significant differences were found for PM10

between boulevard and the Road Types – highway, motorway, tunnel and

urban, and between boulevard and dynamometer. However, it should be

noted that these statistically significant differences found for motorway

and boulevard Road Types may have been influenced by high vehicle

speed scenarios. Therefore, the relatively few statistically significant

differences found between published emission factors derived from

442

dynamometer and most Road Types suggests that these two methods provide

generally similar results.

13. No statistically significant differences were found for particle number, total

particle mass, and PM1, PM2.5, PM10 between the mean values for published

emission factors for dynamometer studies and four Road Classes (based on

Speed Limit on the Road, where urban roads had limits of ≤ 60 km/hr, non-

urban > 60 km/hr; highway ≥ 80 km/hr and non-highway < 80 km/hr).

These statistical tests did not include total particle mass for highway and

non-highway or dynamometer for particle volume due to insufficient data

being available. Although few statistical differences were found, it is

important that future studies investigate the Average Vehicle Speeds

associated with their derived emissions factors in on-road studies, to provide

more realistic information about actual driving conditions and vehicle

speeds, in addition to reporting the Speed Limit on the Road.

14. For Vehicle Type statistically significant differences were found between the

means of published emission factors for Fleet and HDV for particle number,

PM1, PM2.5; between the means for Fleet and LDV for PM2.5; and between

LDV and HDV for all particle metrics (particle number, particle volume,

PM1, PM2.5 and PM10). This finding highlights the wide gap between the

mean particle emission factors for LDVs and HDVS, and emphasizes the

high emission rates of HDVs and the need to limit population exposure to

emissions from this Vehicle Type.

443

15. For buses statistical tests showed that the mean values of published

emission factors for particle number, total particle mass, and PM2.5, PM10

were statistically similar to the mean values for Fleet, LDV and HDV.

The mean values for buses for particle number are likely to be influenced

by the fact that the bus emission factors were derived from Scanning

Mobility Particle Sizer measurements (Instrumentation which places a

major focus on determining particle size distribution) and which measures

a different size range to Condensation Particle Counter (CPC). Whereas

the Fleet, LDV and HDV emission factors for particle number were

derived using the Condensation Particle Counter (Instrumentation which

focuses on determining total particle count, including down to the very

small particle size range of 3 nm, where particle numbers tend to be very

prolific). In addition, the same size of values for buses were considerably

lower than those for Fleet, LDV and HDV. Greater differences between

mean values for buses and other Vehicle Types would be expected in the

future when vehicle standards are introduced that require the fitting of

technologies such as particle filters to diesel vehicles. No data was

available to examine the means for PM1 and particle volume for buses.

16. No statistically significant differences were found between the means

of published emission factors for different Fuel Types for particle

number, or between the means for total particle mass. However,

statistically significant differences were found between petrol and

diesel in PM10. Fuel Types were not reported in particle volume

444

studies and as PM1 had fewer than three Fuel Type groups, multiple

comparison tests were not able to be conducted.

17. Statistical tests showed no significant difference between the mean values

for published emission factors measured by different Instrumentation for

PM2.5 and total particle mass. However, a very important finding of this

research was that statistically significant differences were found between

the mean values for published emission factors measured by Condensation

Particle Counter (CPC) and the Scanning Mobility Particle Sizer (SMPS)

for particle number of 22.69 x 1014 particles per vehicle per km and 2.08 x

1014 particles per vehicle per km, respectively. This finding highlights

major differences in the measurements of these two Instrumentation, and

is a finding that needs to be addressed as a much broader issue. One

possible explanation, however, is that the lower and upper size window for

measurement is determined by instrument capability and operator choice.

In the case of the lower size window for the SMPS this is usually set

higher, commonly in the range 0.010-0.02 µm, than that for the CPC

which ranges from 0.002-0.01 µm, which means that generally the CPC

measures the nucleation mode and the SMPS does not. The nucleation

mode tends to be prolific in terms of particle number emissions for motor

vehicles; as evidenced by larger emission factor values derived for LDVs

and HDVs motor vehicles relating in the < 0.018 µm size range

(Morawska et al. 2008).

445

In relation to Size Range Measured for particle number no statistically

significant differences were found between the means of published

emission factors for the lower and upper size ranges measured for the

various levels of each of the categorical variables, after accounting for the

associated variability of these estimates. This means that for studies which

reported both the lower and upper size ranges for Instrumentation used, no

statistically significant differences were found between the means in terms

of the different size ranges measured. However, further studies may be

needed that focus on the nucleation mode, in particular less than 18 nm,

where particle number emissions tend to be very prolific (please refer

point 17 above).

OTHER PARTICLE EMISSION INVENTORIES

18. A detailed review of the international literature found only one example of

a study that attempted to develop an inventory of particle emissions for a

motor vehicle fleet. This was prepared for the UK in 1996, 1998 and 2001

(Group 1999; Goodwin et al. 2000; AQEG 2005). However, this estimate

used emission factors for the smaller particle size ranges derived by

applying distribution profiles for these size ranges to PM10 estimate data,

and not based on specific, individual measurement data for the different

particle size ranges. As these UK estimates were based on such few input

data and were not comprehensively assessed, they cannot be considered

one inventory.

446

Therefore, another very significant and unique contribution of the work of

this PhD study is derivation of a comprehensive set of particle emission

factors for different vehicle and road type combinations covering the full

size range of particles emitted. These were derived for both particle

number and particle mass, based on measurement data for different

particle size ranges, and used to develop the a comprehensive inventory of

motor vehicle particle emissions presented in this PhD research study,

which is the first comprehensive inventory that has been published.

GAPS IN OUR KNOWLEDGE OF MOTOR VEHICLE EMISSION

FACTORS

19. The research identified a number of important gaps in our current

scientific knowledge about motor vehicle particle emissions. Limited

emission factor data was found for particle volume, particle surface area,

PM1, brake and tyre wear, road grade, engine power and for vehicles

travelling in congested traffic conditions at speeds < 50 km per hour.

There was also limited information available in the literature on methods

which would enable discrimination of resuspended road dust from motor

vehicle tailpipe emissions. Further research work on these areas is needed

to develop emission factors related to these different factors for tailpipe

and non-exhaust emissions, mentioned above.

447

20. No relevant emission factors were found for buses for particle number

where Instrumentation had measured total particle count down to the

smaller size range of 3 nm, nor for particle volume or PM1. Further

research is needed to derive emission factors for buses for particle number

(including relevant size sub-classes in the range 3 nm-1 µm) and for

particle volume, particle surface area, and for different mass size fractions.

This research needs to derive emission factor values for different vehicle

speeds and engine loads, including for speeds less than 50 km/hr, so that

bus emissions can be accurately modelled for congested traffic conditions;

as well as for different road environment conditions such as different road

or tunnel gradients (slopes).

21. Very limited emission factor data is available in the international literature

for speed-related emission factors for vehicles travelling at speeds < 50

km/hr. Few studies have reported the Average Vehicle Speed of on-road

fleets or different vehicle types, such as for LDVs, HDVs and buses, or

even reported the Speed Limit on the Road. Authors may have assumed

that Speed Limit on the Road would be implied by the Road Type studied.

These practices have contributed to a lack of available data on speed-

related emission factors in the international literature. New Drive Cycle

tests are needed which focus specifically on simulating different driving

speeds and driving patterns commonly associated with congested traffic

conditions, to bridge the gap in knowledge.

448

DEVELOPMENT OF A COMPREHENSIVE PARTICLE

INVENTORY

22. The inventory developed in this PhD project quantified total particulate

matter emissions for the urban South-East Queensland fleet for 2004

covering the full size range of particle emitted, and included particle

number and different particle mass fractions. The method developed in

this research project was used and a comprehensive set of particle

emission factors (derived in this research) were combined with transport

modelling data, to quantify emissions for different vehicle and road type

combinations for particle number, PM1, PM2.5 and PM10. Traffic data used

in the inventory related to vehicle kilometres travelled (VKT) and was

sourced from a government prototype model, the Brisbane Strategic

Transport Model, which covered an area of approximately 4600 square

kilometres. The study region had a resident population of 1.7 million and

1.2 million motor vehicles (ABS 2004 a,b). Analysis of the VKT travelled

in the region revealed that 93.3% of VKT was contributed by LDVs, 6.3%

by HDVs and 0.4% by buses. The bus fleet in 2004 comprised 89%

Diesel and 11% CNG buses (Translink 2007). It was found that although

LDVs dominated total VKT, HDVs made the most significant contribution

to particle emissions in the smaller size ranges, including PM1 and particle

number (please refer to in point no 24).

449

23. A major finding of this research was that the development of the urban

South-East Queensland inventory revealed that HDVs were the major

emitters of particle number and PM1 emissions, and although they only

travelled about 6% of the total distance travelled by the vehicle fleet, they

contributed more than 50% of the region’s particle number and PM1

emissions. This is a very significant finding, particularly also given that

the freight task is expected to double in the next 20 years (SKM 2006).

This finding strongly emphasizes the need to expend major effort to

reduce HDV emissions both now and in the future, particularly as HDVs

tend to emit around 20 times more particles than LDVs in the smaller

particle sizes ranges < 1 µm.

24. The developed inventory for urban South-East Queensland showed

that in 2004 total annual particle emissions emitted by the fleet were

for :-

(i) particle number were 1.08 (0.54-1.97) x 1025 per annum. Of these

total particle number emissions total HDVs contributed

approximately 54%, total LDVs 45% and total buses close to 1%.

No studies were found which can be compared to the urban SEQ

particle number inventory, and the only one available estimated

particle flux from all sources, from both natural and

anthropogenic sources and as data is not available on particle flux

from natural sources in urban SEQ, the studies cannot be

compared.

450

The inventory quantified in this PhD study for particle number for

urban South-East Queensland constitutes the first detailed, published

particle number inventory of motor vehicle particle emissions.

(ii) PM1 were 477 (233-964) tonne, which did not include Diesel or CNG

buses as relevant PM1 emission factors were not available. Although

total HDVs only contributed around 6% of the total VKT travelled in

the region, they contributed 55% of total PM1 emissions, and total

LDVs contributed 45%. Again, this high contribution by total HDVs to

submicrometre particle emissions in the region further supports the

need to focus on reducing HDV particle emissions in the study region.

As discussed previously, research conducted as part of this PhD

project found that PM1 and PM10 are likely to be a more relevant and

discerning combination of air quality standards than the current

standards of PM2.5 and PM10 for combustion sources such as motor

vehicles. It is therefore important to conduct further studies to derive

emission factors for bus and other vehicle types for PM1 mass, and to

take steps to introduce air quality standards for motor vehicles for the

PM1 mass size range;

451

(iii) PM2.5 were 736 (225-1436) tonne, and CNG buses were not included as

relevant emission factors were not available. Total LDVs contributed 61%

of PM2.5 emissions, total HDVs 37% and total Diesel buses 2%; and

(iv) PM10 were 2614 tonne, with an upper 95% confidence interval value of

9668 tonne. Total LDVs contributed 81%, total HDVs 18% and total

buses contributed approximately 1%.

25. The developed inventory for urban South-East Queensland found that total

LDVs emissions dominated the total PM2.5 and PM10 inventories and total

emissions on lower speed roads (urban roads). This was influenced by the

fact that total LDV VKT was almost double that of total HDV VKT on

these roads. Whereas, total HDVs were found to dominate the total

particle number and PM1 emission inventories and total emissions on

higher speed roads (urban-major roads). This was influenced by the

higher value for the HDV emission factor, which was almost 6 times

higher than the emission factor for LDVs, and total HDV VKT was

slightly higher than total LDV VKT on this Road Type. Although total

buses only contributed around 1-2% of total emissions, nevertheless

quantification of these emissions at the local scale is important due to high

localized exposure by populations in busways, tunnels and on roads.

452

VALIDATION OF THE DEVELOPED INVENTORY

26. The study found no local models available which had quantified total

annual particle number, PM1 or PM2.5 that could be compared with the

quantifications derived in this research. However, a comparison was able

to be made between the PM10 inventory developed for urban South-East

Queensland in this research of 2614 tonne per annum and the inventory

derived by the Queensland Environmental Protection Agency for South-

East Queensland for the year 2000 of 2249 tonne (EPA 2004). These two

PM10 inventories are in close agreement and suggest confidence can be

had in the total particle inventory developed for urban South-East

Queensland in this PhD research.

SCENARIO MODELLING FINDINGS & LIKELY FUTURE PARTICLE

EMISSIONS

27. Other important findings of this research related to the modelling of

the particle emission implications of future scenarios for urban South-

East Queensland. This modelling was considered important as the

region has major busway, tunnel and road infrastructure under

construction and many initiatives promoting shifts from passenger car

travel to buses. The modelling conducted in this research found that

reductions of:-

453

(i) 3-4%, 1-2%, 1-6% in particle number, PM2.5 and PM10 emissions,

respectively, were associated with each 10% reduction in LDV

VKT, where 70-100% of these passengers were moved to new

buses; and

(ii) reductions of 2%, 2-3% and 3-4% in particle number, PM2.5 and

PM10 emissions, respectively, were associated with each 10%

reduction in work trips (trips to and from work from home), where

50% of these passengers were moved to new buses.

It could be expected, therefore, that even greater reductions in particulate

matter emissions could be achieved by moving these LDV passengers to

existing buses and increasing the average vehicle occupancy rates of

buses. In 2004 the average vehicle occupancy rate for LDVs was 1.5

passengers and 15.5 passengers for buses in the average 24 hour period

(Translink 2007). This bus occupancy rate is considerably less than half

the maximum carrying capacity of most buses in the fleet. In addition, a

small reduction in the percent of HDV VKT would lead to substantial

particle emission reductions, particularly in the region’s total particle

number and PM1 emissions.

28. The research found that on a per passenger per km basis particle number

emissions from LDVs were 1-2 orders of magnitude higher than those for

buses. LDV emissions per passenger per km were similar to Diesel buses

for PM2.5; and for PM10 were close to 5 times those of Diesel buses, and

several orders of magnitude higher than CNG buses. Even greater

454

differences between LDV and bus emission factors per passenger per km

could be expected if bus occupancy rates increased. These findings

highlight major opportunities for reducing particle number and particle

mass emissions by shifting proportions of LDV passengers to buses. As

relevant PM1 emission factors for buses were not available, these were not

included in the modelling.

29. Scenario modelling undertaken in this research to estimate expected

emissions in urban South-East Queensland in 2026 revealed that, when

compared to the 2004 inventory, an 100-fold increase could be expected in

particle number emissions in 2026 and reductions of 38%, 36% and 31%

in PM1, PM2.5 and PM10 respectively. The results of this modelling

further emphasize the need to focus on strategies to dramatically reduce

HDV emissions, such as mandatory fitting of particle filters, limiting HDV

access to roads situated in close proximity to populations, and finding

alternative lower polluting options for moving freight. The majority of

HDVs presently are diesel-fuelled; and in Switzerland diesel exhaust is

classified as a carcinogen (www.dieselnet.com/standards/ch/). The very

high proportion of the HDV fleet contributions to the smaller particle size

range in urban South-East Queensland is a major concern from a health

effects perspective. More action is needed to regulate HDV emissions and

limit population exposure.

455

REVIEW AND SYNTHESIS OF CURRENT KNOWLEDGE ON

ULTRAFINE PARTICLES, WITH A SPECIFIC FOCUS ON MOTOR

VEHICLES

30. The review and synthesis of existing knowledge on ultrafine particles in

the air, with a specific focus on those originating from motor vehicles,

showed that vehicles are a significant source of ultrafine particles, and are

commonly the most significant source of air pollution in general in

populated urban areas. For this reason, it is critical to understand the

magnitude and characteristics of ultrafine particles in urban air generated

by motor vehicles, and hence this type of environment is the most likely to

be considered as a target for future air quality regulations in relation to

particle number.

31. No standard methods are currently available for conducting size classified

particle number measurement, and ultrafine particles are most commonly

measured in terms of their number concentrations. The review found that

the term “ultrafine particles” is often used imprecisely, and can be taken to

mean various ranges of particle number concentration in a subset of the

submicrometer range. Particle number concentrations reported in the

literature were found to depend on the instrument used and its setting.

456

The review showed that the mean and the median measurements by CPCs

are 32% and 56%, respectively, higher than those for DMPS/SMPS, and

while differences for specific environments could be expected to vary (eg.,

larger differences may be expected for environments where the nucleation

mode is present, and smaller differences where aged aerosol dominates),

this finding nevertheless shows the overall magnitude of difference that

can be expected when comparing results using these different measuring

techniques. The significance of this finding is that it highlights the

importance of keeping these differences in mind when attempting to

establish a quantitative understanding of variation in particle

concentrations reported by different studies. Secondly, the finding further

emphasizes the need for use of standardized measurement procedures to

enable meaningful comparison between results from different studies, and

this has particular significance for human exposure and epidemiological

studies.

32. Despite differences found in reporting measured concentration levels, the

review found that it was possible to quantify the differences between

background concentrations of ultrafine particles in clean environments,

with the levels in vehicle-influenced environments, and that the latter can

span over two orders of magnitude higher than levels in clean

environments. Clean background levels were found to be, on average, of

the order of 2.67 ± 1.79 x 103 cm-3 , while levels at urban sites are 4 times

higher and levels at street canyons, roadside, road and tunnel

457

sites are 27, 18, 16 and 64 times higher, respectively. This finding is of

profound significance in relation to human exposure assessment and

epidemiological studies, and emphasizes the importance of exposure

assessments being conducted in locations where exposures actually occur,

and at time scales that elucidate the temporal nature of the exposure.

These findings suggest that it is unlikely that epidemiological studies

would provide answers based only on monitoring in central locations, and

that central monitoring data alone underestimates exposure and may lead

to inappropriate management of public health risk.

33. The current lack of answers from epidemiological studies in relation to

ultrafine particles and exposure-response relationships is hampering the

development of health guidelines and national regulations. Recent World

Health Organization Air Quality Guidelines set for particulate matter in

relation to mass concentration related to annual mean values for PM2.5 and

PM10 of 10 and 20 µg m-3, respectively (WHO 2005), and these guidelines

are not substantially higher than the concentration levels encountered most

commonly in natural environments (while some locations, and under some

circumstances, concentrations in natural environments may be well below

or above those cited). While the lack of exposure-response relationship

for ultrafine particles makes it impossible to propose health guidelines for

ultrafine particles, it is important to point out that this review found in

vehicle-influenced environments ultrafine particle emissions were up to an

order of magnitude higher than in the natural environments. Hence, it is

suggested that in the absence of a threshold level in response to exposure

458

to ultrafine particles, that future control and management strategies should

target a decrease in ultrafine particles in urban environments of more than

one order of magnitude.

34. The review found large uncertainties currently exist in relation to vehicle

emission factors for particle number and different particle size ranges; no

emission inventories for ultrafine particles generated by motor vehicles are

available; and only very limited data is available on long term trends in

ultrafine particle concentrations in urban environments. The implications

of these findings are that in order to control and manage this major

pollution source, it is critical that significant research effort be expended

to fill this gap in our current knowledge.

35. Findings of the review suggest that although estimations of pollution

concentration in the air are commonly derived based on source emission

inventories, which in turn are derived using the source emission factors,

with respect to the process of secondary particle formation, estimation of

ultrafine particle concentration cannot be derived solely based on vehicle

emission factors, (as these are more likely to reflect emissions of primary

particles), but predictions for secondary particle formation in exhaust

plumes and particle formation by nucleation processes in the wider

atmosphere may need to be considered.

459

Secondary particle formation can result in a rapid increase in particle

number concentrations by one to two orders of magnitude to the

concentration levels of the order of 105 particles cm-3, and most of the new

particles are formed by ion-induced or binary nucleation of sulphuric acid

and water or by ternary nucleation involving a third molecule followed by

condensation of semi-volatile organics, with photochemistry playing an

important role in some of these processes. The mechanisms of new

particle formation strongly depend on local meteorological factors, and

hence models of the dynamics of particle formation in urban environments

need to include all factors involved and be location-specific.

Significant peaks in particle number concentration can occur due to

secondary particle formation, and if future regulations considered were

based on particle number, then the implications of these findings would be

that issues relating to whether the regulations should be set around the base

line concentrations without the peak concentrations, or whether they should

include the peaks and how peaks should be are defined would need to be

resolved. Hence, the significance of this finding is that a much better

understanding of particle formation dynamics in different environments

needs to be obtained, including in traffic-influenced environments, and this

understanding would greatly assist regulation formulation, if secondary

particle formation we found to be relevant and appropriate to include in

regulations.

460

36. The review found that there have been only a relatively small number

of studies which have focused on ultrafine particle chemistry, and

that large differences in particle chemical composition can relate to

factors such as particle solubility, volatility and elemental differences,

for example. Differences can depend on a number of factors,

including fuel, after-treatment devices and vehicle technology used,

and also on post-formation processes occurring during atmospheric

transport. Since particle composition may be a factor determining

particle toxicity, there exists a need to develop a more detailed

understanding on ultrafine particle chemistry in different

environments.

Table 8.1 follows, and provides a précis of principal findings referred to in

1-36 above and their significance in terms of application.

461

Table 8.1 Précis of the principal findings of this PhD research and their significance in terms of application RESEARCH COMPONENT AND PRINCIPAL FINDINGS

SIGNIFICANCE AND APPLICATION FOCUS

Modality within particle size distributions

1. A clear and distinct separation was found between the accumulation and coarse modes at 1 µm in

different worldwide environments for particle number, mass, volume and surface area metrics.

Introduction of a PM1 mass ambient air quality standard is likely to suit the majority of environments worldwide.

2. PM2.5 measurements in the coarse mode related mainly to mechanically-generated sources, but for some environments were from combustion sources, making it complex to distinguish between these sources in different environments. 3. PM1 measurements provided very good information about nucleation and accumulation modes and enabled a clear distinction to be made between combustion and mechanically-generated aerosols. 4. PM10 measurements were found to be suitable for discriminating mechanically-generated particles in the coarse mode. 5. PM1 and PM10 measurements enabled clearer distinction to be made between emissions from combustion and mechanically-generated sources than PM2.5 and PM10.

Provides evidence that modality is a useful basis for developing air quality regulations. A combination of PM1 and PM10 mass ambient air quality standards are likely to be more suitable than the current air quality standards of PM2.5 and PM10 for distinguishing between emissions from combustion and mechanically-generated sources, such as emitted from motor vehicles.

6. Most modes in traffic-influenced environments occurred in the submicrometre size range (diameters < 1 µm) and were dominant in the ultrafine size range (< 0.1 µm).

Particle number standards related to the submicrometre size range, and smaller size ranges such as ultrafine size, need to be introduced to control motor vehicle particle emissions.

462

RESEARCH COMPONENT AND PRINCIPAL FINDINGS


A new method for developing comprehensive particle inventories 7. A new method was developed for estimating comprehensive particle emission inventories for different Vehicle and Road Type combinations for different particle metrics which are suitable to use in transport modelling.

This method provides a comprehensive set of particle emission factors which cover the full size range of particles emitted, including particle number and mass.

8. The comprehensive set of average emission factors derived from the statistical models developed in this study explain:- 86% of the variation in published emission factors for particle number;

93% of the variation in published emission factors for particle volume;

87% of the variation in published emission factors for in PM1;

65% of the variation in published emission factors for in PM2.5;

47% of the variation in published emission factors for in PM10.

Very good correlations were found between the average emission factors derived in this study and those in the international literature. Hence confidence can be had in the average emission factors derived in this study. The lower correlation for PM10 may be confounded by the influence of resuspended road dust at this size range. More methods are needed to enable road dust to be discriminated from exhaust tailpipe emissions.

9. The explanatory variables in the statistical models developed to derive average emission factors were found to be:-

Particle number: Vehicle Type and Instrumentation Particle volume : Vehicle Type, Size Range Measured, Speed Limit on the Road PM1: Vehicle Type and Fuel Type, PM2.5 : Vehicle Type and Instrumentation PM10: Vehicle Type and Road Type

These explanatory variables highlight critical components of emission factor studies, which may need to be given special attention and consideration in study design, conduct and reporting

463



A new method for developing comprehensive particle inventories (continued) 9, 10. The most suitable average emission factors to use in transport modelling were found to be:-

Particle Number: Fleet, HDV, LDV – measured by the Condensation Particle Counter (CPC) Diesel Buses – measured by the Scanning Mobility Particle Sizer (SMPS)

Particle Volume: Fleet, HDV, LDV:- Speed Limit on the Road ≤ 60 km/hr and Size Range Measured 18-300nm Speed Limit on the Road > 60 km/hr and Size Range Measured 18-700nm

PM1: Fleet, HDV, LDV – based on Fuel Type

PM2.5 : Fleet – measured by the Tapered Element Oscillating Microbalances and Differential Mobility Analyzer HDV – overall average of all Instrumentation used LDV – measured by DustTrak

PM10: Fleet, HDV, LDV – freeway, highway, motorway, rural road, tunnel and urban Road Types Buses – boulevard and urban Road Types and dynamometer

These average emission factors are suitable for

developing inventories in any urban areas of the

developed world, particularly where there is no, or

insufficient data, available to derive emission factors.

They can be used to develop road-link based

inventories or region-wide inventories of the spatial

distribution of particle concentrations, to assess

current and predicted pollution levels and potential

health effects.

464



Statistically significant differences found between mean values of published emission factors

11, 12. Few statistically significant differences were found between mean values of published emission factors for Country of Study, Study Location (dynamometer, tunnel or vicinity

of the road studies) or different Road Types.

These statistical results reinforce the suitability and universal application of the comprehensive set of average emission factors derived in this study.

13. Insufficient data was available to conduct statistical tests related to average vehicle

speeds categories, as few studies reported average vehicle speeds at study sites. Most on-road studies only reported Speed Limit on the Road.

Studies need to quantify actual average vehicle speeds so that more realistic information can be obtained, rather than just reporting the Speed Limit on the Road.

14. Statistically significant differences were found between LDV and HDV mean emission

factors for all particle metrics.

Mitigation efforts need to focus on limiting population exposure to HDV emissions, and on reducing HDV emission rates.

15. Mean values for published emission factors for particle number for buses were restricted to SMPS measurements.

Bus studies using the CPC are needed to include the nucleation mode, where particle number tends to be very prolific, as SMPS does not usually measure this mode.

16. Statistically significant differences were found between mean emission factors for petrol and diesel-fuelled vehicles for PM10.

Highlights the significant differences between LDV petrol and HDV diesel emission rates (as 14 above).

17. Statistically significant differences were found between mean emission factors related to CPC and SMPS measurements (22.69 and 2.08 x 1014 particles per km respectively). No statistically significant differences were found in relation to Size Range Measured for particle number between the means of published emission factors for the lower and upper size ranges measured for the various levels of each of the categorical variables.

This difference needs to be addressed as a broader issue. CPC measure the nucleation mode and SMPS generally does not. Particle number in this mode is very prolific. Studies may be needed that focus on the < 18nm size range where particle number tends to be very high.

465



Other particle emission inventories

18. No comprehensive inventories of particle emissions covering the full size range of particles emitted from motor vehicles currently exist. The only study which attempted to develop such an inventory applied distribution profiles for smaller particles to PM10 estimate data, and did not base their emission factors

on measurement data for different particle size ranges.

Motor vehicle inventories need to quantify the full size range of particles emitted and include particle number and PM1, in addition to PM2.5 and PM10, because most particle emissions are found in the submicrometre and ultrafine size range.

Gaps in our knowledge re motor vehicle emission factor data

19. Limited emission factor data is available for particle volume, particle surface area, PM1, brake and tyre wear, road grade, engine power and for vehicles travelling at lower speeds, such as < 50 km/hr.

Few methods were found for discriminating resuspended road dust from motor vehicle tailpipe emissions.

Further research is needed to quantify emission factors for motor vehicles related to particle volume, particle surface area, PM1, brake and tyre wear, road grade, engine power and for vehicles travelling at lower speeds, such as < 50 km/hr. Methods are needed to enable discrimination of resuspended road dust from motor vehicle tailpipe emissions.

20. No relevant bus emission factors were identified in the international literature where the Instrumentation used to derive the emission factors had measured down to 3 nm, nor which quantified particle volume, PM1, particle surface area or derived emission factors for vehicles travelling under congested conditions, eg., < 50 km/hr.

More studies are needed to derive bus emission factors that measure the nucleation mode, including down to 3nm. More emission factor studies are needed for particle volume, PM1, particle surface area and for vehicles travelling under congested conditions, eg < 50 km/hr.

21. Few Drive Cycle tests focus on speeds which emulate congested driving conditions. More Drive Cycle tests are needed which derive emission

factors for vehicles travelling at lower speeds, eg., < 50 km/hr to accurately model congested driving conditions.

466



Development of a comprehensive particle emissions inventory for urban SEQ 22. The 2004 inventory revealed that 93.3% of VKT is related to LDVs, 6.3% to HDVs and

0.4% to buses.

Strategies are needed to reduce LDV VKT and increase vehicle occupancy rates, which were 1.5 passengers in 2004, in order to reduce particle emission levels.

23. HDVs in urban South-East Queensland although they contributed only around 6% of total regional VKT, contributed more than 50% of particle number and PM1 emissions. HDV average emission factors were 20 times higher than LDVs in the submicrometre size range.

Greater effort is needed to focus on reducing population exposure to HDV particle emissions, and to introducing technologies that will reduce emission rates, particularly in the submicrometre size range.

24. The urban South-East Queensland fleet in 2004 was found to emit: 1.08 (0.54-1.97) x 1025 particle number (HDVs contributed 54%) 477 (233-964) tonne of PM1 (HDVs contributed 55%) 736 (225-1436) tonne of PM2.5 (LDVs contributed 61%) 2614 (9668a) tonne PM10 (LDVs contributed 81%) a only the upper 95% confidence level value is available for this particle metric

Particle number inventories for motor vehicle fleets are needed in addition to particle mass inventories.Increasing average vehicle occupancy rates and shifting trips from private car travel to buses have the potential to reduce particle emissions. In 2004 the average vehicle occupancy rate for LDVs was 1.5 passengers and buses 15.5 passengers per vehicle, indicating ample opportunity for increases to occur.

25. In the 2004 urban South-East Queensland inventory:-

total LDV emissions dominated PM2.5 and PM10 emissions on roads with speed limits of < 80 km/hr (urban roads);

and total HDVs dominated particle number and PM1 on higher speed roads with speed limits of ≥ 80 km/hr (urban-major roads).

Greater efforts are needed to reduce on-road LDV and HDV VKT. These results were influenced by the fact that LDV VKT on urban roads was almost double HDV VKT; and the HDV emission factor for urban-major roads was almost 6 times that for LDVs and total HDV VKT was slightly higher than total LDV VKT on this Road Type.

467



Validation of the developed inventory 26. The only inventory able to be compared to the particle number and particle mass inventories for urban South-East Queensland developed in this study related to a PM10 invnetory developed by the Queensland Environmental Protection Agency (QEPA) for SEQ, which was 2249 tonne per annum for 2000. The QEPA estimate compares well with the urban South-East Queensland quantification of 2614 tonne for PM10 found in this study.

The inventory quantified for PM10 for urban South-East Queensland derived in this study compares well with the Queensland Environmental Protection Agency PM10 inventory, and therefore confidence can be had in the total inventory for urban South-East Queensland developed in this study.

Scenario modelling using the urban South-East Queensland inventory data – findings and likely future emissions 27. Each 10% reduction in LDV VKT, where 70-100% of these passengers moved to new buses, was associated with reductions of 3-4% for particle number, 1-2% for PM2.5 and 1-6% for PM10, respectively.

Each 10% reduction in Work Trip LDV VKT (trips to and from work from home) where 50% of these passengers moved to new buses, was associated with reductions of 2% for particle number; 2-3% for PM2.5 and 3-4% for PM10 respectively.

Average vehicle occupancy rates for LDVs and buses in urban South-East Queensland could be increased to reduce particle emission levels, which were 1.5 passengers and 15.5 passengers respectively in 2004. Bus occupancy rates of around 35-45 passengers are possible with the SEQ bus fleet.

28. Emissions per passenger per km travelled were:-

LDVs emissions were 1-2 orders of magnitude higher than buses for particle number; PM2.5 emissions were similar for LDV and Diesel buses; LDV emissions for PM10 were almost 5 times those for Diesel buses and several orders of

magnitude higher than Compressed Natural Gas buses.

Even greater reductions in particle emissions rates could be achieved by increasing bus occupancy rates (Refer 28 above).

29. Particle emissions predicted for urban SEQ in 2026 indicate an 100-fold increase in

particle number; and reductions of 38% for PM1, 36% for PM2.5 and 31% for PM10.

More action is needed to reduce HDV particle number and submicrometre emissions, and limit population exposure to HDVs and identify environmentally-sustainable lower polluting options for moving freight.

468



Review and synthesis of current knowledge on ultrafine particles, with a specific focus on motor vehicles 30. Motor vehicles were found to be a significant source of ultrafine particles and of air pollution generally in urban populated areas.

It is likely that ultrafine particles need to be the target for future air quality regulation in relation to particle number in urban populated areas.

31. No standardized methods are available for measuring particle number and those currently used show differences in measurement outputs. Particle number concentrations reported were found to depend on the instrument used and its setting, and the term “ultrafine particles” was often used imprecisely, and taken to mean various ranges of particle number concentration.

Differences in outputs of measurement techniques used currently for measuring ultrafine particles need to be borne in mind when interpreting variations in particle concentrations reported by different studies. Standardized measurement procedures are needed, and these have particular significance for epidemiological studies and human exposure studies.

32. The extent of difference found between background concentrations of ultrafine particles in

clean environments as compared to those in vehicle-influenced environments spanned over two orders of magnitude. In vehicle-influenced environments, this difference was found to be dependent on the type of road environment and where measurements were conducted, eg., in a street canyon, roadside, on a road or in a tunnel.

This difference is of profound significance for epidemiological and human exposure studies, and highlights the importance of assessing exposure where exposure occurs, and at appropriate time scales, which means it is likely that epidemiological studies providing answers based solely on monitoring data obtained in central locations (such as central monitoring data) may lead to underestimates of exposure, and inappropriate management of public health risk.

469



Review and synthesis of current knowledge on ultrafine particles, with a specific focus on motor vehicles 33. A lack of epidemiological studies on ultrafine particles in terms of exposure-response

relationships is hindering development of health guidelines and ultrafine particle regulation, and the extent of difference found between natural environments and vehicle-influenced environments was up to an order of magnitude.

This finding as to the order of magnitude difference between natural and vehicle-influenced environments in terms of ultrafine particles may form the basis of an important target for future management and control strategies which could seek to reduce ultrafine particle emissions by this amount. A target in reduction of ultrafine particles of more than one order of magnitude in populated urban areas could be suitable in the absence of a threshold level in response to exposure to ultrafine particles being available.

34. Large uncertainties were found in relation to vehicle emission factors for particle

number and different particle size ranges, and an inventory of ultrafine particles emitted from motor vehicles is not currently available. Few data is available on the long term monitoring of ultrafine particles in urban environments.

Further research is needed in relation to vehicle emission factors for particle number and different particle size ranges, and in development of an ultrafine particle inventory for a motor vehicle fleet. More studies are needed on the long term trend monitoring of ultrafine particles. Knowledge in these areas is critical for the control and management of ultrafine particles.

35. Secondary particle formation was found to affect levels of ultrafine particle

concentrations, and these formations may need to be considered if future particle number regulations are formulated.

More research is needed relating to the mechanisms and dynamics of secondary particle formation, as well as the effects of local meteorology, and the relevance of these formations if formulating air quality regulation for particle number and in developing vehicle emission inventories of ultrafine particles.

36. Particle chemical composition needs to be considered when characterizing ultrafine particles, and few studies were found that focused on ultrafine particle chemistry.

Further research is needed on ultrafine particle composition and chemistry to inform air quality regulation and epidemiological and health impact assessments, and for determining particle toxicity.

470

8.4. GENERAL CONCLUSIONS FROM THIS STUDY

The general conclusions from this PhD study related to:-

8.4.1. Modality in ambient particle size distributions

Examination of the modes in particle size distributions was found to have

potential as a basis for developing air quality standards and guidelines, as modes

provided extremely valuable information about contributions from different

pollution sources and particle mechanisms.

The research found that a clear and distinct separation occurred at around 1 µm in

600 modal location values examined for particle number, surface area, volume

and mass size distributions in a wide range of environments worldwide; and a

similar separation was also found in all urban South-East Queensland

environments examined between the accumulation and coarse modes for particle

volume and number size distributions at around 1µm.

8.4.2. A new mass air quality standard for PM1, and its combination with

PM10

Based on the results of the urban South-East Queensland study and the other

studies conducted around the world, it is concluded that PM1 and PM10 offer

greater potential as a combination of particle mass standards than the current mass

standards of PM2.5 and PM10 for combustion sources, such as motor vehicles, for

discriminating between combustion and mechanically-generated particles.

Although presently few data exist on PM1 concentrations, measurement

technologies are available which are similar to those used for PM2.5 monitoring.

471

The analysis revealed that:-

(i) PM10 measurements provided information almost entirely on particles

generated from mechanical processes and belonging to the coarse mode

(eg., mechanical wear and tear or tyres, particles resuspended by vehicle

traffic);

(ii) PM2.5 measurements (coarse mode) provided information mainly on

mechanically-generated particles, but for some environments

contributions from combustion process modes (nucleation and

accumulation modes) was significant. This finding led to the conclusion

that interpreting PM2.5 data and distinguishing contributions from

different sources could become complex, and suggests that PM2.5, as a

basis for a standard, may be inadequate for controlling particle emissions

and concentrations; and

(iii) PM1 measurements (nucleation and accumulation modes) provided very

good information about combustion-generated processes and enabled a

clear discrimination to be made between combustion and mechanically-

generated aerosols.

472

More discussion and research is needed on the best combination of particle

number and mass concentration ambient air quality standards to control motor

vehicle emissions, including consideration of standards for submicrometre and

smaller particle size ranges, such as ultrafine particles; and substantial progress is

currently being made in development of monitoring technologies suitable to

measure particle number concentration.

8.4.3. A comprehensive set of particle emission factors for motor vehicles

The research was successful in deriving a set of average particle emission factors

for different Vehicle and Road Type combinations which can be used in transport

modelling and air quality assessments to model urban fleet emissions in the

developed world. These emission factors enable estimation of comprehensive,

size-resolved inventories covering the full size range of particles emitted by motor

vehicles, including particle number and particle mass.

The emission factors were derived from statistical models developed from a

review of more than 600 particle emission factors in the international published

literature and explain 86%, 93%, 87%, 65% and 47% of the variation in published

emission factor values for particle number, particle volume, PM1, PM2.5 and PM10

respectively. These emission factors are particularly suitable to use in regions

which do not have measurement data, or funding to undertake measurements, or

where experimental data is of insufficient scope; and can be used to develop road

link-based inventories and quantification of the spatial distribution of particle

concentrations in urban regions.

473

Inventories provide useful benchmark information on current pollution levels and

can inform the development of air quality guidelines and regulations to control

particle emissions, and modelling of transport planning scenarios which evaluate

the air quality implications of future transport and land use strategies, and

initiatives such as the introduction of new vehicle standards, new fuels and

technologies.

The comprehensive set of particle emission factors derived in this study provide

land use and transport planners with a straightforward method for evaluating the

air quality implications of a wide range of urban forms - from high density living

areas to areas with increased urban sprawl - and for evaluating the impact of new

busway and transport infrastructure developments. Without such a method and set

of emission factors their ability to minimize the important and potentially severe

health effects of particulate matter exposure in their planning and decision-making

would be severely limited.

In addition, the very high values of average emission factors derived for HDVs

emphasize that HDV movement requires very careful consideration in land use

and transport planning, and special attention in the setting of air quality standards

and regulations. HDV movement on roads situated close to populations requires

careful monitoring to minimize exposure to these gross emitters.

474

8.4.4. The first published comprehensive particle emissions inventory

for a motor vehicle fleet

This study presents the first published comprehensive inventory of motor vehicle

particle emissions covering the full size range of particles emitted by motor

vehicles and including particle number and particle mass. Conclusions from its

development and from analyses which modelled the air quality implications of

different future passenger and freight vehicle scenarios are as follows:-

• It was found that HDVs contributed only around 6% of regional VKT, but

contributed more than 50% of particle number and PM1 emissions which

means that urgent action is needed to reduce HDV diesel vehicle

emissions. Similar findings would be expected in other areas with high

HDV diesel VKT and, as urban South-East Queensland is not highly

industrialized region and is more service and tourism oriented, this means

that even larger HDV particle emissions could be expected in areas with

higher levels of industrialization. HDV particle emissions are a global

problem which requires reduction strategies such as mandatory fitting of

particle filters, regular emissions testing, and identification of freight

options and freight routes that produce lower emissions per tonne-

kilometre and result in lower exposures for populations in close proximity

to truck routes.

475

• This study demonstrated the value of examining and modelling changes in

average vehicle occupancy rates of LDVs and buses, which can be useful to

identify the extent to which small changes in travel mode choice and

occupancy rates can achieve reductions in particle emission levels.

• It was found that modelling future scenarios, such as modelled for 2026 for

urban South-East Queensland, which predicted an 100-fold increase in particle

number and 31-36% reduction in particle mass, offer opportunities to design

mitigation efforts tailored to expected changes in travel demand and vehicle

technologies.

• To adequately control particle emissions emitted by motor vehicles, guidelines

and standards need to be introduced for both particle number and PM1 to

complement existing mass-based standards.

• Another conclusion of this study is that developing regular particle inventories

will help identify hotspots, roads, busways and tunnels that pose a significant

health risk to exposed populations. Inventories are also needed to evaluate the

particle emission implications of new transport infrastructure developments

both pre and post-construction and initiatives such as high density living and

transit oriented developments.

476

• Urban congestion is a global problem which affects travel time and has

environmental implications. It contributes to pollution levels and climate

change effects. Specific speed-related emission factors need to be derived

to enable the modelling of traffic congestion and vehicles travelling at

lower speeds in urban areas.

• It is considered important that the work in this inventory be extended to

quantify the spatial distribution of particle concentrations, and that an

understanding be gained of the socioeconomic characteristics of

populations affected by hot-spots and their corresponding land use

classifications.

• This study also provides new knowledge that can be used in climate

models to assess the impact of motor vehicle particle emissions on the

global airshed, including particle concentrations reaching into the

troposphere and stratosphere, and their potential contribution to effects

such as the cooling and dimming of the planet, and climate change effects.

477

8.4.5. Synthesis of current knowledge on ultrafine particles in

relation to motor vehicles

The review and synthesis found that ultrafine particles are likely to be the target

for development of future air quality standards for particle number in relation to

motor vehicles in populated urban areas, and that a better understanding of

ultrafine particles is needed in terms of their temporal and spatial distribution,

long term trend monitoring and particle composition and chemistry. In addition,

it found that a motor vehicle emissions inventory is not currently available for

particle number or ultrafine particles in the published literature.

Standardized methods need to be developed for measuring particle number, and

discrepancies were found in terms of the outcomes of measurement techniques

currently used to measure ultrafine particles. This has implications for reviewing

variations in particle concentrations reported in different studies, and for

epidemiological studies and human exposure assessments.

The work provides direction for future management and control of ultrafine

particles and suggests a general target for reducing ultrafine particles generated by

motor vehicles in populated urban areas. It highlights the importance of

investigating the relevance of secondary particle formations in the formulation of

future particle number air quality regulation.

478

8.5. SCIENTIFIC CHALLENGES AND THE NOVEL

CONTRIBUTION OF THIS PHD STUDY

It is very important to quantify the extent of particle emissions emitted from

motor vehicle fleets because they are a major source of pollution and there are

known adverse health effects associated with exposure. To control this ever-

growing pollution source air quality regulations and development of regular

inventories are needed to quantify the extent and distribution of this pollution.

This PhD study contributes new knowledge and understanding to the field by:-

• Firstly, developing a new method for deriving inventories of motor

vehicle particle emissions covering the full size range of particles

emitted, and including particle number and particle mass emissions. This

method involves combining knowledge from two different disciplines -

from aerosol science and transport modelling, which is a novel approach

because it has never been attempted before.

• Secondly, devising new concepts for identifying suitable emission

factors to use in developing inventories. These included analysing a very

large set of emission factor data from Australia and overseas and

developing statistical models to derive average emission factors for

different vehicle and road type combinations for different particle sizes

and metrics, which have never been developed before.

479

• Thirdly, designing a new approach and new concepts for examining

modality within particle size distributions that is novel and has

never been done before, which deepened our understanding about

the fraction of mass contributed to different particle size ranges by

different sources and identified a new particle mass standard, PM1,

which was found to be suitable for the majority of worldwide

environments.

• Fourthly, providing evidence that a combination of PM1 and PM10

standards provide a more suitable combination of ambient particle

mass emission standards for discriminating between mechanical

and combustion-generated sources, such as emitted by motor

vehicles, than the present mass standards of PM2.5 and PM10.

• Fifthly, this study presents the first published comprehensive

inventory for a motor vehicle fleet which covers the entire size

range of particles emitted and includes particle number and size

fractions of particle mass for different vehicle and road type

combinations.

• Sixthly, developing new approaches for modelling future scenarios of

travel demand and their particle emission implications.

480

• Seventhly, current knowledge on ultrafine particles as it relates to

motor vehicles has been extensively reviewed and synthesized, and

this analysis provides clear guidelines on future directions needed in

research to ensure accurate epidemiological studies and human

exposure assessment and quantification of particle number and

ultrafine particles in populated urban areas.

8.6 COMPARISON OF EMISSION FACTORS DERIVED IN THIS PHD

STUDY WITH A SELECTION OF CANADIAN, EUROPEAN, UK

AND USA EMISSION FACTORS

Table 8.2 compares a selection of emission factors recommended for use in the

Australian National Pollutant Inventory (NPI) and from Australian Diesel

NEPM preparatory work, Canadian, European, UK, and USA studies, to those

derived in this PhD study. Particle volume and fleet emission factors were not

included in this comparison as they are not considered relevant for developing

inventories. At the time of this study, the majority of LDVs were petrol-fuelled

and HDVs diesel-fuelled.

When reviewing the analysis presented in Table 8.2, it should be borne in mind

that the emission factors derived in this PhD study using advanced statistical

analysis were found to explain 86%, 87%, 65% and 47% of the variation in

published emission factors for particle number, PM1, PM2.5 and PM10

respectively.

481

Table 8.2 Comparison of Australian National Pollutant Inventory (NPI), Australian Diesel NEPM Preparatory Work, and a selection of Canadian, European, UK and USA particle emission factors, with emission factors derived in this PhD study Particle metric

Researchers Method HDV emission factor (diesel-

fuelled)

LDV (petrol-fuelled)

Bus (Diesel-fuelled) b

emission factor emission factor emission factor Particle number


This study Keogh et al. 2009 Derived using advanced statistical analysis of measurement data sourced from an extensive worldwide literature review

65 (60.19-69.81) 3.63 (a-9.85) 3.08 (a-9.30)

Australia Jamriska et al. 2004 Tunnel measurements n/a n/a 2.27, 3.11

Australia Morawska et al. 2001

Dynamometer measurements 5.9 n/a n/a

Australia Ristovski et al. 2002

Dynamometer measurements n/a n/a 3.87

Australia Morawska et al. 2005

On-road measurements 1.53 to 7.17 2.18 to 6.08 n/a

Austria Imhof et al. 2005a

Tunnel measurements 3.94 0.59 n/a

European drive cycles

CONCAWE 1998 Dynamometer measurements n/a 0.362 to 1.59 n/a

Sweden Gidhagen et al. 2003

Tunnel measurements 73.3 10.1 n/a

482

Particle metric

Researchers

Method

HDV emission factor (diesel-

fuelled)

LDV (petrol-

fuelled)

Bus (Diesel-

fuelled) b

emission factor

emission factor emission factor

Particle number (c’td)


Germany Imhof et al. 2005c On-road measurements 7.79 1.22 n/a

Sweden Gidhagen et al. 2004 On-road measurements 52 1.4 n/a

Switzerland Imhof et al. 2005a On-road measurements 55, 73 3.2, 6.9 n/a

UK Jones and Harrison 2006 On-road measurements 6.36 7.05 n/a

UK Imhof et al. 2005a Tunnel measurements 6.84 0.59 n/a

USA Cadle et al. 2001 Dynamometer measurements n/a 0.04 to 2.36 n/a

PM1 mg per km This study Keogh et al. 2009 Derived using advanced statistical analysis of

measurement data sourced from an extensive worldwide literature review

287 (257-317)

16 (a-50)

n/a

Australia Department of Environment & Heritage (DOEH 2003)

Dynamometer, composite urban drive cycle 257 to 364

n/a n/a

Switzerland Imhof et al. 2005b On-road measurements 187 to 413 4, 29 n/a

483

Particle metric


fuelled)



emission factor emission factor emission factor PM2.5

mg per km

This study Keogh et al. 2009 Derived using advanced statistical analysis of measurement data sourced from an extensive worldwide literature review

302 (236-367) 33 (a-80) 299b (205-394)

Australia NEPC (2000) - Diesel preparatory project

Dynamometer testing on composite urban drive cycle

397 to 672 n/a 695, 1068

Australia National Pollutant Inventory (NPI 2008)

Calculated using a US PM profile from the California Emission Inventory and Reporting System (CEIDARS 2000)

n/a

7.45, 9.11

n/a

Australia

Jamriska et al. 2004

Tunnel measurements

n/a

n/a

201-583

Australia

Tran et al. 2003 Tunnel measurements d

526

7

n/a

UK Imhof et al. 2005a Tunnel measurements

381 19 n/a

USA

Abu-Allaban et al. 2003

On-road measurements on different Road Types

57 to 480

7 to 90

105 to 260

USA Gertler et al. 2002 Tunnel measurements

135 14 n/a

USA

Wayne et al. 2004

Dynamometer measurements

n/a

n/a

42-124

484

Particle metric


fuelled)



emission factor emission factor emission factor

PM10

mg per km

PM10 for different Road Types

Australia National Pollutant Inventory (NPI 2008)

Based on UK 2000 g/L data from UK Emissions Inventory Database (NAEI 2007)

n/a

9.82 n/a

Australia National Pollutant

Inventory (NPI 2008) Sourced from UK Emissions Inventory Database (NAEI 2007), based on g/L, petrol density 739 kg/kL.

n/a

8.03 n/a

This study

Keogh et al. 2009

Boulevard c

4815 (3459-6171)

454 (a-1413)

4130 (2774-5486)

USA Abu-Allaban et al. 2003 Boulevard c 530 to 9100 90 to 850 460, 650, 7800 This study

Keogh et al. 2009

Freeway

2500 (1144-3856)

285 (a-1244)

n/a

USA Abu-Allaban et al. 2003 Freeway 1300, 3700 100, 260, 680 n/a

This study

Keogh et al. 2009

Highway

840 (a-1947)

141 (a-924)

n/a

Australia NSW EPA 2003 cited in Hibberd 2005 Highway (arterial road)

186-311

17 n/a

Switzerland Gehrig et al. 2004 Highway 344 33

n/a

Switzerland Imhof et al. 2005b Highway 275 26

n/a USA

Abu-Allaban et al. 2003

Highway

1900, 9100 90, 215 n/a

485

Particle metric


fuelled)



emission factor emission factor emission factor PM10 (c’td) mg per km PM10 for different Road Types

This study Keogh et al. 2009 Motorway 213 (a-1568) 63 (a-1419) n/a

Switzerland Imhof et al. 2005b Motorway 158 63 n/a

Switzerland Gehrig et al. 2004 Motorway 267 63 n/a This study

Keogh et al. 2009

Rural Area

394 (a-2312)

46 (a-1964) n/a

Switzerland Gehrig et al. 2004 Rural area 394 46 n/a This study

Keogh et al. 2009

Tunnel

1019 (236-1802)

14 (a-797) n/a

Australia Tran et al. 2003 Tunnel measurements d 615 9 n/a

Australia Hibberd 2005 Tunnel measurements 2000 5, 10, 22 n/a

Austria Imhof et al. 2005a Tunnel measurements 944 n/a n/a

Austria Schmid et al. 2001 Tunnel measurements 394 30 n/a

USA Gertler et al. 2002 Tunnel measurements 181 10 n/a

486

Particle metric


fuelled)



emission factor emission factor emission factor

PM10 (c’td)

mg per km

PM10 for different Road Types

This study

Keogh et al. 2009

Urban e

538 (a-1145)

156 (a-635)

1089 (306-1872)

Australia NEPC (2000) - Diesel preparatory project

Urban - dynamometer composite urban drive cycle classed as Urban Road Type e n/a n/a 698, 1100

Canada

Environment Canada cited in Lowell et al. 2003

Urban - dynamometer urban drive cycles, classed as Urban Road Type e n/a n/a

149, 230, 863, 1224, 3416,

Switzerland Gehrig et al. 2004 Urban - on-road measurements 496, 703, 1268 30, 49, 104 n/a

Switzerland Imhof et al. 2005b Urban - on-road measurements 652 51 n/a

USA Abu-Allaban et al. 2003 Urban - on-road measurements c 750 100, 180 650

USA Cadle et al 1997 Urban - dynamometer urban drive cycles classed as Urban Road Type e n/a 3.9 n/a

USA Cadle et al 2001 Urban - dynamometer urban drive cycles classed as Urban Road Type e n/a 32 n/a

This study

Keogh et al. 2009

Dynamometer

n/a

n/a

313 (a-753)

UK

Romilly 1999

Dynamometer measurements

n/a

n/a

347 to 668

487

a The lower bound 95% confidence interval value calculated to be negative and therefore is not valid. These values, although physically uninterpretable, can be obtained as a

consequence of the normal assumptions underlying the models, and hence are not reported. b Buses – principally diesel-fuelled buses. Some studies did not specify fuel used, and

due to the timing and location of studies, these can be assumed to be diesel-fuelled. c These on-road Boulevard and Urban Road Type studies were reported to be affected by very

high levels of resuspended road dust and the influence of variation in acceleration and speed (Abu-Allaban et al. 2003). d In-stack pollution monitoring data and hourly vehicle

counts. e Due to the small sample size of emission factors derived from measuremens on urban roads, dynamometer measurements of urban and Central Business District (CBD)

drive cycles were classed as Urban Road Type. (NAEI) UK National Atmospheric Emissions Inventory “Emissions Factor Database” retrieved August 2007 from the NAEI

website (www.naei.org.uk/emissions/index.php) (CEIDARS) California Emission Inventory and Reporting System (2002), Particulate Matter (PM) Speciation Profiles,

26/09/2002.

488

The particle emission factors presented in Table 8.2 show that when

comparing Australian emission factors to those derived in this study:-

• Australian emission factors for LDVs for PM2.5 are at least an order of

magnitude lower, and for PM10 are several orders of magnitude lower

than those derived in this study;

• for HDVs the Australian emission factors for particle number and

PM10 highway are substantially lower, and for PM10 tunnel

measurements are both substantially higher and lower than those

derived in this study;

• in terms of diesel-fuelled vehicles for PM2.5 bus and HDVs,

Australian emission factors were substantially higher than the

values in this study.

This analysis suggests that Australian emission factors for LDV petrol-fuelled

vehicles may be considerably underestimated. LDV PM2.5 and PM10 emission

factors are likely to have been influenced by study site conditions, where vehicle

fleet speeds may not have been excessively high (for example ≤ 60 km/hr) or

where measurements were undertaken in tunnels, leading to the incidence of

lower levels of resuspended road dust at these particle size ranges. Methods used

for deriving emission factors may have also influenced results. The Australian

NPI emission factors for LDVs for PM2.5 were calculated using a US PM profile

from the California Emission Inventory and Reporting System (CEIDARS 2000)

and the LDV emission factors for PM10 were based on UK g/L and petrol density

data from the UK Emissions Inventory Database (NAEI 2007), and the resultant

489

Australian NPI emission factors are up to one order of magnitude lower than most

other worldwide emission factor values derived for LDVs cited in Table 8.2. This

highlights a major discrepancy in the values presented for LDVs in the Australian

NPI emission inventory manuals.

For PM2.5 diesel-fuelled HDVs and buses, differences may relate to the nature of

the composite driving cycle used for dynamometer testing for the Australian

emission factors, which included six different CUEDC cycles developed to

represent driving in a range of urban traffic conditions, from congested to

highway/freeway (NEPC 2000).

In the case of particle number emission factors, these lower value emission factors

may be due to the choice of instrumentation used and the size range measured,

particularly in cases where the lower size range of the nucleation mode has not

been measured (< 10 nm).

The particle emission factors presented in Table 8.2 demonstrate that when

comparing Canadian, European, UK and USA emission factors with those derived

in this study for:-

• Particle number: Considerable variation was found. HDV particle

emission factors derived in Austria, Germany and the UK were an

order of magnitude lower than those derived in this study. LDV

emission factors from Austria were an order of magnitude lower;

European drive cycle, German and USA studies were lower; and

UK emission factors were both higher and lower than this study.

490

• PM1: Swiss emission factors for LDV and HDV were variable and were

both higher and lower than those derived in this study.

• PM2.5 : USA emission factors for LDVs and HDVs tended to vary and

were both higher and lower than this study, and UK LDV emission factors

were lower.

• PM10: for Boulevard and Freeway Road Types for LDV and HDV USA

emission factors were both higher and lower than this study; Highway

LDV and HDV Swiss emission factors were considerably lower; whereas

USA emission factors were considerably higher. Motorway – Swiss HDV

were slightly lower and slightly higher. Tunnel – Austrian and USA

emission factors were lower and for Urban Road Type Bus – Canadian

emission factors varied, as did the Swiss HDV emission factor. For LDV

on Urban Roads the Swiss emission factor was lower, but for the USA for

LDV it was both higher and lower.

The analysis above confirms the diversity in terms of the range of emission factor

values reported in the international published literature, and the complex nature of

deciding which emission factors are the most suitable to use in developing motor

vehicle emission inventories.

In terms of particle number, instrumentation choice and size range measured

strongly influence the derived emission factor value, particularly if the lower size

range of the nucleation mode has not been measured. In the case of particle mass,

a very wide range of different instrumentation and methods were used in these

studies, which contribute to the range of different results. The lack of accurate

491

methods available, at this time, for discriminating motor vehicle tailpipe particle

emissions from resuspended road dust, particularly at the PM2.5 and PM10 size

range, further add to the complexity of interpreting on-road measurements,

particularly under high-speed scenarios where high levels of road dust may be

resuspended.

8.7. FUTURE RESEARCH FOCUS

A number of recommendations are made for future studies which could be based

on the data, knowledge and methods developed in this PhD research project, and

on gaps in existing knowledge identified in this study in terms of ultrafine

particles as they relate to motor vehicles. These studies could relate to

quantifying particulate matter pollution; developing air quality monitoring tools,

guidelines, standards and policies; undertaking scenario modelling of particle

emissions; and freight and fleet-related studies; and the conduct of further

research related to ultrafine particles. The recommended studies are outlined in

Table 8.2 and also discussed below.

492

Table 8.3 Future studies recommended which use the data, knowledge and methods developed in this PhD study

Data, knowledge and methods developed in this PhD study

No

Focus of the recommended study

Particle emission factors

Inventory method

Method for examining modes

Scenario modelling variables

Scale of the project

1 Develop a transport emission database for climate models to quantify motor vehicle particle contributions to upper atmospheres (eg., troposphere and stratosphere).

Global

2 Develop an inventory of vehicle emissions in relation to particulate matter, with a specific focus on nano and ultrafine particles (diameters < 0.05 µm and < 0.1 µm respectively). (These small particle size ranges are currently at the centre of air quality, health impact and climate change debates).

Regional & Local

3 Determine the spatial distribution of particle concentrations for different particle sizes and metrics, and overlay these quantifications on GIS maps to identify emission hotspots, and the socioeconomic profiles of populations affected, particularly at distances of 100, 200 and 300m from roads. Compare these quantifications with current air quality guidelines and standards. This research would link transport emission and pollution dispersion modelling and its outcome would include visualization of the results, which would be a useful tool in the management of transport emission decision-making.

Regional & Local

493

No




Inventory method




4 Design an air quality alert system which automatically calculates, on a daily basis, the particle contributions from LDVs and HDVs on selected roads using traffic data.

Vehicle-based

5 Predictive modelling of particle and gaseous emissions for different scenarios of pre and post-construction of busways, tunnels and major transport infrastructure to determine their air quality implications. This would include research into the components of the transport system which are the most sensitive in terms of effecting changes in regional particle emission levels (eg., vehicle occupancy rates).

Local airshed

Road-link based

Vehicle-based

6 Quantify HDV VKT and average freight loads to determine particle emissions per tonne of freight carried per km, and identify other low polluting options for transport (eg., rail, intermodal options, B-triples etc)

HDV Vehicle-based

7

Research into, and develop recommendations for the most appropriate air quality guidelines and standards, in terms of both particle number and particle mass, to control motor vehicle particle emissions, with a special focus on the smallest particles, in the nano and ultrafine size ranges.

Vehicle-based

494

No




Inventory method




8

Ultrafine particles are a likely target for future air quality regulation in terms of particle number in populated urban areas.

Local airshed

Road-link based

Vehicle-based

9

Standardized methods need to be devised to measure particle number, and these have particular significance for epidemiological and health exposure assessments.

Local airshed

Road-link based

Vehicle-based

10

Develop a target to reduce ultrafine particle concentrations generated by motor vehicles in urban areas by more than one order of magnitude.

Local airshed

Road-link based

Vehicle-based

495

No




Inventory method




11

More studies are needed related to the mechanisms and dynamics of secondary particle formation of ultrafine particles generated by motor vehicles, as these have particular relevance for formulating air quality regulation, epidemiological and human exposure assessments. Include consideration of the effects of localized meteorological conditions.

Local airshed

Road-link based

Vehicle-based

Particle physics

12

More studies are needed on the particle composition and chemistry of ultrafine particles, to better understand these in terms of different environments.

Local airshed

Road-link based

Vehicle-based

Particle physics & chemistry

496

National and state policies and programs which could be considered for

development include:-

• Implementing the requirement for development of regular particle number

emission inventories for motor vehicles; and investigating the extent of

gross emitters in motor vehicle fleets.

• Evaluating the feasibility of introducing low emission zones to limit

HDVs not fitted with particulate filters or aftertreatment devices which

have high emission rates.

• Developing a communication strategy which raises the awareness of the

public, policymakers and health professionals of the implications of

individual transport choices, eg., how much particulate matter is emitted

per passenger per km for different transport modes.

• Developing national policies and regulations which specifically focus on

reducing HDV emissions, particularly with respect to particle number and

PM1 emissions, eg., providing financial incentives, mandatory fitting of

particulate filters etc.

497

8.8. REFERENCES

ABS., 2004. Population by Age and Sex. Australian Bureau of Statistics, Canberra.

ABS., 2004. Survey of Motor Vehicle Use Australia. Australian Bureau of

Statistics, Canberra.

Abu-Allaban, M., Gillies, J.A., Gertler, A.W., 2003. Application of a multi-lag



AQEG., 2005. Particulate Matter in the UK. Department for Environment, Food

and Rural Affairs, London.

Cadle, S.H., Mulawa, P., Groblicki, P., Laroo, C., Ragazzi, R. A., Nelson, K.,



35(1), 26-32.

Cadle, S.H., Mulawa, P.A., Ball, J., Donase, C., Weibel, A., Sagebiel, J. C.,



& Technology 31(12), 3405-3412.

498



european driving cycle conditions.

CONCAWE, Brussels Report no. 98/51. Diesel Vehicles in Australia, Department

of the Environment and Heritage, Canberra.



Dora, C., 1999. A different route to health: implications of transport policies

British Medical Journal 318(7199), 1686-1689.

EPA., 2004. Air Emissions Inventory South-east Queensland Region. Brisbane,

Environmental Protection Agency.





Gertler, A.W., Gillies, J.A., Pierson, W.R., Rogers, C.F., Sagebiel, J. C., Abu-




499


Hansson, H.C., 2003. Model simulation of ultrafine particles inside a road tunnel.


Gidhagen, L., Johansson, C., Omstedt, G., Langner, J., Olivares, G., 2004. Model

simulations of NOx and ultrafine particles close to a Swedish highway.


Goodwin, J.W.L., Salway, A.G., Murrells, T. P., Dore, C. J., Passant, N.R.,











500


Rodler, J., Sturm, P., McCrae, I., Sjodin, A., Baltersperger, U., 2005a. Aerosol

and NOx Emission Factors and Submicron Particle Number Size Distributions in

Two Road Tunnels with Different Traffic Regimes. Atmospheric Chemistry and

Physics Discussions 55127-55166.


Baltensperger, U., 2005b. Real-world emission factors of fine and ultrafine

aerosol particles for different traffic situations in Switzerland. Environmental


Imhof, D., Weingartner, E., Vogt, U., Drei seidler, A., Rosenbohm, E., Scheer, V.,

Vogt, R., Nielsen, O.J., Kurtenbach, R., Corsmeier, U., Kohler, M.,

Baltensperger, U., 2005c. Vertical distribution of aerosol particles and NOx close




38(24), 6701-6709.




501

Keogh, D.U., Kelly, J., Mengersen, K., Jayaratne, R., Ferreira, L., Morawska, L.,

2009. Derivation of motor vehicle tailpipe particle emission factors suitable for



0.1007/s11356-009-0210-9.

Lowell, D.M., Parsley, W., Bush, C., Zupo, D., 2003. Comparison of Clean

Diesel buses to CNG Buses. 9th Diesel Engine Emissions

Reduction (DEER) Workshop, Newport, RI, USA, 24-28 August.

Macharis, C., Mierlo, J. van., Bossche, P. van. den., 2007. Transportation

Planning and Technology. Combining Intermodal Transport With Electric

Vehicles: Towards More Sustainable Solutions 30(2-3), 311-323.

Morawska, L., Ristovski, Z., Jayaratne, E.R., Keogh, D. U., Ling, X., 2008.


characteristics, ambient processing and implications on human exposure.

Accepted for publication in Atmospheric Environment.




39(23), 9130-9139.

502

Morawska, L., Moore, M. R., Ristovski, Z.D., 2004. Health Impacts of Ultrafine






University of Technology, Brisbane.




NPI (National Pollutant Inventory), Department of the Environment, Water,

Heritage and the Arts, Australian Government, http://www.npi.gov.au/index.html.



Thurston, G.D., 2002. Lung cancer, cardiopulmonary mortality, and long-term


Association 287(9), 1132 1141.

Ristovski, Z.D., Morawska, L., Ayoko, G.A., Jayaratne, E.R., Lim, M., 2002.




503







SKM (Sinclair Knight Merz), 2006. Twice the Task: A review of Australia's

freight transport tasks Melbourne, Victoria, National Transport Commission.





Brisbane.




WHO., 2005. Guidelines for Air Quality. World Health Organization, Geneva.

development of a particle number and …eprints.qut.edu.au/30297/1/diane_keogh_thesis.pdf · jurgen...

Documents