UNIVERSITE OF CALIFORNIA, SAN DIEGO - eScholarship.org

UNIVERSITY OF CALIFORNIA, SAN DIEGO

Investigations into the Impact of Transported Particles on Air Pollution

and Climate Using Aerosol Time-of-Flight Mass Spectrometry

A dissertation submitted in partial satisfaction of the requirements for the degree

Doctor of Philosophy

in

Chemistry

by

Andrew Phillip Ault

Committee in charge:

Professor Kimberly A. Prather, Chair Professor Katja Lindenberg Professor David R. Miller

Professor Amitabha Sinha Professor William C. Trogler

2010

Copyright

Andrew Phillip Ault, 2010

All rights reserved

This dissertation of Andrew Phillip Ault is approved, and it is

acceptable in quality and form for publication on microfilm and

electronically:

__________________________________________________

__________________________________________________

__________________________________________________

__________________________________________________

__________________________________________________

Chair

University of California, San Diego

2010

iii

DEDICATION

Dedicated to my parents Bruce and Helene Ault and sister Julie Ault for all of their love

and support.

iv

EPIGRAPH

Life is what happens while you are busy making other plans.

John Lennon

v

TABLE OF CONTENTS

Signature Page iii

Dedication iv

Epigraph v

Table of contents vi

List of figures xiii

List of tables xvii

Acknowledgements xviii

Vita xxv

Abstract of the Dissertation xxvii

1 Introduction 1

1.1 The Importance of Aerosols 1

1.2 Aerosol Transport and Processing – Local Scale 1

1.3 Aerosol Transport and Processing – Regional Scale 2

1.4 Aerosol Transport and Processing – Intercontinental Scale 3

1.5 Aerosol Time-of-Flight Mass Spectrometry (ATOFMS) 4

1.5.1 ATOFMS Background 4

1.5.2 The ATOFMS Instrument 5

1.5.3 ATOFMS Aerosol Inlet 7

1.5.4 ATOFMS Light Scattering and Particle Sizing Region 8

1.5.5 Laser Desorption/Ionization Time-of-Flight Mass

Spectrometry 13

vi

1.6 Research Objectives and Scope of the Thesis 13

1.7 References 18

2 Impact of Emissions from the Los Angeles Port Region on San

Diego Air Quality During Regional Transport Events 25

2.1 Synopsis 25

2.2 Introduction 26

2.3 Experimental 27

2.3.1 Scripps Institution of Oceanography Pier 2006 Study 27

2.3.2 Real-Time Single-Particle Mass Spectrometry 28

2.3.3 Data Analysis 29

2.3.4 Scaling ATOFMS to Mass Concentrations 29

2.4 Results and Discussion 30

2.4.1 Mass Concentrations Across San Diego County 34

2.4.2 Identification of Regional Events 34

2.4.3 Air Mass History during Regional Events 35

2.4.4 Particle Chemistry during Regional Transport Events 38

2.4.5 General submicron particle types 40

2.4.6 Sulfate Formation and Aging Processes During Transport 42

2.4.7 Temporal Variations by Particle Type 43

2.4.8 Correlation between Major Particle Types 45

2.4.9 Size-Resolved Chemistry of Regional Events 46

2.4.10 Correlation between Ships at Sea and Selected

Particle Types 49

vii

2.4.11 Impact of Regional Events 51

2.5 Acknowledgements 52

2.6 References 53

3 Characterization of the Single Particle Mixing State of Individual

Ship Plume Events Measured at the Port of Los Angeles 58

3.1 Synopsis 58

3.2 Introduction 59

3.3 Experimental 60

3.3.1 Sampling Information 60

3.3.2 Gas and Particle Peripheral Instrumentation 61

3.3.3 Aerosol Time-of-Flight Mass Spectrometry (ATOFMS) 61

3.3.4 Single Particle Analysis 62


3.4.1 Ship Identification 62

3.4.2 Plume Characterization 66

3.4.3 In Plume Gas Phase Chemistry and Concentrations 68

3.4.4 Unique Plume Chemistry 70

3.4.5 Details on the OC-V-sulfate and Fresh Soot

Particle Types 74

3.4.6 Description of the Ca-ECOC Particle Type 77

3.4.7 Background Particle Types During Plumes 79

3.4.8 Temporal Trends 82

3.4.9 Size-resolved Chemistry 84

viii

3.4.10 Correlation between sulfate and vanadium particles 87

3.4.11 Sulfate and Sulfuric Acid in Additional Plumes 90

3.4.12 Conclusions 93


3.6 References 95

4 Mobile Laboratory Observations of Temporal and Spatial

Variability within the Coastal Urban Aerosol 100

4.1 Synopsis 100

4.2 Introduction 101

4.3 Experimental 104

4.3.1 Mobile ATOFMS Laboratory 104

4.3.2 2009 San Diego Bay Measurements 105

4.3.3 Instrumentation 105

4.3.4 Aerosol Time-of-Flight Mass Spectrometry 108

4.3.5 Single Particle Data Analysis 108

4.3.5 Scaling Single Particle Measurements to Mass

Concentrations 109


4.4.1 Temporal Patterns of Meteorological and Gas

Phase Concentrations 110

4.4.2 Variability of Particle Mass and Number Concentrations 113

4.4.3 Variation in Aerosol Size Distributions 115

4.4.4 Differeing Aerosol Sources Determined by Particle

ix

Chemistry 119

4.4.5 Differences in Particle Aging and Mixing State 122

4.4.6 Variations in Chemistry and Implications for Models

and Regulations 124


4.6 References 127

5 Observations of Transported Asian Dust in California Precipitation

during CalWater 131

5.1 Synopsis 131



5.3.1 Meteorological Methods 134

5.3.2 Precipitaation Sampling 135

5.3.3 FLEXPART Model 135

5.4 Results and discussion 136

5.4.1 Atmospheric Rivers Impacting California 136

5.4.2 Meteorology and Precipitation Properties for Storms 1

and 2 144

5.4.3 Transported Asian Dust in Precipitation Samples 146

5.4.4 Back Trajectory Analysis 152

5.4.5 Conclusions 156


5.5 References 158

x

6 Characterization of Asian Outflow at Gosan, Korea by

Single Particle Mass Spectrometry 162

6.1 Synopsis 162



6.3.1 Location and Instrumentation 166

6.3.2 Single particle mass spectrometry 167


6.4.1 Air Mass Characteristics and Meteorology 168

6.4.2 Aerosol Size Distributions and Bulk Mass Concentrations 175

6.4.3 Aerosol Chemistry 181

6.4.4 Aerosol Chemical Mixing States 186

6.5 Conclusion 192


6.7 References 196

7 An Intercontinental Overview of Size-Resolved Chemical Mixing

State by Single Particle Mass Spectrometry 202

7.1 Synopsis 202



7.3.1 Aerosol Time-of-Flight Mass Spectrometry (ATOFMS) 207

7.3.2 Particle Classification 211


xi

7.4.1 Size-Resolved Chemical Composition 218

7.4.1.1 Urban Areas 218

7.4.1.1.1 Aged Urban Areas 218

7.4.1.1.2 Fresh Urban Areas 221

7.4.1.2 Remote Continental Sites 223

7.4.1.3 Marine Studies 225

7.4.2 Internal Mixing with Secondary Speces 229

7.4.2.1 Urban Areas 230

7.4.2.1.1 Aged Urban Areas 230

7.4.2.1.2 Fresh Urban Areas 233

7.4.2.2 Remote Continental Sites 234

7.4.2.3 Marine Studies 236

7.5 Conclusion 241


7.7 References 246

8 Conclusion and Future Directions 258

8.1 Conclusion 258

8.2 Los Angeles Basin Mobile Study (LABMS) Result 262

8.3 Characterization of Aerosol Optical Properties at Gosan, Korea 267

8.4 Future Directions 272

8.5 Final Thought 277

8.6 References 278

xii

LIST OF FIGURES

Figure 1.1 Schematic of the transportable ATOFMS 6

Figure 1.2 Schematic of the ATOFMS light scattering region 9

Figure 1.3 Different refractive indices calculated from Mie Theory 12

Figure 2.1 Time series of mass and number concentrations at the SIO Pier 31

Figure 2.2 Fraction of Scaled PM2.5 corresponding to PM1 hourly 33

Figure 2.3 HYSPLIT back trajectories during transport and non-transport time 37

Figure 2.4 Mass spectra of ECOC and V-Ni-Fe particle type and time series

of both types 39

Figure 2.5 Positive and negative ion mass spectra for the six most

abundant particle types (excluding ECOC and V-Ni-Fe already shown) 41

Figure 2.6 Relative fractions and scaled mass concentrations at the SIO Pier 44

Figure 2.7 Comparison of the average size-resolved mass distributions

during regional and non-event time periods 48

Figure 2.8 Time series of ECOC and V-Ni-Fe and ship radio contacts an event 50

Figure 3.1 Map showing the Port of LA 65

Figure 3.2 Gas and particle measurements of the ship Container1b 67

Figure 3.3 Average mass spectra for OC-V-sulfate and fresh soot 71

Figure 3.4 Digital color stack for the OC-V-sulfate particle type 75

Figure 3.5 Digital color stack for the fresh soot particle type 76

Figure 3.6 Average mass spectrum of the Ca-ECOC particle type 78

Figure 3.7 Average mass spectra for background particle types before

xiii

and after the plume from Container 1a 80

Figure 3.8 Average mass spectra of background particle types from before

and after the plume from Container4 81

Figure 3.9 Time series of the OC-V-sulfate and fresh soot particle types 83

Figure 3.10 Size-resolved number fractions of the different particle types

before and after Container4 and Container1a 86

Figure 3.11 Sulfate and sulfuric acid absolute peak areas by particle type 88

Figure 3.12 Sulfate and sulfuric acid absolute peak areas for additional plumes 92

Figure 4.1 Map of sampling sites and HYSPLIT back trajectories 106

Figure 4.2 Comparison of meteorological conditions between Feb. 12th and

Mar. 13th for San Diego Bay 111

Figure 4.3 Mass and number concentration comparison across sites 114

Figure 4.4 Size distribution measurements for Feb. 12th and Mar. 13th 116

Figure 4.5 Chemically resolved mass fractions across the eight sites 121

Figure 4.6 Secondary marker information across eight sites 123

Figure 5.1 Sampling sites used during CalWater Early Start 133

Figure 5.2 Integrated water vapor and precipitation rates during Calwater

Early Start 138

Figure 5.3 Wind profiles and surface observations from Colfax, California 142

Figure 5.4 Chemical and meteorological measurements during Storms 1 and 2 145

Figure 5.5 ATOFMS mass spectra from precipitation samples 150

Figure 5.6 FLEXPART back trajectory analysis during each precipitation sample 153

Figure 5.7 Schematic summary of Storm 2 case study 155

xiv

Figure 6.1 HYSPLIT back trajectory analysis during different air mass types 171

Figure 6.2 Meteorological data from the Gosan site 172

Figure 6.3 HYSPLIT forward trajectory analysis during different air mass types 174

Figure 6.4 Time series of bulk and size distribution mesaurements 177

Figure 6.5 Size distributions and mass concentrations from each air mass type 178

Figure 6.6 Size and chemically-resolved mass distributions by air mass type 179

Figure 6.7 Number fraction of particle types and secondary marker fractions

by size 180

Figure 6.8 Average mass spectra for the seven main particle types 184

Figure 6.9 Matrix of the number fraction of particles mixed with secondary

markers 185

Figure 6.10 Matrices of the number fraction of particles mixed with secondary

markers for each particle type versus air mass type 188

Figure 6.11 Ternary plots of ammonium, sulfate, and nitrate for each particle

type and air mass type 189

Figure 6.12 Ternary plot of ammonium, sulfate, and nitrate as a function of size 192

Figure 7.1 Global and US maps of ATOFMS study sites 210

Figure 7.2 ATOFMS particle classes 214

Figure 7.3 Size-resolved particle type number fractions for aged urban areas 219

Figure 7.4 Size-resolved particle type number fractions for fresh urban areas 222

Figure 7.5 Size-resolved particle type number fractions for remote continental

sites 224

Figure 7.6 Size-resolved particle type number fractions for marine studies 226

xv

Figure 7.7 Size-resolved particle type number fractions for clean and marine

polluted time periods during marine studies 228

Figure 7.8 Secondary matrices from aged urban areas 231

Figure 7.9 Secondary matrices from fresh urban areas 232

Figure 7.10 Secondary matrices from remote continental 235

Figure 7.11 Secondary matrices from marine studies 237

Figure 7.12 Secondary matrices from aged and polluted marine studies 239

Figure 8.1 PM2.5 concentrations at sampling sites across the Los Angeles Basin 263

Figure 8.2 Back trajectory analysis from each sampling site 264

Figure 8.3 Size-resolved chemical mixing state from each site 265

Figure 8.4 Single particle scattering data from each particle type at Gosan, Korea 269

Figure 8.5 Single particle scattering response as a function of relative humidity 270

Figure 8.6 Top fit of scattering response as a function of size for aged sea salt

at different relative humidities 271

xvi

LIST OF TABLES

Table 2.1 Table of regional events observed at the SIO Pier 32

Table 3.1 Characteristics of ship plumes 64

Table 3.2 Gas phase concentrations during different plume events 69

Table 4.1 Mobile laboratory sampling locations 107

Table 5.1 Atmospheric river conditions for both storms 139

Table 5.2 Precipitation sampling period information 149

Table 6.1 The fraction of time observed for each air mass type 173

Table 7.1 ATOFMS studies used for comparison 209

Table 7.2 Table of characteristic particle types 215

xvii

ACKNOWLEDGMENTS

It is a bit overwhelming to think back over the last 5 years at the number of

people that have supported me during my time in graduate school. Be it with a kind word,

a helpful piece of advice, or lending a hand, these acts of generosity and friendship have

made an intense endeavor a fulfilling and enjoyable experience.

The most important source of support and guidance during my time at UCSD has

been from my advisor Prof. Kim Prather. Looking back it is remarkable the amount of

trust and faith she has placed in me and the rest of the group as we have taken million

dollar instruments around the country and the world. Kim has given me incredible

freedom to explore different aspects of aerosol chemistry ranging from ship plumes to

intercontinentally transported dust. This independence has always been balanced by a

keen scientific insight and an encouraging word when things have gotten tough and

discouraging. This support has taken many different forms, including my personal

favorite - slipping away for a mid-afternoon drink at the pub. Her passion has served as

an inspiration as I pursue my career objective of becoming a university professor. I look

forward to bouncing scientific ideas off of one another, reminiscing about the old days,

and discussing the latest exploits of the Giants and Reds at a corner seat in a local

watering hole during conferences for years to come.

One of the greatest pleasures I have had during graduate school has been the

chance to work with members of the Prather group both past and present. I feel fortunate

to have learned something from all of the members of the Prather group that I have had

the pleasure of working with over the last 5 years including: Dr. Hiroshi Furutani, Dr.

xviii

Steve Toner, Dr. Matt Spencer, Dr. Ryan Moffet, Dr. Laura Shields, Dr. Sharon Qin,

John Holecek, Dr. Thomas Rebotier (and Leah), Dr. Ryan Sullivan, Dr. Kerri Pratt, Dr.

Robert Moision, Dr. Ying Wang, Dr. Yongxuan Su, Meagan Moore, Liz Fitzgerald,

Lindsay Hatch, Cassandra Gaston, Melanie Zauscher, Jessie Creamean, Kaitlyn Suski,

Jack Cahill, Doug Collins, Jessie Charrier, Rene Sanchez, Maggie Yandell, Danny

Phung, and Brandon Heilman. First and foremost I have to thank Joe Mayer for

encouragement and support over the last 5 years. Joe has been a constant source of calm

and logic no matter how harried a field study situation has become. His ability to digest a

situation and devise a solution, often a beautifully simplistic solution, has been an

invaluable asset to all Prather group members over the past nearly two decades. Dr.

Hiroshi Furutani was a constant source of support and guidance during my first two years

in the Prather Lab; he taught me when to push through something, when to reassess and

change tactics, and when to take a deep breath and have a laugh. Dr. Matt Spencer and

Dr. Steve Toner showed me the ropes early on in grad school emphasizing how

imperative it was to make sure you were collecting the best data possible and that you

have to be willing to tear things apart to really learn how they work. The other older

group members when I joined, Dr. Thomas Rebotier, Dr. Ryan Moffet, Dr. Laura Shields,

Dr. Sharon Qin, and John Holecek, always had a kind and encouraging word and were

always willing to share their vast knowledge with me. Though Dr. Yongxuan Su has

moved on from the Prather group, he has remained close by with a watchful eye over all

of us and has always been a source of support and advice over the years. Dr. Kerri Pratt

was always willing to provide her honest opinion and help with a study or a paper. She

also encouraged me to chase fellowships I didn’t think I had a chance to receive. Cassie

xix

Gaston, Liz Fitzgerald, and Lindsay Hatch have all been there to provide an opinion or

listen to a gripe over the years and have been great fellow travelers on the journey that is

graduate school. I feel fortunate to have had the chance to mentor Jessie Creamean during

the last 2 years; through field studies and writing, I have learned as much from her

through this process as she has learned from me. Maggie Yandell was a good

undergraduate to work with and has remained a good friend.

Beyond the Prather Group, many people have lent a helping hand during my years

at UCSD. First and foremost among them has been Myra Kosak who has always been

able to handle the complicated monetary and travel related situations we have produced

over the years. I learned an incredible amount during my time teaching with Dr. Sergio

Guazzotti and Prof. Skip Pomeroy, who through very different styles taught me an

incredible amount about instrumentation over a very small amount of time. Prof.

Pomeroy has always been willing to take a break and talk through whatever is happening

in my grad career or in the world at large. I would also like to thank my thesis committee

for their guidance and ideas over the years: Prof. Amit Sinha, Prof. Katja Lindenberg,

Prof. Bill Trogler, and Prof. Ralph Keeling.

I have been fortunate to work with a fantastic group of fellow scientists and

collaborators during my time in the Prather Group. Our CalWater collaborators at NOAA

Drs. Marty Ralph, Allen White, Christopher Williams, and Paul Neiman have been

fantastic to work with, and I have learned an incredible amount about meteorology and

clouds through working with them. We have also worked during field studies in

California over the years with Prof. Mark Thiemans and Dr. Gerardo Dominguez at

UCSD, Dr. Richard Peiper and Carrie Wold at the Southern California Marine League,

xx

Dr. Robert Loeschen at CSU-LB, Prof. Paul Ziemann at UCR, and many more. During

the Pacific Dust Experiment in Gosan, Korea, I was fortunate to work with Prof. Sang-

Woo Kim, Kyeongsik Kim, Ji Hyoung Kim, Man Hae Kim, Jong Hwan Kim, Prof. Seong

Soo Yum, Aihua Zhu, and Prof. V. Ramanathan. The travel and science involved in these

field studies were incredibly exciting, and I appreciate the incredible amount of

assistance and patience that our scientific collaborators have shown through these studies.

Before arriving at UCSD many people helped influence me during my prior

academic endeavors. During high school chemistry, Ms. Chow and Mr. Lazar both

helped me discover an interest in chemistry that continues to grow. During the summer of

2003, Dr. Paul Fleitz, Dr. Joy Rogers, Dr. Ben Hall, and the rest of the research group at

Wright Patterson Air Force Base provided me with my first research experience. At

Carleton College, my research advisor Prof. Deborah Gross helped nurture my interest in

the atmosphere and mass spectrometry. She has continued to serve as a mentor,

collaborator, and friend over the years since I’ve left Northfield. Prof. Will Hollingsworth

inspired an awe at the quirky and complicated world of quantum mechanics. Working

with my comps (senior thesis) advisor Prof. Daniela Kohen and comps partner Dr. Janel

Uejio to study the world of surface adsorption was an incredible experience that helped

me appreciate research at an in-depth level and prepared me for graduate school. I have

valued Janel’s continued friendship and support as we have wound our way through our

respective graduate programs.

Many people from my personal life have provided an incredible amount of

support to me through my graduate school journey. My parents Prof. Bruce Ault and

Helene Ault have been an incredible source of support and encouragement throughout my

xxi

life. My mom’s relentless positivity and my dad’s calm, reasoned analysis have been

fantastic complementary resources during graduate school. My sister Julie Ault is

probably my most loyal advocate and is always looking out for her older brother. Visits

with Ju have provided needed breaks during graduate school and talks with her have been

a great way to escape the bubble of scientific research. My extended family, especially

those in San Diego (the Blankenships: Kathy, Jay, Carrie, Richard, Corey, Nate, and

Cammie and the Aults: Frank, Evie, Jenny (and Mark), Heather (and Dave)) have been

great sources of support as well. In particular, Uncle Frank and Aunt Evie, have

supported me during my time in SD from the moment I got here all the way through my

defense. My good friend Dave Kayser helped me to develop an interest in math and

science in high school and has remained a good friend. My two running partners from

Carleton, Micah and Becky Johnson, who’ve run a marathon or half marathon together

with me each year for the last 6 years, have provided an excellent outlet and motivation

to keep in shape physically, as well as mentally, and are true friends. In San Diego, Chris

Johnson and Diana Schlamadinger have been constant sources of friendship and a good

laugh. The last two years have been much more fun after three friends from Carleton

moved to San Diego (Andy Ryan, Lexi Gelperin Ryan, and Yui Takeshita). I’ll miss our

trips to Pt. Loma Seafood and other adventures. My roommate of four years Andrew

Markley has been a great source of friendship and patience. Lastly, I must thank Dr. Kerri

Pratt who has inspired me with her quest for knowledge, passion for life, and devotion to

those she truly cares for. You have motivated me and believed in me when I didn’t

believe in myself; I can’t thank you enough for your faith in me, in us, and in our future. I

look forward to continuing our adventure together in the years to come more than you

xxii

can possibly imagine. All of the people listed above have supported and encouraged me

through the years in a variety of ways, and I feel truly blessed to have had their support.

The work in this dissertation was supported by the California Air Resources

Board (CARB), Department of Energy (DOE), Unified Port of San Diego, California

Energy Comission (CEC), and the Atmospheric Brown Cloud project funded under the

United Nations Environmental Programme and the National Oceanic and Atmospheric

Administation (NOAA). I am very grateful for the DOE Global Change Education

Program (GCEP) Graduate Research Environmental Fellowship (GREF) that has

supported me at the end of my graduate school career. Specifically, Prof. Jeff Gaffney,

Milton Constantin, and Rose Etta Cox have been fantastic to work. My involvement in

the GCEP fellowship program has been a tremendously positive experience.

Chapter 2 is reproduced with permission from the American Chemical Society:

Ault, A.P., Moore, M.J., Furutani, H., and K.A. Prather, Impacts of emissions from the

Los Angeles port region on San Diego air quality during Regional Transport Events.

Environmental Science and Technology, 43 (10), 3500-3506, 2009.

Chapter 3 is reproduced with permission from the American Chemical Society:

Ault, A.P., Gaston, C.J., Wang, Y., Dominguez, G.l Thiemens, M.H., and K.A. Prather,

Characterization of the single particle mixing state of individual ship plume events

measured at the Ports of Los Angeles and Long Beach, Environmental Science &

Technology, 44 (6), 1954-1961, 2010.

Chapter 4 is in preparation for submission to Atmospheric Environment: Ault,

A.P., Creamean, J.M., and Prather, K.A., Mobile laboratory observations of temporal and

spatial variability within the coast urban aerosol.

xxiii

Chapter 5 has been submitted to Nature Geoscience: Ault, A.P., Williams, C.R.,

White, A.B., Neiman, P.J., Creamean, J.M., Gaston, C.J., Ralph, F.M., and K.A. Prather.

Potential Changes to California Orographic Precipitation Due to Transported Asian Dust.

Chapter 6 is in preparation for submission to the Atmosperhic Environment: Ault,

A.P., Zhu, A., Ramanathan, V., and K.A. Prather. Analysis of regionally transported

particles in Gosan, Korea by single particle mass spectrometry.

Chapter 7 is in preparation for submission to the Journal of Geophysical

Research: Ault, A.P., Pratt, K.A., Gross, D.S., Dall’Osto, M., Gaelli, M., Harrison, R.M.,

and K.A. Prather. An intercontinental overview of size-resolved chemical mixing state by

single particle mass spectrometry.

xxiv

VITA

2005 B.A. Chemistry, Carleton College

2005-2008 Teaching Assistant, Dept. of Chemistry & Biochemistry, University of

California San Diego

2005-2010 Graduate Research Assistant, Dept. of Chemistry & Biochemistry,

University of California, San Diego

2010 Ph.D. in Chemistry, University of California, San Diego

PUBLICATIONS

Rogers, J. E.; Hall, B. C.; Hufnagle, D. C.; Slagle, J. E.; Ault, A. P.; McLean, D. G.; Fleitz, Paul A.; Cooper, T. M. Effect of platinum on the photophysical properties of a series of phenyl-ethynyl oligomers. Journal of Chemical Physics (2005), 122(21). Ault, A.P.; Moore, M.J.; Furutani, H.; Prather, K.A. Impact of emissions from the Los Angeles port region on San Diego air quality during regional transport events. Environmental Science and Technology. 2009, 43(10), 3500-3506. Ault, A.P.; Gaston, C.J.; Wang, Y.; Dominguez, G.; Thiemens, M.H.; Prather, K.A. Characterization of the single particle mixing state of individual ship plume events measured at the Port of Los Angeles. Environmental Science Technology. 2010, 44(6), 1954-1961. Ault, A.P.; Creamean, J.M.; Prather, K.A. Mobile laboratory observations of temporal and spatial variability within the coastal urban aerosol, Atmospheric Environment, 2010, in preparation Ault, A.P.; Williams, C.R.; White, A.B.; Neiman, P.J.; Creamean, J.M.; Gaston, C.J.; Ralph, F.M.; Prather, K.A. Potential Changes to California Orographic Precipitation Due to Transported Asian Dust Nature Geoscience, 2010, submitted Creamean, J.M.; Ault, A.P.; Gaston, C.J.; Roberts, G.; Prather, K.A. Measurements of aerosol chemistry during new particle formation events at a remote rural site. Environmental Science and Technology, 2010, in preparation Ault, A.P.; Pratt, K.A.; Gross, D.S.; Dall’Osto, M.; Gaelli, M.; Harrison, R.M.; Prather, K.A. An intercontinental overview of size-resolved chemical mixing state by single particle mass spectrometry. Journal of Geophysical Research, 2010, in preparation

xxv

Creamean, J.M.; Hatch, L.E.; Ault, A.P.; Prather, K.A. Signature marine aerosol chemistry observed in an inland urban location during tropical cyclones. Geophysical Research Letters, 2010, in preparation Ault, A.P.; Zhu, A.; Ramanathan, V.; Prather, K.A. Analysis of regionally transported particles in Gosan, Korea by single particle mass spectrometry. Atmospheric Environment, 2010, in preparation Ault, A.P.; Moffet, R.C.; Prather, K.A. Chemical and optical properties of transported carbonaceous and crustal species by single particle mass spectrometry. Geophysical Research Letters, 2010 in preparation

FIELDS OF STUDY

Major Field: Chemistry

Studies in Mass Spectrometry: Professor Kimberly A. Prather

Studies in Atmospheric Chemistry:

Professor Kimberly A. Prather

xxvi

ABSTRACT OF THE DISSERTATION

Investigations into the Impact of Transported Particles on Air Pollution and Climate

Using Aerosol Time-of-Flight Mass Spectrometry

by

Andrew Phillip Ault

Doctor of Philosophy in Chemistry

University of California, San Diego, 2010

Professor Kimberly A. Prather, Chair

Atmospheric aerosols have a significant impact on human health and climate, yet

the full scope of these influences are only beginning to be discovered and characterized.

To understand these impacts, detailed in-situ measurements of the physical, chemical,

and optical properties of aerosols are necessary. Aerosol time-of-flight mass spectrometry

(ATOFMS) provides the ability to measure chemical, physical, and optical properties of

single particles in real-time. This dissertation uses ATOFMS to explore both the

properties and evolution of particles as they are transported over local to global distances.

The results of numerous field studies are utilized to explore the changes to these particles

as they travel through the atmosphere from their source to eventual deposition.

xxvii

xxviii

Local to regional scale transport of particles was observed from a number of

perspectives in this dissertation. Particles regionally transported from the Ports of Los

Angeles and Long Beach to San Diego were identified chemically as ship and truck

emissions and shown to overwhelm local sources during peak transport conditions. Ship

emissions were studied in detail at the Port of Los Angeles by characterizing individual

ship plumes at a site adjacent to the main channel. Mobile laboratory measurements

demonstrated the variation in particle concentrations and composition on a local-to-

regional level.

On the intercontinental-to-global scale, Asian dust was observed in precipitation

samples collected in the Sierra Nevada Mountains during orographic precipitation. The

incorporation of the long range-transported dust might enhance precipitation, which may

alter California’s precipitation patterns and water supply. The outflow of particles from

Asia to North America were measured on a remote island off Korea, and the sources and

aging of particles in Chinese urban, Chinese dust, and Korean air masses were compared

to marine air masses. Lastly, ATOFMS studies from sites across North America, Asia,

Europe, and Africa were compared to determine similarities and differences in size-

resolved chemical mixing state of particles across numerous types of sampling sites, with

the objective being to provide information for global climate models to more accurately

represent particles. Taken together these results provide an increased understanding of

particle chemistry and transport on the scale of meters-to-continents.

1 Introduction

1.1 The Importance of Aerosols

Aerosols have an enormous impact on our world through human inhalation,

scattered solar radiation, and cloud droplet and ice crystal nucleation [Poschl, 2005].

From a human health perspective, high mass concentrations have been linked to

increased mortality [Pope, 2007] and certain types of aerosols have been linked to

specific deleterious health effects [Campen, et al., 2001;Pinkerton, et al., 2004]. Aerosols

have also been shown to have significant influences on climate, but our scientific level of

understanding of these processes is still low [Solomon, et al., 2007]. A small scale

example of these issues is understanding the role of aerosols on buffered systems such as

clouds [Stevens and Feingold, 2009]. While our understanding of aerosols has

dramatically increased in the last decade, a great deal remains to be learned as to their

effect on the whole earth system.

1.2 Aerosol Transport and Processing – Local Scale

One of the challenges in determining the effects of aerosols on climate and human

health is that they are constantly changing from the point of emission to the point of

deposition or inhalation. This occurs through processes such as condensation,

coagulation, heterogeneous reactions, and aqueous phase processing [Finlayson-Pitts and

Pitts, 2000]. Emissions from light duty vehicles (cars) and heavy duty vehicles (trucks)

have been studied extensively to determine changes as emissions age from the tailpipe

[Shields, et al., 2007;Sodeman, et al., 2005;Toner, et al., 2005]. Plumes from ships have

been studied to determine the size, concentrations, and lifetime of plumes in the boundary

1

2

layer [Petzold, et al., 2008]. Studies of power plant and refinery plumes at a coastal site

have observed marine, industrial, and crustal particles coagulating with one another

[Choel, et al., 2010]. Biomass emissions have been studied to look at changes in

chemistry [Yan, et al., 2008] and hygroscopicity [Petters, et al., 2009]. Aircraft emissions

have been investigated to look at changes due to heterogeneous chemistry during plume

evolution [Meilinger, et al., 2005]. Other local scale investigations of aerosol properties

include grass mowing [Drewnick, et al., 2008] and indoor air pollution [Dall'Osto, et al.,

2007].

1.3 Aerosol Transport and Processing – Regional Scale

The transport of aerosols on the time scale of hours to days after emission is

currently a topic of considerable interest in atmospheric chemistry. Recent studies have

investigated the changes occurring to particles as they are transported downwind from

urban areas [Moffet, et al., 2010], volcanic eruptions [Karagulian, et al., 2010], dust

storms [Sullivan, et al., 2007a], and other large sources like biomass burning [Paris, et

al., 2010]. Downwind of urban areas, the mass of organic carbon per particle increases

and the number of double bonds associated with elemental carbon decreases [Moffet, et

al., 2010]. Volcanic eruptions are found to initially have high concentrations of ash

particles, but after 1-2 days, large concentrations of sulfuric acid droplets are found in

these plumes [Karagulian, et al., 2010]. Asian dust and pollution outflow have also been

observed downwind of Asia within days to hours over both land [Chang, et al., 2010] and

the ocean [Bates, et al., 2004]. During dust storms, Asian outflow particles continue to

evolve after leaving the continent by taking up different secondary species as a function

3

of size [Sullivan, et al., 2007a] and serving as sinks for species that form in the plumes

[Sullivan, et al., 2007b].

1.4 Aerosol Transport and Processing – Intercontinental Scale

The fact that particles from one continent contribute significantly to the particle

concentrations of downwind continents has been established through multiple

measurements and models in the last decade [Uno, et al., 2009;VanCuren and Cahill,

2002;Yu, et al., 2008]. Particles transported from Asia have been shown to significantly

contribute to North American aerosol loadings through ground based measurements

[VanCuren and Cahill, 2002] and satellite measurements [Yu, et al., 2008]. In fact, dust

storms from Asia have been shown recently to circumnavigate the earth under certain

conditions [Uno, et al., 2009]. Recent advances in satellite data allow for the detailed

monitoring of Asian dust outbreaks during intercontinental transport [Yumimoto, et al.,

2009]. Outflow from Asia can also take a multilayered form with dust and pollution

mixed near the surface and a “clean” dust layer further aloft [Hara, et al., 2009]. The

chemical composition of aerosols transported intercontinentally has also recently

received considerable attention and multiple papers on this phenomenon have emerged

from the INTEX-B field campaign [Dunlea, et al., 2009;Fairlie, et al., 2010;Leaitch, et

al., 2009].

Although there have been recent advancements within the last 10-15 years about

the chemistry of transported particles, both locally and over vast distances, a great deal

remains to be learned, especially with respect to mixing state, which describes on a single

particle level which species are mixed. Aerosol uncertainties are still considerable from a

climate perspective [Solomon, et al., 2007] and also for more specific problems like

4

modeling intercontinental dust transport [Rastigejev, et al., 2010]. The research presented

in this thesis uses state-of-the-art instrumentation, specifically aerosol time-of-flight mass

spectrometry (ATOFMS), to provide detailed mixing state information of single particles

in real-time. These results provide a clearer picture of particle transport and evolution in

the atmosphere and assist modelers by providing experimental results for constraining

and improving aerosol representation in models.

1.5 Aerosol Time-of-Flight Mass Spectrometry (ATOFMS)

1.5.1 ATOFMS Background

ATOFMS was originally developed as a way to study the variable and complex

chemistry of atmospheric particles. The initial version of the instrument had two

significant limitations: it was not field portable, limiting its atmospheric range to

Riverside, California, and it was single polarity, allowing only positive or negative ions

to be studied at one time [Prather, et al., 1994]. A great deal of information about aerosol

chemistry was obtained from these initial experiments including: automobile emissions

[Silva and Prather, 1997], pyrotechnic emissions (a.k.a. fireworks) [Liu, et al., 1997],

biomass burning particles [Silva, et al., 1999], organic carbon particles [Silva and

Prather, 2000], and size-resolved chemical properties [Noble and Prather, 1996].

The desire to measure the diversity of aerosols in the world led to the

development of the transportable field instrument [Gard, et al., 1997] (Figure 1.1). Since

the development of the portable instrument, ATOFMS studies have been carried out on 3

continents (North America [Moffet, et al., 2008a], Europe [Dall'Osto and Harrison,

2006], and Asia [Spencer, et al., 2008]), on multiple research cruises [Furutani, et al.,

2008;Guazzotti, et al., 2003;Sullivan, et al., 2007a], and at numerous field sites across the

5

United States [Guazzotti, et al., 2001;Liu, et al., 2003;Pastor, et al., 2003;Shields, et al.,

2008;Su, et al., 2006]. A recently-developed version of the ATOFMS, the aircraft (A)-

ATOFMS permits sampling throughout the troposphere [Pratt, et al., 2009b] and has

yielded important findings through the first flights of a dual polarity, single-particle mass

spectrometer [Pratt, et al., 2009a]. ATOFMS has been utilized in over 100 peer reviewed

publications to date and continues to have a significant impact on our understanding of

aerosols.

1.5.2 The ATOFMS instrument

The converging nozzle inlet ATOFMS shown in Figure 1.1a [Gard, et al., 1997]

and the ultrafine (UF)-ATOFMS shown in Figure 1.1b [Su, et al., 2004] are the principle

instruments utilized in this thesis. These two instruments as well as the A-ATOFMS

[Pratt, et al., 2009b] draw particles into a differentially pumped vacuum chamber,

stepping down from atmospheric pressure to 10-7 torr or lower. Particles pass through an

inlet region, a light scattering and particle sizing region, and finally the mass

spectrometer region. Each of these regions is described below.

6

Figure 1.1 Schematic of the transportable ATOFMS with the nozzle inlet a) and aerodynamic lens inlet b). (Figure 1.1a Reprinted with permission from Gard, E., J.E. Mayer, B.D. Morrical, T. Dienes, D.P. Fergenson, and K.A. Prather, Real-time analysis of individual atmospheric aerosol particles: Design and performance of a portable ATOFMS, Analytical Chemistry, 69 (20), 4083-4091, Copyright 1997 American Chemical Society. Figure 1.1b Reprinted with permission from Su, Y.X., M.F. Sipin, H. Furutani, and K.A. Prather, Development and characterization of an aerosol time-of-flight mass spectrometer with increased detection efficiency, Analytical Chemistry, 76 (3), 712-719. Copyright 2004 American Chemical Society)

7

1.5.3 ATOFMS Aerosol Inlet

Initial versions of the ATOFMS utilized a converging nozzle inlet. The

converging nozzle inlet has its highest transmission in the 0.2 – 3.0 μm size range and

was utilized for the measurements reported in Chapters 2, 4, 5, 6, 7, and 8. After the

nozzle, particles pass through 2 skimmers (1.5 mm apart) that are used to collimate the

particle beam. As particles are initially accelerated in the instrument, they undergo a

supersonic expansion and are accelerated to a size-dependent terminal velocity. While

passing through the skimmer region, most gases are pumped away. After passing through

both skimmer stages, particles enter the light scattering region described below.

An aerodynamic lens inlet [Liu, et al., 1995a;Liu, et al., 1995b] was mated with

the ATOFMS to increase transmission at lower sizes than was possible with the

converging nozzle inlet and is referred to as the UF-ATOFMS [Su, et al., 2004]. The

peak transmission in the UF-ATOFMS is between 0.2-0.3 μm, with a range from 0.07-

1.00 μm. With the UF-ATOFMS, observations have been made of single particles in the

ultrafine size range (0.07 – 0.10 μm) during ambient studies [Shields, et al., 2008]. As

particles enter the instrument, they pass through a critical orifice (~ 100 μm inner

diameter) followed by a series of expansions and contractions through larger orifices that

serve to focus the particle beam along the central axis of the lens. The particles then pass

through a skimmer at which point most gases are pumped away and enter the particle

sizing region of the instrument. The UF-ATOFMS was utilized for measurements shown

in Chapter 3 and portions of Chapter 8.

8

1.5.4 ATOFMS Light Scattering and Particle Sizing Region

Particles that enter the light scattering region after passing through the particle

inlet and skimmer regions pass through two continuous wave Nd:YAG lasers placed 6

cm apart and operating at 532 nm. Scattering signal for single particles is collected with

photomultiplier tubes (PMTs) orthogonal to the incident laser beam (Figure 1.2). The

scattered light is focused by an ellipsoidal mirror for each scattering laser with the focal

point being the PMT. The electrical signal produced by the first PMT is used to initiate a

timing circuit. After receiving a signal from the second scattering laser, the time between

scatters is calculated and the particle speed is determined based off the calculated time

and set distance between the two lasers. The speed is then used to calculate the time when

the particle will arrive in the mass spectrometer source region 12 cm below the second

scattering laser. Not only is the particle speed used to calculate the timing for the mass

spectrometer region, but by running polystyrene latex spheres (PSLs) of known size and

density into the instrument, a calibration curve can be determined to convert particle

speed into aerodynamic diameter (Da). The calibration curve is frequently fit with a third-

order polynomial, although this does not work across the full range of ATOFMS and UF-

ATOFMS. Since the calibration curve is empirical, different fits (i.e. power law or fifth

order polynomial) are also used at times.

9

Figure 1.2 Schematic of ATOFMS light scattering region. (Figure 1.2 Reprinted with permission from Moffet, R.C., Qin, X., Rebotier, R., Furutani, H., and Prather, K.A., Chemically segregated optical and microphysical properties of ambient aerosols measured in a single-particle mass spectrometer, Journal of Geophysical Research, 113, D12213, Copyright 2008 American Geophysical Union.

10

A second feature of the light scattering region is that the light scattering signal can

be used to determine the optical properties (i.e. refractive index) of a subset of particles.

This method has been shown previously [Moffet and Prather, 2005;Moffet, et al., 2008b],

but is described briefly below. The area of the particle scattering response is converted to

a partial scattering cross section using Mie theory. This partial scattering cross section is

then plotted as a function of size, frequently revealing Mie “ripples” related to the

scattering pattern off of spherical particles. The upper limit of this distribution is then fit

and compared using a least squares model with a Mie model. As shown in Figure 1.3, as

the real part of the refractive index increases from that of water (1.33) the curves are seen

to “fall over” at lower scattering cross sections. The model fits to geometric diameter,

which is related to aerodynamic diameter by the Equation 1.1:

where Da is aerodynamic diameter, Dp is geometric diameter, χd is the dynamic shape

factor, C(Dp) is the Cunningham slip coefficient for the geometric diameter, C(Da) is the

Cunningham slip coefficient for the aerodynamic diameter, ρp is the particle density, and

ρ0 is the standard density. Thus, the density of a particle group can be determined by

using different density values to minimize the difference between the geometric

theoretical fit and the aerodynamic data. It is important to note that particles below 0.6

μm can have very similar refractive indices and it is useful to have a population of

particles above 0.6 μm to make the most accurate fit. This technique has been used in

11

multiple publications to explore the optical properties of atmospheric particles [Moffet

and Prather, 2009;Moffet, et al., 2008b].

12

Figure 1.3 Diagram demonstrating that many different combinations of refractive index (n) and size (or density) can give the same value for scattered intensity at sizes below 0.6 μm. Figure 1.3 from Moffet, R.C., Qin, X., Rebotier, R., Furutani, H., and Prather, K.A., Chemically segregated optical and microphysical properties of ambient aerosols measured in a single-particle mass spectrometer, Journal of Geophysical Research, 113, D12213, Copyright 2008 American Geophysical Union.

13

1.5.5 Laser Desorption/Ionization Time-of-Flight Mass Spectrometry

The timing calculated in the particle sizing region is used to determine when the

particle will be in the source region of the mass spectrometer. The timing signal is used to

fire a Q-switched Nd:YAG laser (266 nm) that desorbs and ionizes the particles in one

step. The ions created in the source region are then accelerated by sources plates in

opposite directions based on polarity. Each set of ions then passes through a field free

region before reaching reflectrons at the end of each flight tube. The reflectron serves to

increase mass resolution by decreasing differences in the initial kinetic energy

distribution. Ions then pass back through the field free region before striking a set of

multichannel plates (MCP) that transmit the signal to data acquisition boards in the data

collection computer. Ion optics and other improvements to the A-ATOFMS referenced

above have led to higher mass resolution, greater sensitivity, higher mass range, and

fewer mass spectral artifacts [Pratt, et al., 2009b]. All of the data in this instrument was

collected on the transportable dual polarity reflectron time-of-flight mass spectrometers

(ATOFMS and UF-ATOFMS).

1.6 Research Objectives and Scope of this Thesis

Ships represent one of the least regulated sources of anthropogenic aerosols and

efforts to regulate them have been limited by the difficulty of enforcing national laws on

international waters (i.e. > 20 km offshore) and the international scope of the businesses

involved in international shipping. Ships are well known to emit large quantities of

SO2(g), estimated at 2.4 tons annually on a global [Eyring, et al., 2005]. Large amounts of

this SO2 are converted to particulate sulfate, which represents up to 10% of sulfate mass

14

globally. While SO2 and NOx emissions have received considerably attention with

respect to ships [Capaldo, et al., 1999;Corbett and Fischbeck, 1997], metals that are

present in the low grade fuel that ocean-going ships often burn have not been studied in

nearly as much detail [Agrawal, et al., 2008a;Agrawal, et al., 2008b]. Those that have

studied it have focused on bulk techniques that cannot determine the particle mixing state

[Agrawal, et al., 2008a;Agrawal, et al., 2008b]. Chapter 2 of this thesis demonstrates the

impact that ship emissions can have regionally when transported. Also, a particle type

with a strong vanadium signature is identified and suggested to represent a marker for

ship emission transport from the Los Angeles port region. Chapter 3 follows up on

Chapter 2 by measuring individual ship plumes with ATOFMS at the Ports of Los

Angeles and Long Beach. Vanadium is shown to correlate highly with ships burning

bunker fuel and is hypothesized to facilitate the oxidation of SO2 to sulfate. This

potentially catalytic reaction helps to explain the order of magnitude greater

concentrations of sulfate and sulfuric acid on vanadium-containing particles in the

plumes. This work has the potential to improve global models predicting sulfate

concentrations by giving a more detailed description of the processes present in a typical

ship plume. Combined, these chapters strengthen our understanding of the nature of ship

emissions near the source and their long range impacts on coastal communities.

The detailed information measured by ATOFMS provides a snapshot of the

complex processes occurring in the atmosphere in real-time. One of the limitations of

these instruments is that there are only a limited number of them in the world and thus

only a few attempts have been made to measure the variation in particle chemistry over

short spatial distances with high time resolution. One such effort involved ATOFMS

15

instruments operating at Long Beach, Fullerton, and Riverside, California [Hughes, et al.,

2000]. Data were compared during trajectories that crossed all 3 sites, providing insight

into particle aging processes [Hughes, et al., 2000]. Chapter 4 in this thesis takes a

different approach by utilizing the ATOFMS mobile laboratory, which is powered by

generators to gain maximal flexibility in where and when sampling can take place. This

mobile laboratory developed as a key element of this thesis project provides the ability to

sample in multiple locations over the course of one day, thereby gaining multiple snap

shots that can be compared to study temporal and spatial variability of aerosols. The

mobile laboratory sampled at 5 sites in 1 day while sampling around San Diego Bay, a

record unlikely to be broken. The results of this study give us insight into how particle

physical and chemical properties vary around San Diego Bay on clean and polluted days.

In addition, the mobile laboratory is now a staple of ATOFMS field studies in California.

The influence of aerosols on clouds and precipitation is a relatively new focus of

study due to the difficulty in teasing apart the influence of dynamics versus aerosols in a

buffered system [Stevens and Feingold, 2009]. One challenge has been that the new and

complex tools necessary to measure each of these components with the resolution

necessary for comparison have not been combined until recently. Chapter 5 describes a

collaborative effort with the National Oceanic and Atmospheric Administration (NOAA)

to combine state-of-the-art meteorological measurements with detailed chemical analysis

from the ATOFMS. This chapter compares two storms with atmospheric rivers, a feature

responsible for the majority of California’s wintertime precipitation [Neiman, et al.,

2008]. While their integrated water vapor contents were similar, one storm was impacted

by transported Asian dust and one was not. This was observed experimentally by

16

ATOFMS analysis of precipitation samples, matching predictions based on the

Lagrangian transport model, FLEXPART [Stohl, et al., 2005]. The results described in

this chapter demonstrate the potential influence of aerosols on precipitation patterns and

intensity and establish a framework for exploring this field further on future field

campaigns.

Given the potential importance of Asian dust on North American climate as

described in Chapter 5, it is important to understand the properties of these transported

aerosols en route on their transcontinental journey. Chapter 6 describes measurements

made at Gosan on Jeju Island off the southern coast of South Korea. This sampling

location provides an optimal observational location since it is on the west coast of the

island and there is nothing but ocean between the site and China. With transport times of

roughly one day from China, the measurements at Gosan provide a detailed look at

particle composition and aging early in the process of long range transport. These

properties are compared with time periods when particles at the site came from sources

other than China.

Chapter 7 details an effort to describe similarities and differences across different

types of sampling locations by combining results from numerous ATOFMS field

campaigns from multiple groups. Field campaigns from 3 continents (Europe, North

America, and Asia) are compared and details on the size-resolved chemistry and

secondary mixing state are presented. It is hoped that by presenting both the strengths and

limitations of ATOFMS data that this paper can correct numerous persistent

misconceptions in the atmospheric community at large about single particle mass

spectrometry. This collaborative effort is aimed at providing a reference to modelers and

17

other aerosol and atmospheric scientists attempting to understand single particle mass

spectrometry measurements.

18

1.7 References

H. Agrawal, Q. G. J. Malloy, W. A. Welch, J. Wayne Miller and D. R. Cocker III, In-use gaseous and particulate matter emissions from a modern ocean going container vessel, Atmospheric Environment, 42 (21), 5504, 2008a.

H. Agrawal, W. A. Welch, J. W. Miller and D. R. Cocker, Emission Measurements from

a Crude Oil Tanker at Sea, Environ. Sci. Technol., 2008b. T. S. Bates, P. K. Quinn, D. J. Coffman, D. S. Covert, T. L. Miller, J. E. Johnson, G. R.

Carmichael, I. Uno, S. A. Guazzotti, D. A. Sodeman, K. A. Prather, M. Rivera, L. M. Russell and J. T. Merrill, Marine boundary layer dust and pollutant transport associated with the passage of a frontal system over eastern Asia, Journal Of Geophysical Research-Atmospheres, 109 (D19), 2004.

M. J. Campen, J. P. Nolan, M. C. J. Schladweiler, U. P. Kodavanti, P. A. Evansky, D. L.

Costa and W. P. Watkinson, Cardiovascular and thermoregulatory effects of inhaled PM-associated transition metals: A potential interaction between nickel and vanadium sulfate, Toxicological Sciences, 64 (2), 243-252, 2001.

K. Capaldo, J. J. Corbett, P. Kasibhatla, P. Fischbeck and S. N. Pandis, Effects of ship

emissions on sulphur cycling and radiative climate forcing over the ocean, Nature, 400 (6746), 743-746, 1999.

S. C. Chang, C. C. K. Chou, W. N. Chen and C. T. Lee, Asian dust and pollution

transport-A comprehensive observation in the downwind Taiwan in 2006, Atmospheric Research, 95 (1), 19-31, 2010.

M. Choel, K. Deboudt and P. Flament, Development of Time-Resolved Description of

Aerosol Properties at the Particle Scale During an Episode of Industrial Pollution Plume, Water Air And Soil Pollution, 209 (1-4), 93-107, 2010.

J. J. Corbett and P. Fischbeck, Emissions from ships, Science, 278 (5339), 823-824, 1997. M. Dall'Osto and R. M. Harrison, Chemical characterisation of single airborne particles

in Athens (Greece) by ATOFMS, Atmospheric Environment, 40 (39), 7614-7631, 2006.

M. Dall'Osto, R. M. Harrison, E. Charpantidou, G. Loupa and S. Rapsomanikis,

Characterisation of indoor airborne particles by using real-time aerosol mass spectrometry, Science Of The Total Environment, 384 (1-3), 120-133, 2007.

F. Drewnick, M. Dall'Osto and R. Harrison, Characterization of aerosol particles from

grass mowing by joint deployment of ToF-AMS and ATOFMS instruments, Atmospheric Environment, 42 (13), 3006-3017, 2008.

19

E. J. Dunlea, P. F. DeCarlo, A. C. Aiken, J. R. Kimmel, R. E. Peltier, R. J. Weber, J.

Tomlinson, D. R. Collins, Y. Shinozuka, C. S. McNaughton, S. G. Howell, A. D. Clarke, L. K. Emmons, E. C. Apel, G. G. Pfister, A. van Donkelaar, R. V. Martin, D. B. Millet, C. L. Heald and J. L. Jimenez, Evolution of Asian aerosols during transpacific transport in INTEX-B, Atmospheric Chemistry Physics, 9 (19), 7257-7287, 2009.

V. Eyring, H. W. Kohler, J. van Aardenne and A. Lauer, Emissions from international

shipping: 1. The last 50 years, Journal Of Geophysical Research-Atmospheres, 110 (D17), 2005.

T. D. Fairlie, D. J. Jacob, J. E. Dibb, B. Alexander, M. A. Avery, A. van Donkelaar and

L. Zhang, Impact of mineral dust on nitrate, sulfate, and ozone in transpacific Asian pollution plumes, Atmospheric Chemistry And Physics, 10 (8), 3999-4012, 2010.

B. J. Finlayson-Pitts and J. N. Pitts, Chemistry of the Upper and Lower Atmosphere:

Theory, Experiments, and Applications, Academic Press, San Diego, 2000. H. Furutani, M. Dall'osto, G. C. Roberts and K. A. Prather, Assessment of the relative

importance of atmospheric aging on CCN activity derived from field observations, Atmos. Environ., 42 (13), 3130-3142, 2008.

E. Gard, J. E. Mayer, B. D. Morrical, T. Dienes, D. P. Fergenson and K. A. Prather, Real-

time analysis of individual atmospheric aerosol particles: Design and performance of a portable ATOFMS, Analytical Chemistry, 69 (20), 4083-4091, 1997.

S. A. Guazzotti, D. T. Suess, K. R. Coffee, P. K. Quinn, T. S. Bates, A. Wisthaler, A.

Hansel, W. P. Ball, R. R. Dickerson, C. Neususs, P. J. Crutzen and K. A. Prather, Characterization of carbonaceous aerosols outflow from India and Arabia: Biomass/biofuel burning and fossil fuel combustion, Journal Of Geophysical Research-Atmospheres, 108 (D15), 2003.

S. A. Guazzotti, J. R. Whiteaker, D. Suess, K. R. Coffee and K. A. Prather, Real-time

measurements of the chemical composition of size-resolved particles during a Santa Ana wind episode, California USA, Atmospheric Environment, 35 (19), 3229-3240, 2001.

Y. Hara, K. Yumimoto, I. Uno, A. Shimizu, N. Sugimoto, Z. Liu and D. M. Winker,

Asian dust outflow in the PBL and free atmosphere retrieved by NASA CALIPSO and an assimilated dust transport model, Atmospheric Chemistry And Physics, 9 (4), 1227-1239, 2009.

20

L. S. Hughes, J. O. Allen, P. Bhave, M. J. Kleeman, G. R. Cass, D. Y. Liu, D. F. Fergenson, B. D. Morrical and K. A. Prather, Evolution of atmospheric particles along trajectories crossing the Los Angeles basin, Environmental Science & Technology, 34 (15), 3058-3068, 2000.

F. Karagulian, L. Clarisse, C. Clerbaux, A. J. Prata, D. Hurtmans and P. F. Coheur,

Detection of volcanic SO2, ash, and H2SO4 using the Infrared Atmospheric Sounding Interferometer (IASI), Journal Of Geophysical Research-Atmospheres, 115, 2010.

W. R. Leaitch, A. M. Macdonald, K. G. Anlauf, P. S. K. Liu, D. Toom-Sauntry, S. M. Li,

J. Liggio, K. Hayden, M. A. Wasey, L. M. Russell, S. Takahama, S. Liu, A. van Donkelaar, T. Duck, R. V. Martin, Q. Zhang, Y. Sun, I. McKendry, N. C. Shantz and M. Cubison, Evidence for Asian dust effects from aerosol plume measurements during INTEX-B 2006 near Whistler, BC, 9 (11), 3523-3546, 2009.

D. Y. Liu, D. Rutherford, M. Kinsey and K. A. Prather, Real-time monitoring of

pyrotechnically derived aerosol particles in the troposphere, Analytical Chemistry, 69 (10), 1808-1814, 1997.

D. Y. Liu, R. J. Wenzel and K. A. Prather, Aerosol time-of-flight mass spectrometry

during the Atlanta Supersite Experiment: 1. Measurements, Journal Of Geophysical Research-Atmospheres, 108 (D7), 2003.

P. Liu, P. J. Ziemann, D. B. Kittelson and P. H. McMurry, Generating Particle Beams Of

Controlled Dimensions And Divergence .1. Theory Of Particle Motion In Aerodynamic Lenses And Nozzle Expansions, Aerosol Science And Technology, 22 (3), 293-313, 1995a.

P. Liu, P. J. Ziemann, D. B. Kittelson and P. H. McMurry, Generating Particle Beams Of

Controlled Dimensions And Divergence .2. Experimental Evaluation Of Particle Motion In Aerodynamic Lenses And Nozzle Expansions, Aerosol Science And Technology, 22 (3), 314-324, 1995b.

S. K. Meilinger, B. Karcher and T. Peter, Microphysics and heterogeneous chemistry in

aircraft plumes - high sensitivity on local meteorology and atmospheric composition, Atmospheric Chemistry And Physics, 5, 533-545, 2005.

R. C. Moffet, B. de Foy, L. T. Molina, M. J. Molina and K. A. Prather, Measurement of

ambient aerosols in northern Mexico City by single particle mass spectrometry, Atmospheric Chemistry And Physics, 8 (16), 4499-4516, 2008a.

R. C. Moffet, T. R. Henn, A. V. Tivanski, R. J. Hopkins, Y. Desyaterik, A. L. D.

Kilcoyne, T. Tyliszczak, J. Fast, J. Barnard, V. Shutthanandan, S. S. Cliff, K. D.

21

Perry, A. Laskin and M. K. Gilles, Microscopic characterization of carbonaceous aerosol particle aging in the outflow from Mexico City, Atmospheric Chemistry And Physics, 10 (3), 961-976, 2010.

R. C. Moffet and K. A. Prather, Extending ATOFMS measurements to include refractive

index and density, Analytical Chemistry, 77 (20), 6535-6541, 2005. R. C. Moffet and K. A. Prather, In-situ measurements of the mixing state and optical

properties of soot with implications for radiative forcing estimates, Proceedings Of The National Academy Of Sciences Of The United States Of America, 106 (29), 11872-11877, 2009.

R. C. Moffet, X. Y. Qin, T. Rebotier, H. Furutani and K. A. Prather, Chemically

segregated optical and microphysical properties of ambient aerosols measured in a single-particle mass spectrometer, Journal Of Geophysical Research-Atmospheres, 113 (D12), 2008b.

P. J. Neiman, F. M. Ralph, G. A. Wick, J. D. Lundquist and M. D. Dettinger,

Meteorological characteristics and overland precipitation impacts of atmospheric rivers affecting the West Coast of North America based on eight years of SSM/I satellite observations, Journal of Hydrometeorology, 9 (1), 22-47, 2008.

C. A. Noble and K. A. Prather, Real-time measurement of correlated size and

composition profiles of individual atmospheric aerosol particles, Environmental Science & Technology, 30 (9), 2667-2680, 1996.

J. D. Paris, A. Stohl, P. Ciais, P. Nedelec, B. D. Belan, M. Y. Arshinov and M. Ramonet,

Source-receptor relationships for airborne measurements of CO2, CO and O-3 above Siberia: a cluster-based approach, Atmospheric Chemistry And Physics, 10 (4), 1671-1687, 2010.

S. H. Pastor, J. O. Allen, L. S. Hughes, P. Bhave, G. R. Cass and K. A. Prather, Ambient

single particle analysis in Riverside, California by aerosol time-of-flight mass spectrometry during the SCOS97-NARSTO, Atmospheric Environment, 37, S239-S258, 2003.

M. D. Petters, C. M. Carrico, S. M. Kreidenweis, A. J. Prenni, P. J. DeMott, J. L. Collett

and H. Moosmuller, Cloud condensation nucleation activity of biomass burning aerosol, J. Geophys. Res.-Atmos., 114, 2009.

A. Petzold, J. Hasselbach, P. Lauer, R. Baumann, K. Franke, C. Gurk, H. Schlager and E.

Weingartner, Experimental studies on particle emissions from cruising ship, their characteristic properties, transformation and atmospheric lifetime in the marine boundary layer, Atmos. Chem. Phys., 8 (9), 2387-2403, 2008.

22

K. E. Pinkerton, Y. M. Zhou, S. V. Teague, J. L. Peake, R. C. Walther, I. M. Kennedy, V. J. Leppert and A. E. Aust, Reduced lung cell proliferation following short-term exposure to ultrafine soot and iron particles in neonatal rats: Key to impaired lung growth?, Inhalation Toxicology, 16, 73-81, 2004.

C. A. Pope, Mortality effects of longer term exposures to fine particulate air pollution:

Review of recent epidemiological evidence, Inhalation Toxicology, 19, 33-38, 2007.

U. Poschl, Atmospheric aerosols: Composition, transformation, climate and health

effects, Angewandte Chemie-International Edition, 44 (46), 7520-7540, 2005. K. A. Prather, T. Nordmeyer and K. Salt, Real-Time Characterization Of Individual

Aerosol-Particles Using Time-Of-Flight Mass-Spectrometry, Analytical Chemistry, 66 (9), 1403-1407, 1994.

K. A. Pratt, P. J. DeMott, J. R. French, Z. Wang, D. L. Westphal, A. J. Heymsfield, C. H.

Twohy, A. J. Prenni and K. A. Prather, In situ detection of biological particles in cloud ice-crystals, Nature Geoscience, 2 (6), 397-400, 2009a.

K. A. Pratt, J. E. Mayer, J. C. Holecek, R. C. Moffet, R. O. Sanchez, T. P. Rebotier, H.

Furutani, M. Gonin, K. Fuhrer, Y. X. Su, S. Guazzotti and K. A. Prather, Development and Characterization of an Aircraft Aerosol Time-of-Flight Mass Spectrometer, Analytical Chemistry, 81 (5), 1792-1800, 2009b.

Y. Rastigejev, R. Park, M. P. Brenner and D. J. Jacob, Resolving intercontinental

pollution plumes in global models of atmospheric transport, Journal Of Geophysical Research-Atmospheres, 115, 2010.

L. G. Shields, X. Y. Qin, S. M. Toner and K. A. Prather, Detection of ambient ultrafine

aerosols by single particle techniques during the SOAR 2005 campaign, Aerosol Science And Technology, 42 (8), 674-684, 2008.

L. G. Shields, D. T. Suess and K. A. Prather, Determination of single particle mass

spectral signatures from heavy-duty diesel vehicle emissions for PM2.5 source apportionment, Atmospheric Environment, 41 (18), 3841-3852, 2007.

P. J. Silva, D. Y. Liu, C. A. Noble and K. A. Prather, Size and chemical characterization

of individual particles resulting from biomass burning of local Southern California species, Environmental Science & Technology, 33 (18), 3068-3076, 1999.

P. J. Silva and K. A. Prather, On-line characterization of individual particles from

automobile emissions, Environmental Science & Technology, 31 (11), 3074-3080, 1997.

23

P. J. Silva and K. A. Prather, Interpretation of mass spectra from organic compounds in

aerosol time-of-flight mass spectrometry, Analytical Chemistry, 72 (15), 3553-3562, 2000.

D. A. Sodeman, S. M. Toner and K. A. Prather, Determination of single particle mass

spectral signatures from light-duty vehicle emissions, Environmental Science & Technology, 39 (12), 4569-4580, 2005.

S. Solomon, D. Qin, M. Manning, Z. Chen, M. Marquis, K. B. Averyt, M. Tignor and H.

L. Miller, Climate Change 2007: The Physical Science Basis. Contribution of Working Group 1 to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change,Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 2007.

M. T. Spencer, J. C. Holecek, C. E. Corrigan, V. Ramanathan and K. A. Prather, Size-

resolved chemical composition of aerosol particles during a monsoonal transition period over the Indian Ocean, 113 (D16), 2008.

B. Stevens and G. Feingold, Untangling aerosol effects on clouds and precipitation in a

buffered system, Nature, 461 (7264), 607-613, 2009. A. Stohl, C. Forster, A. Frank, P. Seibert and G. Wotawa, Technical note: The

Lagrangian particle dispersion model FLEXPART version 6.2, Atmos. Chem. Phys., 5, 2461-2474, 2005.

Y. X. Su, M. F. Sipin, H. Furutani and K. A. Prather, Development and characterization

of an aerosol time-of-flight mass spectrometer with increased detection efficiency, Analytical Chemistry, 76 (3), 712-719, 2004.

Y. X. Su, M. F. Sipin, M. T. Spencer, X. Y. Qin, R. C. Moffet, L. G. Shields, K. A.

Prather, P. Venkatachari, C. H. Jeong, E. Kim, P. K. Hopke, R. M. Gelein, M. J. Utell, G. Oberdorster, J. Berntsen, R. B. Devlin and L. C. Chen, Real-time characterization of the composition of individual particles emitted from ultrafine particle concentrators, Aerosol Science And Technology, 40 (6), 437-455, 2006.

R. C. Sullivan, S. A. Guazzotti, D. A. Sodeman and K. A. Prather, Direct observations of

the atmospheric processing of Asian mineral dust, Atmospheric Chemistry And Physics, 7, 1213-1236, 2007a.

R. C. Sullivan, S. A. Guazzotti, D. A. Sodeman, Y. H. Tang, G. R. Carmichael and K. A.

Prather, Mineral dust is a sink for chlorine in the marine boundary layer, Atmospheric Environment, 41 (34), 7166-7179, 2007b.

24

S. M. Toner, L. G. Shields, D. A. Sodeman and K. A. Prather, Using mass spectral source signatures to apportion exhaust particles from gasoline and diesel powered vehicles in a freeway study, Abstracts Of Papers Of The American Chemical Society, 229, U127-U127, 2005.

I. Uno, K. Eguchi, K. Yumimoto, T. Takemura, A. Shimizu, M. Uematsu, Z. Y. Liu, Z. F.

Wang, Y. Hara and N. Sugimoto, Asian dust transported one full circuit around the globe, Nature Geoscience, 2 (8), 557-560, 2009.

R. A. VanCuren and T. A. Cahill, Asian aerosols in North America: Frequency and

concentration of fine dust, J. Geophys. Res.-Atmos., 107 (D24), 2002. B. Yan, M. Zheng, Y. T. Hu, S. Lee, H. K. Kim and A. G. Russell, Organic composition

of carbonaceous aerosols in an aged prescribed fire plume, Atmospheric Chemistry And Physics, 8 (21), 6381-6394, 2008.

H. B. Yu, L. A. Remer, M. Chin, H. S. Bian, R. G. Kleidman and T. Diehl, A satellite-

based assessment of transpacific transport of pollution aerosol, Journal Of Geophysical Research-Atmospheres, 113 (D14), 2008.

K. Yumimoto, K. Eguchi, I. Uno, T. Takemura, Z. Liu, A. Shimizu and N. Sugimoto, An

elevated large-scale dust veil from the Taklimakan Desert: Intercontinental transport and three-dimensional structure as captured by CALIPSO and regional and global models, Atmos. Chem. Phys., 9 (21), 8545-8558, 2009.

2 Impact of Emissions from the Los Angeles Port Region on

San Diego Air Quality during Regional Transport Events

2.1 Synopsis

Oceangoing ships emit an estimated 1.2–1.6 million metric tons (Tg) of PM10 per

year and represent a significant source of air pollution to coastal communities. As shown

herein, ship and other emissions near the Los Angeles and Long Beach Port region

strongly influence air pollution levels in the San Diego area. During time periods with

regional transport, atmospheric aerosol measurements in La Jolla, California show an

increase in 0.5-1 m sized single particles with unique signatures including soot, metals

(i.e. vanadium, iron, and nickel), sulfate, and nitrate. These particles are attributed to

primary emissions from residual oil sources such as ships and refineries, as well as traffic

in the port region, and secondary processing during transport. During regional transport

events, particulate matter concentrations were 2-4 times higher than typical average

concentrations from local sources, indicating the health, environmental, and climate

impacts from these emission sources must be taken into consideration in the San Diego

region. Unless significant regulations are imposed on shipping-related activities, these

emission sources will become even more important to California air quality as cars and

truck emissions undergo further regulations and residual oil sources such as shipping

continue to expand.

25

26

2.2 Introduction

Atmospheric aerosols regionally and globally impact climate and health.

Regionally transported particles contribute significantly to PM2.5 as observed in multiple

studies [Allen and Turner, 2008;Lall and Thurston, 2006]. Oceangoing ships emit an

estimated 1.2–1.6 million metric tons (Tg) of PM10 per year [Corbett and Koehler, 2003]

and the 2006 California Air Resources Board emissions inventory estimates that ships

emit more PM2.5 (4.94 tons/day) than refineries and power plants combined (3.45

tons/day), representing the second highest PM mobile source in Los Angeles County

behind diesel trucks (5.87 tons/day) [C.A.R.B., 2006]. In coastal regions, models show a

significant impact from the emissions of ships and residual fuel combustion on air

pollution levels. Strong implications exist for human health [Capaldo, et al.,

1999;Corbett, et al., 2007;Vutukuru and Dabdub, 2008], but major uncertainties exist

regarding the spatial and seasonal variability of particulate pollution that need further

investigation [Martien, et al., 2003]. Field observations have confirmed that ships

produce significant amounts of soot, vanadium, nickel, and sulfate [Lu, et al., 2006;Xie,

et al., 2006]. Ships typically burn residual fuel oil, which produces higher concentrations

of heavy metals and soot than distillate fuels such as gasoline and diesel [International

Standards Organization, 2005;Pattanaik, et al., 2007;Vallius, et al., 2003]. Additional

residual fuel oil sources, such as refineries, are often located near large ports and emit

metal-containing particles that are particularly harmful to human health [Eatough, et al.,

1984;Fang, et al., 2008;Jang, et al., 2007]. These deleterious effects can be explained by

the chemistry of the emissions as well as proposed synergistic relationships between

coemitted chemical species such as soot and iron or vanadium and nickel [Campen, et al.,

27

2001;Pinkerton, et al., 2004;Zhou, et al., 2003]. Recent studies in the Los Angeles and

Long Beach (LA-LB) port region have shown vanadium as having the strongest

correlation to reactive oxygen species using a macrophage assay [Hu, et al., 2008]. These

negative health effects combined with the transport of particles from residual fuel

combustion make it imperative to determine their contributions to the air pollution in

coastal locations along the California coast.

Regionally transported air pollution has been shown to influence the San Diego

region in previous studies. Aircraft measurements have shown O3 transport from the Los

Angeles (LA) basin and SO2 transport from offshore to San Diego [Kauper and

Niemann, 1977]. Particles transported from LA have also been observed in San Diego at

an urban freeway location by aerosol time-of-flight mass spectrometry (ATOFMS)

[Toner, 2007]. This paper discusses distinct episodes quantifying the impact of regional

transport when the air masses originated near the vicinity of the Port of LA observed by

real-time single particle mass spectrometry at the Scripps Institution of Oceanography

(SIO) Pier in La Jolla, California. During these transport periods, the highest

concentrations of submicron carbonaceous and transition metal particles (V, Ni, and Fe)

were measured. This study examines the impact of these transported aerosols on PM

concentrations in the San Diego region.

2.3 Experimental

2.3.1 Scripps Institution of Oceanography Pier 2006 Study

From August 17 to October 3, 2006, sampling was conducted at the end of the

SIO Pier, three hundred meters from shore. Ambient particles were sampled through a

four meter mast, seven meters above the pier, and fifteen meters above the ocean.

28

Different instruments sampled in parallel through insulated sampling lines. A five hour

period of sampling on September 28, 2006 was dominated by submicron particles from a

local wildfire, as determined by the ATOFMS mass spectral signatures, and because the

focus of this paper is on regional transport impacts, this period is not discussed. Particle

size distributions were optically measured using an aerodynamic particle sizer

spectrometer (APS 3321, TSI Inc.), which measures size distributions between 0.5-3 m.

Air mass back trajectory calculations were made using HYSPLIT 4.8 [Draxler and

Rolph, 2003].

2.3.2 Real-Time Single-Particle Mass Spectrometry

The aerodynamic size and chemical composition of individual particles between

0.2 and 3.0 μm were measured in real-time using an aerosol time-of-flight mass

spectrometer (ATOFMS); the design and details have been described in detail elsewhere

[Gard, et al., 1997]. Briefly, particles are introduced into the ATOFMS through a

converging nozzle into a differentially pumped vacuum chamber where they are

accelerated to a terminal velocity. The particles then pass through two continuous wave

lasers (diode pumped Nd:YAG lasers operating at 532 nm) located 6 cm apart. The

velocity of each particle is used to determine the aerodynamic size using speeds from

calibration with polystyrene latex spheres of known aerodynamic size, shape, and

density. The sized particles are then desorbed and ionized by irradiation from a Q-

switched Nd:YAG laser (266nm, 1.2-1.4 mJ/pulse) that is triggered when the particle

enters the center of the mass spectrometer ion source. Positive and negative ions from

each particle are detected using a dual-polarity, reflectron time-of-flight mass

spectrometer. During this study, ambient particles were dried using a silica gel diffusion

29

drier prior to analysis with the ATOFMS to decrease particle water content and,

therefore, increase negative ion signal intensities [Neubauer, et al., 1997]. The ATOFMS

sized and chemically analyzed 3,502,907 particles during the study.

2.3.3 Data Analysis

Single particle size and mass spectral information were analyzed with YAADA

1.2 (www.yaada.org) a data analysis toolkit for MATLAB 6.5.1 (The MathWorks, Inc.).

Particles were analyzed via two approaches: (1) searching mass spectral, aerodynamic

size, and temporal features and (2) clustering mass spectra using an adaptive resonance

theory based neural network algorithm (ART-2a) at a vigilance factor of 0.8 [Song, et al.,

1999]. ART-2a combines particles into clusters based on the intensity of ion peaks in

individual mass spectra. General particle types are defined by the characteristic chemical

species or possible source to simplify the naming scheme; these labels do not reflect all

of the species present within a specific particle type, but reflect the most intense ion

peaks. Peak identifications within this paper correspond to the most probable ions for a

given m/z ratio.

2.3.4 Scaling ATOFMS to Mass Concentrations

ATOFMS counts were scaled to number and mass concentrations using APS size-

resolved number concentrations, a method shown previously to yield quantitative mass

concentrations [Qin, et al., 2006]. Briefly, ATOFMS counts are binned into sizes that

correspond to the APS size bins (0.5-2.5 m) and scaling factors for each hour and size

bin are determined to account for ATOFMS transmission biases and busy time. After

scaling the number concentrations, density is used to convert from number to mass

concentrations. The bins were summed to give the mass of particulate matter less than 1

30

m (PM1) or less than 2.5 m (PM2.5). Scaled ATOFMS PM2.5 mass concentrations have

been shown previously to be well correlated with standard mass measurements, including

the beta attenuation monitor (BAM) and micro-orifice uniform deposit impactor

(MOUDI) [Qin, et al., 2006]. ATOFMS scaled PM1 concentrations have been shown to

track mass concentrations [Liu, et al., 2000;Qin, 2007]. Previous work has shown that

with high relative humidity [Moffet, et al., 2008] or liquid water content conditions

[Spencer, et al., 2007] particle density is low; thus, a density of 1.2 μg/m3 has been used

for the scaling herein due to an average relative humidity of 93% during sampling.

2.4 Results and Discussion

Figure 2.1 shows the PM concentrations measured over the full study,

highlighting the three high concentration regional transport periods. During many of the

regional events, daily PM concentration maxima were observed over several days due to

diurnal changes in wind speed and direction. Details on the impacts of the regional

events on PM concentrations are described below. Table 2.1 lists each identified regional

event with accompanying properties, discussed in detail in the following sections. The

fraction of PM1 contributing to PM2.5 is shown per hour in Figure 2.2. During submicron

events, PM1 mass accounted for 60% of the PM2.5 mass compared to 36% on average

during non-event periods.

31

Figure 2.1 A) The lower portion of the graph shows time series of scaled PM2.5 mass concentrations at the SIO Pier (red line) and PM2.5 data from downtown San Diego (black line). B) Hourly size distributions measured by an aerodynamic particle sizer (APS) from 0.5-1 m from August 17, 2006 – October 2, 2006. The overlaid blue line is the sum of ATOFMS counts in the 0.5-1 m size range per hour. C) Scaled PM1 (green line) and PM2.5 (red line) mass concentrations are shown over the sampling period. The gray dashed line represents the average PM1 mass concentration during the study, and the black dashed line represents the PM1 threshold set for regional events.

32

Table 2.1 Properties of different regional events are listed including: peak PM1 and PM2.5 mass concentrations, fraction of PM1 contributing to PM2.5 mass, and the mass percentage of V-Ni-Fe and ECOC particles contributing to the PM mass. 1

33

Figure 2.2 Fraction of scaled PM2.5 corresponding to PM1 hourly.

34

2.4.1 Mass Concentrations across San Diego County

To verify that the scaled PM values at the SIO Pier were reasonable, the PM2.5

measurements were compared to 3 other sites in San Diego County operated by the San

Diego Air Pollution Control District (SDAPCD). The average PM2.5 measured at the SIO

pier during the study (16 μg/m3) was similar to the other sites (downtown 10 μg/m3,

Escondido 18 μg/m3, and Alpine 16 μg/m3). The downtown San Diego site is the most

comparable site to the SIO Pier due to its close proximity and air mass back trajectory

patterns. Despite 12 miles separating the SIO Pier and downtown sites, similar trends in

PM2.5 mass were observed at both sites during the study (Figure 2.1a). This suggests that

these transport events described herein likely influence even larger portions of San Diego

County.

2.4.2 Identification of Regional Events

During sampling at the SIO pier, time periods with high submicron particle

concentrations were observed using average hourly submicron (0.5-1 m) size-resolved

particle concentrations from an APS. Trends in submicron ATOFMS particle counts are

correlated with submicron APS particle concentrations as shown in Figure 2.1b. The

relationship between the APS and ATOFMS submicron counts shows that, despite

transmission biases from the ATOFMS inlet, temporal variations in submicron ATOFMS

counts directly correspond to changes in atmospheric concentrations [Qin, et al.,

2006;Spencer, et al., 2008]. Multiple time periods showed elevated PM1 mass

concentrations. Figure 2.1c shows PM1 (blue line) versus PM2.5 (red line) mass

concentrations calculated using the APS, as described above. The average PM1

concentration during the study was 7 μg/m3 (dashed gray line); this represents an upper

35

limit because certain time periods with low particle concentrations were excluded from

the analysis as they did not have the requisite statistics for scaling the ATOFMS counts to

mass. Regional events are defined as any period when the peak PM1 mass was at least

twice the average PM1 concentration (14 g/m3). During regional events, PM1 accounted

for 52% of PM2.5 mass and 70% of PM2.5 number concentrations.

2.4.3 Air Mass History during Regional Events

To determine the source(s) of the submicron particles at the SIO pier, HYSPLIT

air mass back trajectories (48 hours) were calculated at 500 meters and patterns were

verified by comparing to back trajectories at 200, 400, 600, 800, and 1000 meters

[Draxler and Rolph, 2003]. Back trajectories calculated for the peak hour of each

submicron concentration spike are shown in Figure 2.3a. Many events had back

trajectories coming from the vicinity of the LA-LB Port region, while a small number of

additional events were more local and passed near the SD Port. These more local events

most likely still had contributions from LA-LB transport that stagnated over the region

for several days. Figure 2.3b shows air mass back trajectories during nonevent periods.

These air masses either came from the west over the ocean or from inland locations but

did not pass near the SD or LA-LB Ports.

Figure 2.3c groups the main air mass back trajectories by type (Port of LA-LB,

San Diego, inland, and oceanic) and includes how frequently each type was observed.

Back trajectories came from the vicinity of the Ports of LA-LB (32% of the sampling

period), the Port of SD (17%), inland (5%), and the ocean (46%). During periods when

the air masses came from the LA-LB Port region, regional events were observed 40% of

the time. Not all air mass back trajectories coming from the LA-LB port region produced

36

high concentration regional events. In order to create these conditions, high levels of

emissions at the port, consistent boundary layer height, relatively short air mass transport

times, and meteorological conditions all have to occur simultaneously. It is important to

note that the amount of ship traffic in the LA-LB Port region remained relatively constant

(25-40 major ships per day) and thus the major factor contributing to the regional events

appears to be proper meteorological conditions [Exchange, 2006]. Based on the

HYSPLIT trajectories, emissions from the Port of SD most likely contributed to two of

the events. When oceanic back trajectories occurred, no regional events were observed;

this indicates that ship traffic in the open ocean west of San Diego, which is much lower

than traffic near the ports, did not contribute significantly to San Diego air quality.

During regional transport periods when trajectories came from the LA-LB Port region, it

is likely additional sources and processes contributed along the way. Also, the very

limited recreational boating in close proximity to the pier and lower ship traffic at the

Port of San Diego (1% of LA-LB) leads to lower emissions compared to the Ports of LA-

LB and helps explain the lack of events during inland or oceanic time periods

[Authorities, 2006].

37

Figure 2.3 HYSPLIT (24 hour) back trajectory analysis at 500 meters for the peak hour of each regional event. Large diamonds represent 24 hours back from the source and small diamonds present 12 hours. A) Trajectories in red pass near the LA Port while trajectories in blue pass over the SD Port. B) Trajectories in purple are representative of non-transport time periods. C) The table included displays the frequency of different air mass types and source regions, as well as the number of regional events observed during each back trajectory type.

38

2.4.4 Particle Chemistry during Regional Transport Events

The most prevalent submicron particle types observed during these regional

events were external mixtures of soot, organic carbon, biomass burning, metals, and sea

salt. Figure 2.4 shows the average mass spectra of the two most common particle types

observed during the regional events. These included an elemental carbon (or soot)

particle type mixed with organic carbon, sulfate, nitrate and water, labeled as

ECOC(sulfate-nitrate). The second type included vanadium-nickel-iron particles mixed

with sulfate, nitrate, carbonaceous species, and water, labeled as V-Ni-Fe(sulfate-nitrate).

The particle labeling scheme used herein first indicates the most dominant peaks in the

positive ion particle spectra, followed by secondary species detected in the negative ion

spectra in parentheses. The positive ion peaks indicate the primary source of the

particles, whereas the negative ions provide insight into the atmospheric aging the

particles have undergone during transport. Notably, the primary source makes up a high

number fraction of particles, but the secondary species make up the bulk of the mass of

the particles. The mass spectra of the ECOC(sulfate-nitrate) or soot particles were

characterized by elemental carbon cluster ions (12C1+, 24C2

+,…, Cn+) with less intense

organic carbon markers (27C2H3+, 29C2H5

+, 37C3H+, and 43C2H3O+) [Spencer and

Prather, 2006]. The V-Ni-Fe(sulfate-nitrate) particle type is characterized by intense

peaks at 51V+ and 67VO+ with less intense iron (56Fe+) and nickel (58,60Ni+) ion peaks. It

should be noted that a small percentage of spectra in the other submicron particle types

showed minor vanadium and vanadium oxide peaks during these events. These mixtures

were most likely formed by coagulation processes, but only represented 2-5% of the total

particle spectra.

39

Figure 2.4 Average negative and positive ion mass spectra for a) ECOC(sulfate-nitrate) and b) V-Ni-Fe(sulfate-nitrate) particles. c) Time series of hourly ECOC(sulfate-nitrate) and V-Ni-Fe(sulfate-nitrate) particle type counts.

40

2.4.5 General submicron particle types

Average mass spectra of each particle type (beyond ECOC(sulfate-nitrate) and V-

Ni-Fe(sulfate-nitrate)) are shown in Figure 2.5 and described below. Elemental carbon

particles were identified by carbon cluster ions (12C1, 24C2,…, Cn) as the dominant peaks

in the positive and negative ion mass spectra, as well as sulfate and nitrate ions (97HSO4-,

62NO3-, and 46NO2

-), markers that are associated with aging [Spencer and Prather, 2006].

Organic carbon particles have intense ion markers at 27C2H3+, 29C2H5

+, 37C3H+, and

43C2H3O+ [Spencer and Prather, 2006]. Biomass burning particles have a dominant

potassium peak (39K+), as well as elemental and organic carbon marker ions. This

particle type also has a significant sulfate ion peak (97HSO4-) and a minor 213K3SO4

+

peak (not shown) which has been linked previously with biomass burning [Qin and

Prather, 2006;Silva, et al., 1999]. Sea salt particles were identified by 23Na+, 62Na2O+,

63Na2OH+, and 81,83Na2Cl+ [Spencer, et al., 2008]. Sea salt with soot particles were

characterized by an intense 23Na+ peak, elemental carbon cluster ions, organic markers,

and other sea salt ion peaks (m/z 62, 63, 81, and 83). This particle type is formed from

coagulation of SS and ECOC particles during transport [Holecek, et al., 2007;Spencer, et

al., 2008].

41

Figure 2.5 Average negative and positive ion mass spectra for the top six particle types (excluding ECOC and V-Ni-Fe which are discussed previously: a) Elemental carbon, b) Organic carbon, c) Biomass Burning, d) Sea salt, e) Sea salt with soot.

42

2.4.6 Sulfate Formation and Aging Processes during Transport

It is important to note that a high percentage (30-80%) of particles did not

produce negative ion mass spectra. Based on optical measurements of single particles

with similar chemical signatures [Moffet, et al., 2008], water has been shown to remain

on the particles which can suppress negative ion formation [Neubauer, et al.,

1997;Neubauer, et al., 1998]. The larger (0.5-1 m) size of these particles also suggests

they have undergone a significant amount of chemical aging during transport. The aging

processes include condensation of secondary species, aqueous phase processing of SO2,

as well as coagulation with pre-existing sulfate marine background particles [Hughes, et

al., 2000].

Due to high SO2 concentrations in the port region coupled with high average

ambient relative humidity (~93%), it is likely that the particles acquired sulfate due to

SO2 oxidative processing when initially formed. In addition, the particles most likely

picked up additional secondary species (sulfate and nitrate), water soluble organic

carbon, and water during transport. This is supported by the fact that for both particle

types, the limited particles with negative ion mass spectra showed a major bisulfate ion

peak (97HSO4), as well as nitrate ion peaks (46NO2 and 62NO3). The observations from

this study, along with a previous study in the southern California coastal region showing

marine background particles with high levels of sulfate, suggest that the ECOC and V-Ni-

Fe particles are likely composed of relatively small primary particle cores mixed with

sulfate, nitrate, and water [Hughes, et al., 2000]. Similar observations in previous studies

of single soot particles, where core sizes of less than 0.2 m were observed [Moffet and

Prather, 2008], show that sulfate and nitrate species contribute the majority of the mass

43

to these aged particles. Previous work has also shown that sulfate is associated with ship

emissions at the source [Agrawal, et al., 2008b;Agrawal, et al., 2008c]. Concurrent filter-

based triple oxygen isotope measurements of sulfate at the SIO Pier estimate that ships

represent 10-44% of non-sea salt sulfate mass concentrations below 1.5 m [Dominguez,

et al., 2008]. These co-located isotope measurements also suggest the ECOC(sulfate-

nitrate) and V-Ni-Fe(sulfate-nitrate) particles acquired significant quantities of sulfate

and water at the point of emission as well as through secondary aging processing during

transport.

2.4.7 Temporal Variations by Particle Type

Figure 2.6 shows the scaled a) relative fractions and b) contributions to PM1 mass

of the different submicron particle types by hour throughout the study. The black line

shows the hourly PM1 mass concentrations (μg/m3). The particle type representing the

majority of submicron mass varies strongly with the concentration of submicron mass.

When PM1 levels were high (>14 μg/m3), soot and V-Ni-Fe particles accounted for 24-

86% (52% average) of submicron mass. Sea salt, the primary background submicron

particle type, represented a significant fraction of submicron particles (75-80%) during

periods when the lowest PM1 mass concentrations were observed and PM1 was the

smallest fraction of PM2.5. These represent marine background periods with high wind

speeds. Biomass burning and organic carbon showed low correlations with PM1 (R2 =

0.09 and R2 = 0.11), suggesting these particle types did not come from the same sources

as most of the submicron particles detected at the pier.

44

Figure 2.6: a) Time series of the relative fractions of submicrometer particles types b) Time series of counts of submicron particle types. The black line on both plots is the scaled PM1 mass concentration (μg/m3).

45

2.4.8 Correlation between Major Particle Types

Metal-containing particles composed of trace amounts of V, Fe, and Ni mixed

with carbonaceous species are primarily emitted from residual fuel combustion processes

and as such have limited sources, including ships and refineries [Agrawal, et al.,

2008a;Reinard, et al., 2007]. Since no refineries exist in San Diego or Orange County

and residual fuel is burned only on large ocean-going vessels, this limits likely sources to

ships from the Ports of LA-LB and Port of SD, as well as refineries near the Ports of LA-

LB. ECOC particles represent a general type produced by most combustion sources,

including diesel truck and automobile emissions [Sodeman, et al., 2005;Toner, et al.,

2008]; however the larger sizes observed (> 0.5 m) and low percentage of particles with

negative ion mass spectra, indicate that the particles were highly aged and not from local

vehicle sources. Figure 2.4c shows the high correlation (R2 = 0.64) observed between the

number concentrations of V-Ni-Fe(sulfate-nitrate) particles and ECOC(sulfate-nitrate)

particles with one hour resolution. Ships burn a variety of fuels (i.e. distillate and

residual) that produce the soot and V-Ni-Fe(sulfate-nitrate) particle types in different

proportions [Agrawal, et al., 2008c;Ault, et al., 2008], that depend on the type of engine,

fuel, and operating conditions. The temporal correlation of the V-Ni-Fe(sulfate-nitrate)

and ECOC(sulfate-nitrate) particle types (Figure 2.4c) and the significant contributions to

PM1 mass (Figure 2.7) during regional events suggests that these two particle types

originate from the same source and/or source region. Figure 2.8 below shows one specific

event where a strong correlation between the mass concentrations of ECOC(sulfate-

nitrate) and V-Ni-Fe(sulfate-nitrate) for a regional event and the number of ship radio

contacts for that period. This occurred during the period with the most rapid (6 hours)

46

transport from the LA-LB Port to the pier. In general, the number of ship counts in the

LA-LB Port region was relatively constant. As described, meteorology, rather than a

change in source emissions, appeared to control the high concentrations and regional

events observed at the SIO Pier during this study.

2.4.9 Size-Resolved Chemistry of Regional Events

The size-resolved chemical mass distributions, calculated using the APS method

discussed above, are shown for the average of regional events (Figure 2.7a) as well as the

average of non-event time periods for comparison (Figure 2.7b). Regional events showed

higher mass concentrations across all submicron sizes (4.7-6.5 µg/m3) when compared to

non-event time periods (2.9-3.8 µg/m3). During regional events, the ECOC(sulfate-

nitrate) particles represent the most significant fraction of submicron mass in each bin

(35-44%), while V-Ni-Fe(sulfate-nitrate) particles contribute 10-19% of particle

submicron mass. Note that this is a unique single particle perspective of the mass

fractions, and it focuses on the mass fraction of the entire particle type. As previously

noted, the majority of the mass of these particles was most likely sulfate, nitrate, and

water, as well as a smaller amount of carbonaceous species from residual oil and

secondary organic carbon. Soot mixed with sea salt, which has been observed in previous

studies during long range transport episodes, showed higher mass concentrations during

the event periods [Holecek, et al., 2007]. These particles were most likely formed by

coagulation processes occurring during the longer transport periods. Non-event time

periods had much lower ECOC(sulfate-nitrate) (4-22%) and V-Ni-Fe(sulfate-nitrate) (2-

10%) mass fractions, showing significantly higher fractions of background particle types.

These background types included more sea salt (16-81%) and biomass burning (9-43%)

47

than sea salt (4-28%) and biomass burning during (11-28%) regional events. During these

periods, PM1 mass represents only 36% of PM2.5 compared with 60% during regional

events. The increase in the overall PM mass and contributions from ECOC(sulfate-

nitrate) and V-Ni-Fe(sulfate-nitrate) during regional events shows the major influence of

transported emissions on San Diego air quality. Notably, these transported aerosols

(ships, refineries, and port vehicle traffic) make much higher contributions to PM

concentrations than local San Diego sources during these events.

48

Figure 2.7: Comparison of the average size-resolved mass distributions of major particle types during regional events and non-events.

49

2.4.10 Correlation between Ships at Sea and Selected Particle Types

Radio transmissions of the location, speed, and heading were recorded in real time

for ships in Southern California waters with an automated identification system (AIS)

antenna in La Jolla, CA and binned as the number of transmissions per hour

[Organization, 2000]. These transmissions are required for ships at sea over 299 gross

tons every minute by the International Maritime Organization (IMO). Figure 2.8 shows

one major regional transport event, based on HYSPLIT back trajectories, from September

12-13 with the black line representing the number of recorded ship positions per hour.

The transport time from the LA-LB port region during this particular event was ~6 hours

and the peak in radioed positions occurred roughly 5-7 hours before the peaks in V-Fe-Ni

and ECOC particles, suggesting ships were a contributor to the increase in particle

concentrations. To observe this strong correlation required many conditions including

favorable meteorology, air mass direction and speed, meteorology favorable to radio

transmissions, and a sharp spike in the number of ships at sea. This correlation occurred

only one time during sampling, but provides further evidence that the LA-LB port region

is a significant source of PM in San Diego.

50

Figure 2.8 Time series of hourly counts of the ECOC and V-Ni-Fe particle types during 6 day period with ship positions recorded per hour.

51

2.4.11 Impact of Regional Events

Regional transport periods resulted in PM1 mass concentrations that were

increased by a factor of 2-4, showing how strongly ships, refineries, vehicles, and other

port combustion sources from the Ports of Los Angeles and Long Beach impact air

quality in the San Diego area. Emissions transported 12 miles north from the Port of San

Diego, while less frequent, most likely mixed with regionally transported emissions and

play a role in lowering the overall air quality of the San Diego region. Particles between

0.5-1 m contribute the most mass to PM1 and PM2.5 for the regional events. The

majority of the mass for these particles is likely a mixture of sulfate, nitrate, and water,

but identifying the core (ECOC or V-Ni-Fe) is critical to identifying the original particle

source and allowing the proper regulations to be established. The adverse health effects

of residual fuel combustion particles, which contain a complex mixture of carbonaceous

and metals, on residents of San Diego county as well as other coastal cities could be quite

significant [Corbett, et al., 2007].

Long-term sampling at a coastal site with speciated aerosol number and mass

concentration measurements will shed further light on the frequency and intensity of

these transport periods and how they vary seasonally. The impact of regionally

transported pollution on receptor site locations beyond San Diego needs to be considered

when evaluating the health impacts of particles in urban and rural areas. These emissions

will become even more important as vehicle emissions become further regulated and

residual oil sources such as shipping continue to grow. The strong influence of regional

transport on San Diego air quality in comparison to the impact of local emissions

52

demonstrates the importance of monitoring and further regulating these sources in the

future.

2.5 Acknowledgements

The authors gratefully acknowledge Dr. Gerardo Dominguez and Prof. Mark

Thiemens for providing the ship tracking data. The authors acknowledge and thank Dr.

Ryan Sullivan, Lindsay Hatch, Margaret Yandell, Elizabeth Fitzgerald, and John Holecek

for their support and assistance with measurements at the SIO Pier. William Brick at the

San Diego Air Pollution Control District is thanked for sharing San Diego County PM2.5

mass concentrations. Funding was provided by the California Air Resources Board under

contracts 04-336 and 05-346. The authors gratefully acknowledge the NOAA Air

Resources Laboratory for the provision of the HYSPLIT transport model and READY

website (http://www.arl.noaa.gov/ready.html) used in this publication.

Chapter 2 is reproduced with permission from A. P. Ault, M. J. Moore, H.

Furutani, and K.A. Prather, Impact of Emissions from the Los Angeles Port Region on

San Diego Air Quality During Regional Transport Events, Environmental Science &

Technology, 43 (10), 3500-3506, 2009. Copyright 2009, American Chemical Society.

53

2.6 References

H. Agrawal, Q. G. J. Malloy, W. A. Welch, J. W. Miller and D. R. Cocker, In-use gaseous and particulate matter emissions from a modern ocean going container vessel, Atmospheric Environment, 2008a.

H. Agrawal, Q. G. J. Malloy, W. A. Welch, J. Wayne Miller and D. R. Cocker III, In-use

gaseous and particulate matter emissions from a modern ocean going container vessel, Atmospheric Environment, 42 (21), 5504, 2008b.


a Crude Oil Tanker at Sea, Environ. Sci. Technol., 2008c. D. T. Allen and J. R. Turner, Transport of atmospheric fine particulate matter: Part 1 -

Findings from recent field programs on the extent of regional transport within North America, Journal Of The Air & Waste Management Association, 58 (2), 254-264, 2008.

A. Ault, C. Gaston, Y. Wang, G. Dominguez, M. H. Thiemens and K. Prather, Single

Particle Mixing State of Individual Ship Plumes at the Port of Los Angeles, Environmental Science & Technology, in prep, 2008.

A. A. o. P. Authorities, 2006 U.S. Port Cargo Tonnage Rankings, American Association

of Port Authorities, 2006. C.A.R.B., Estimated Annual Average Emissions California Air Resources Board, 2006. M. J. Campen, J. P. Nolan, M. C. J. Schladweiler, U. P. Kodavanti, P. A. Evansky, D. L.


K. Capaldo, J. J. Corbett, P. Kasibhatla, P. Fischbeck and S. N. Pandis, Effects of ship

emissions on sulphur cycling and radiative climate forcing over the ocean, Nature, 400 (6746), 743-746, 1999.

J. J. Corbett and H. W. Koehler, Updated emissions from ocean shipping, Journal Of

Geophysical Research-Atmospheres, 108 (D20), 2003. J. J. Corbett, J. J. Winebrake, E. H. Green, P. Kasibhatla, V. Eyring and A. Lauer,

Mortality from ship emissions: A global assessment, Environmental Science & Technology, 41 (24), 8512-8518, 2007.

G. Dominguez, T. Jackson, L. Brothers, B. Barnett, B. Nguyen and M. H. Thiemens,

Discovery and measurement of an isotopically distinct source of sulfate in Earth's

54

atmosphere, Proceedings Of The National Academy Of Sciences Of The United States Of America, 105 (35), 12769-12773, 2008.

R. R. Draxler and G. D. Rolph, HYSPLIT (HYbrid Single Particle Lagrangian Integrated

Trajectory) v. 4.8 Model access via NOAA ARL READY Website NOAA Air Resources Laboratory, pp. (http://www.arl.noaa.gov/ready/hysplit4.html). Silver Spring, MD. , 2003.

D. J. Eatough, N. L. Eatough, M. W. Hill, N. F. Mangelson and L. D. Hansen,

Identification Of Vo2+ In Particles From The Flue Lines Of Oil-Fired Power-Plants, Environmental Science & Technology, 18 (2), 124-126, 1984.

T. M. Exchange, Ship Arrivals-Departures-Transfers Ports of Los Angeles/Long Beach,

The Marine Exchange, San Pedro, CA, 2006. S. C. Fang, E. A. Eisen, J. M. Cavallari, M. A. Mittleman and D. C. Christiani, Acute

changes in vascular function among welders exposed to metal-rich particulate matter, Epidemiology, 19 (2), 217-225, 2008.



J. C. Holecek, M. T. Spencer and K. A. Prather, Analysis of rainwater samples:

Comparison of single particle residues with ambient particle chemistry from the northeast Pacific and Indian oceans, Journal Of Geophysical Research-Atmospheres, 112 (D22), 2007.

S. Hu, A. Polidori, M. Arhami, M. M. Shafer, J. J. Schauer, A. Cho and C. Sioutas,

Redox activity and chemical speciation of size fractioned PM in the communities of the Los Angeles-Long Beach harbor, Atmos. Chem. Phys., 8 (21), 6439, 2008.

L. S. Hughes, J. O. Allen, P. Bhave, M. J. Kleeman, G. R. Cass, D. Y. Liu, D. F.

Fergenson, B. D. Morrical and K. A. Prather, Evolution of atmospheric particles along trajectories crossing the Los Angeles basin, Environmental Science & Technology, 34 (15), 3058-3068, 2000.

International Standards Organization, Petroleum products -- Fuels (class F) --

Specifications of marine fuels, 22 pp., 2005. H. N. Jang, Y. C. Seo, J. H. Lee, K. W. Hwang, J. I. Yoo, C. H. Sok and S. H. Kim,

Formation of fine particles enriched by V and Ni from heavy oil combustion: Anthropogenic sources and drop-tube furnace experiments, Atmospheric Environment, 41 (5), 1053-1063, 2007.

55

E. K. Kauper and P. L. Niemann, Los Angeles to San Diego three dimensional ozone study, ARB Contract Number A-6-090-30, 1977.

R. Lall and G. D. Thurston, Identifying and quantifying transported vs. local sources of

New York City PM2.5 fine particulate matter air pollution, Atmospheric Environment, 40, S333-S346, 2006.

D. Y. Liu, K. A. Prather and S. V. Hering, Variations in the size and chemical

composition of nitrate-containing particles in Riverside, CA, Aerosol Science And Technology, 33 (1-2), 71-86, 2000.

G. Lu, J. R. Brook, M. R. Alfarra, K. Anlauf, W. R. Leaitch, S. Sharma, D. Wang, D. R.

Worsnop and L. Phinney, Identification and characterization of inland ship plumes over Vancouver, BC, Atmospheric Environment, 40 (15), 2767-2782, 2006.

P. T. Martien, R. A. Harley, J. B. Milford and A. G. Russell, Evaluation of incremental

reactivity and its uncertainty in Southern California, Environmental Science & Technology, 37 (8), 1598-1608, 2003.

R. C. Moffet and K. Prather, Aging of soot captured with in situ measurements:

Implications for aerosol climate forcing, Proceedings Of The National Academy Of Sciences Of The United States Of America, Submitted, 2008.


segregated optical and microphysical properties of ambient aerosols measured in a single-particle mass spectrometer, Journal Of Geophysical Research-Atmospheres, 113 (D12), 2008.

K. R. Neubauer, M. V. Johnston and A. S. Wexler, On-line analysis of aqueous aerosols

by laser desorption ionization, International Journal Of Mass Spectrometry And Ion Processes, 163 (1-2), 29-37, 1997.

K. R. Neubauer, M. V. Johnston and A. S. Wexler, Humidity effects on the mass spectra

of single aerosol particles, Atmospheric Environment, 32 (14-15), 2521-2529, 1998.

I. M. Organization, International Convention on Safety of Life at Sea (SOLAS) Chapter

V, 2000. S. Pattanaik, F. E. Huggins, G. P. Huffman, W. P. Linak and C. A. Miller, XAFS studies

of nickel and sulfur speciation in residual oil fly-ash particulate matters (ROFA PM), Environmental Science & Technology, 41 (4), 1104-1110, 2007.

56

K. E. Pinkerton, Y. M. Zhou, S. V. Teague, J. L. Peake, R. C. Walther, I. M. Kennedy, V. J. Leppert and A. E. Aust, Reduced lung cell proliferation following short-term exposure to ultrafine soot and iron particles in neonatal rats: Key to impaired lung growth?, Inhalation Toxicology, 16, 73-81, 2004.

X. Y. Qin, Characterization of Ambient Aerosol Composition and Formation

Mechanisms and Development of Quantification Methodologies Utilizing ATOFMS, Doctoral Thesis thesis, University of California San Diego, La Jolla, 2007.

X. Y. Qin, P. V. Bhave and K. A. Prather, Comparison of two methods for obtaining

quantitative mass concentrations from aerosol time-of-flight mass spectrometry measurements, Analytical Chemistry, 78 (17), 6169-6178, 2006.

X. Y. Qin and K. A. Prather, Impact of biomass emissions on particle chemistry during

the California Regional Particulate Air Quality Study, International Journal Of Mass Spectrometry, 258 (1-3), 142-150, 2006.

M. S. Reinard, K. Adoua, J. M. Martini and M. V. Johnston, Source characterization and

identification by real-time single particle mass spectrometry, Atmospheric Environment, 41 (40), 9397-9409, 2007.





X. H. Song, P. K. Hopke, D. P. Fergenson and K. A. Prather, Classification of single

particles analyzed by ATOFMS using an artificial neural network, ART-2A, Analytical Chemistry, 71 (4), 860-865, 1999.


Resolved Chemical Composition of Aerosol Particles During a Monsoonal Transition Period over the Indian Ocean, Journal Of Geophysical Research-Atmospheres, 2008.

M. T. Spencer and K. A. Prather, Using ATOFMS to determine OC/EC mass fractions in

particles, Aerosol Science And Technology, 40 (8), 585-594, 2006.

57

M. T. Spencer, L. G. Shields and K. A. Prather, Simultaneous measurement of the effective density and chemical composition of ambient aerosol particles, Environmental Science & Technology, 41 (4), 1303-1309, 2007.

S. M. Toner, Anthropogenic Particulate Source Characterization and Source

Apportionment Using Aerosol Time-of-Flight Mass Spectrometry, Doctoral Thesis thesis, University of California San Diego, La Jolla, 2007.

S. M. Toner, L. G. Shields, D. A. Sodeman and K. A. Prather, Using mass spectral source

signatures to apportion exhaust particles from gasoline and diesel powered vehicles in a freeway study using UF-ATOFMS, Atmospheric Environment, 42 (3), 568-581, 2008.

M. Vallius, T. Lanki, P. Tiittanen, K. Koistinen, J. Ruuskanen and J. Pekkanen, Source

apportionment of urban ambient PM2.5 in two successive measurement campaigns in Helsinki, Finland, Atmospheric Environment, 37 (5), 615-623, 2003.

S. Vutukuru and D. Dabdub, Modeling the effects of ship emissions on coastal air

quality: A case study of southern California, Atmospheric Environment, 42 (16), 3751-3764, 2008.

Z. Q. Xie, L. G. Sun, J. D. Blum, Y. Y. Huang and W. He, Summertime aerosol chemical

components in the marine boundary layer of the Arctic Ocean, Journal Of Geophysical Research-Atmospheres, 111 (D10), 2006.

Y. M. Zhou, C. Y. Zhong, I. M. Kennedy, V. J. Leppert and K. E. Pinkerton, Oxidative

stress and NF kappa B activation in the lungs of rats: a synergistic interaction between soot and iron particles, Toxicology And Applied Pharmacology, 190 (2), 157-169, 2003.

3 Characterization of the Single Particle Mixing State of

Individual Ship Plume Events Measured at the Port of Los

Angeles

3.1 Synopsis

Ship emissions contribute significantly to gaseous and particulate pollution

worldwide. To better understand the impact of ship emissions on air quality,

measurements of the size-resolved chemistry of individual particles in ship emissions

were made at the Port of Los Angeles using real-time, single-particle mass spectrometry.

Ship plumes were identified through a combination of ship position information and

measurements of gases and aerosol particles at a site 500 meters from the center of the

main shipping channel of the Port of Los Angeles. Single particles containing mixtures of

organic carbon, vanadium, and sulfate (OC-V-sulfate) resulted from residual fuel

combustion (i.e. bunker fuel), whereas high quantities of fresh soot particles (when OC-

V-sulfate particles were not present) represented distinct markers for plumes from

distillate fuel combustion (i.e. diesel fuel) from ships as well as trucks in the port area.

OC-V-sulfate particles from residual fuel combustion contained significantly higher

levels of sulfate and sulfuric acid than plume particles containing no vanadium. These

associations may be due to vanadium (or other metals such as iron) in the fuel catalyzing

the oxidation of SO2 to produce sulfate and sulfuric acid on these particles. Enhanced

sulfate production on OC-V-sulfate ship emission particles would help explain some of

the higher than expected sulfate levels measured in California compared to models based

on emissions inventories and typical sulfate production pathways. Understanding the

58

59

overall impact of ship emissions is critical for controlling regional air quality in the many

populated coastal regions of the world.

3.2 Introduction

Ship emissions represent one of the least regulated forms of anthropogenic

pollution due, in large part, to the challenges involved in establishing international

policies. Ship emissions impact climate by initiating cloud formation and altering the

Earth’s radiation budget [De Bock, et al., 2000;Kim, et al., 2008]. Ships emit high

concentrations of soot and heavy metals (i.e. vanadium and nickel) [Agrawal, et al.,

2008a;Isakson, et al., 2001], in addition to an estimated 2.4 tons of SO2 globally,

producing up to 10% of sulfate mass globally through atmospheric reactions [Eyring, et

al., 2005]. Over the next century, SO2 levels are predicted to rise significantly as global

commerce expands [Eyring, et al., 2005]. Ship emissions have been shown to have

negative effects on human health through exposure studies and epidemiological models

[Campen, et al., 2001;Corbett, et al., 2007;South Coast Air Quality Management District,

2008]. For example, inhaled vanadium particles are toxic and synergistic effects with

nickel and sulfate have been shown to enhance toxicity [Campen, et al., 2001]. To

develop effective strategies for reducing the impact of shipping on climate and health,

recent studies have focused on improving ship emission inventories, which are used to

estimate future scenarios for gas phase concentrations [Eyring, et al., 2009]. Efforts to

regulate shipping emissions have been difficult due to fuel and upgrade costs and

international dependence on foreign trade [Silverman, 2008;Wang, et al., 2007].

Particulate emissions from ships have been characterized by multiple analytical

methods including energy dispersive x-ray fluorescence measurements of particulate

60

matter on filters [Isakson, et al., 2001], ion chromatography (IC) of particulate matter

extracted from filters [Agrawal, et al., 2008a], real-time mass-based measurements [Lack,

et al., 2009;Murphy, et al., 2009], particle counters [Sinha, et al., 2003], cloud

condensation nuclei measurements [Murphy, et al., 2009], and size distribution

measurements [Isakson, et al., 2001]. These measurements have led to updated emission

inventories and emission factors for particle mass and particle number for ship emissions

[Lack, et al., 2009;Murphy, et al., 2009]. Real-time, single-particle mass spectrometry

has also been used to identify ship emissions at locations both near source regions and

after transport [Ault, et al., 2009;Healy, et al., 2009;Reinard, et al., 2007]. Herein, we

report in situ measurements of the size-resolved chemical mixing state of particles in

individual fresh ship plumes measured at the Port of Los Angeles.

3.3 Experimental

3.3.1 Sampling Information

Ambient air sampling was conducted from November 16–26, 2007 at the Port of

Los Angeles (LA) on Terminal Island in San Pedro, California. Particles were sampled

through a four meter sampling mast, seven meters above the ground, at a sampling site

500 meters east of the center of the main channel. Wind direction and speed were

measured using a R.M. Young wind monitor. Winds exhibited a consistent diurnal

pattern with southerly sea breezes during the day and light nocturnal land breezes from

the north. Radio transmissions of the location, speed, and heading of ships entering and

exiting the port were recorded in real-time with an automated identification system (AIS)

antenna in La Jolla, CA [Organization, 2000]. These transmission signals are required to

61

be sent every minute for ships at sea over 299 gross tons by the International Maritime

Organization.

3.3.2 Gas and Particle Peripheral Instrumentation

Gas phase measurements were made of NOx (Thermo Environmental Instruments

(TEI) Model 42C), O3 (TEI Model 49), and SO2 (TEI Model 43). Particle size

distributions were measured with a scanning mobility particle sizer (SMPS, TSI Model

3936L) operating from 10 – 600 nm at a time resolution of 5 minutes; black carbon (BC)

concentrations (ng/m3) were measured using an aethalometer (Magee Scientific AE31).

3.3.3 Aerosol Time-of-Flight Mass Spectrometry (ATOFMS)

The design and details of the ultrafine (UF)-ATOFMS used in this study are

described in detail elsewhere [Su, et al., 2004]. This instrument measured the size and

chemical composition of individual particles between 100-1000 nm during this study.

Briefly, particles are introduced into the UF-ATOFMS through an aerodynamic lens into

a differentially pumped vacuum chamber where the particles are accelerated to a size-

dependent velocity. Particles pass through two 532 nm continuous wave lasers located 6

cm apart. Particle speed is used to determine vacuum aerodynamic diameter by

calibration with polystyrene latex spheres of known size. Sized particles are individually

desorbed and ionized in the mass spectrometer source region by a 266 nm Q-switched

Nd:YAG laser (1.2-1.4 mJ). Positive and negative ions produced from individual

particles are detected using a dual-reflectron time-of-flight mass spectrometer.

62

3.3.4 Single Particle Analysis

During sampling at the Port of LA, the aerodynamic size and chemical

composition of 1,245,041 particles were analyzed using an UF-ATOFMS. Size and mass

spectral information were imported into MatLab 6.5.1 (The Mathworks Inc.) and

analyzed utilizing YAADA 1.2 (www.yaada.org). Individual particles were analyzed via

two methods: 1) mass spectral ion intensities, aerodynamic size, and temporal

information and 2) clustering via an adaptive resonance theory based neural network

algorithm (ART-2a) at a vigilance factor of 0.8 [Song, et al., 1999]. ART-2a classifies

individual particles into separate clusters based on the presence and intensity of ion peaks

in individual single-particle mass spectra. Peak identifications within this paper

correspond to the most probable ions for a given mass/charge (m/z) ratio. General particle

types are defined by their characteristic chemical species in an attempt to simplify the

naming scheme; these labels do not reflect all of the species present within a specific

particle class.


3.4.1 Ship Identification

Characteristics of each positively identified ship plume over a 10 day period are

listed in Table 3.1 which includes: the peak time of the observed plume, length of time

the plume was observed, calculated transport time from the point of emission to the peak

of the plume at the sampling location using wind speed, peak particle number

concentration, characteristic particle type, number of particles chemically characterized

by the UF-ATOFMS, and number fractions of OC-V-sulfate and fresh soot particles

present during the plume detection period. The times when different ships passed the

63

sampling location were determined through a combination of arrival and departure logs

[Exchange, 2007], AIS ship position recordings, and pictorial documentation. As an

example, Figure 3.1 shows the departure of the vessel Container1b on November 19,

2007 at 03:30 (PST). Red markers (1 minute resolution) represent the ship’s position as

recorded by the AIS, and the green diamond represents the sampling location. The blue

line shows the path of the plume (550 meters) to the sampling location based on

measured wind direction (292°) and speed (0.6 m/s). Ships that could be positively

identified by time, position, and wind speed/direction are analyzed in detail; however,

numerous ships and plumes passed the site during the sampling period that could not be

confidently assigned using the criteria listed above. Due to changing wind direction and

wind speed, calculated plume travel times to the sampling location varied from 5-20

minutes.

64

Table 3.1 Characteristics of ship plumes sampled including: Ship plume, plume peak time, plume duration, plume age, peak number concentration, identifying ATOFMS particle type, number of ATOFMS particles, number fraction of OC-V-sulfate particles, and number fraction of fresh soot particles.

65

Figure 3.1 Map showing the Port of LA. The green diamond represents the sampling location and the red markers and line represent the minute-resolved positions and course of a container ship as it departed from the Port of LA at 03:30 on November 19, 2007. The blue line represents the average wind direction during transport of the plume to the sampling location.

66

3.4.2 Plume Characterization

Characteristics of a representative plume observed on November 19, 2007 at

03:30 from the vessel Container1b are shown in Figure 3.2. The top portion shows SO2,

NO, NO2, and BC concentrations increasing sharply from background conditions at the

onset of the plume and O3 decreasing due to its reaction with NO forming NO2 [Isakson,

et al., 2001;Sinha, et al., 2003]. Gas phase data were used to qualitatively identify the

presence of ship plumes. The duration of this plume at the sampling site was ~15

minutes, which is similar to previously observed timescales (13.6 min) [Isakson, et al.,

2001]. Similar trends for gas phase species were observed in most plumes; Table 3.2 lists

SO2 and NOx concentrations during different plume events and provides a detailed

discussion of these events. The bottom portion of Figure 3.2 shows the temporal

evolution of the size-resolved particle number concentrations with 5 minute resolution

during the plume sampling period. A rapid increase in the total number concentrations of

10-600 nm particles (white line) occurred as the plume passed the sampling location,

doubling background levels from 5000/cm3 to 11,000/cm3. The mode and shape of the

size distribution did not shift significantly while sampling this plume. In other plumes,

changes in the size distribution were also observed.

67

Figure 3.2 Identification of the plume of Container1b. Gas phase measurements included SO2 (orange), NO (green), NO2 (red), and O3 (blue). Particle phase measurements were of black carbon (black), particle number concentration (white), and size-resolved number concentrations over time (color matrix).

68

3.4.3 In Plume Gas Phase Chemistry and Concentrations

The expected loss of O3 shown in Figure 3.2 theoretically should have a 1:1 ratio

to the gain in NOx (assuming NOx is primarily NO in a fresh plume), but this was not

observed. A number of factors may have contributed to this including: additional species

reacting with NO and O3 and the presence of aqueous particles and droplets that could

have scavenged gas phase species. The maximum concentrations of SO2, NOx, and O3

during the different ship plumes are given in Table 3.2, along with background

concentrations and the net concentration increase. The ratio of NOx/SO2 is also given

which can be an indicator of fuel type. Residual fuels have higher concentrations of

sulfur and thus produce higher concentrations of SO2, while distillate fuels have less

sulfur and produce less SO2 [International Standards Organization, 2005]. Changes in

the gas phase data to lower NOx/SO2 (ppbv/ppbv) ratios presented here correspond to

changes in particle chemistry in the plumes with higher levels of OC-V-sulfate particle

levels. In contrast, fresh soot plumes have higher NOx/SO2 ratios.

69

Table 3.2 Gas phase concentrations (max, baseline, and net change) of SO2, NOx, and Ozone during different plume events are shown along with the NOx/SO2 ratio.

NOx/SO2

Max Baseline Peak Max Baseline Peak Min Baseline Peak Ratio

Container1a 7.2 3.5 3.7 26.2 7.3 18.9 30.0 55.1 25.1 5.1

Container2 7.7 3.4 4.3 16.7 7.9 8.8 19.7 54.9 35.2 2.0

Container3 11.2 5.0 6.2 19.1 10.7 8.4 45.0 68.0 23.0 1.4

Container1b 12.5 7.4 5.1 18.9 7.7 11.2 14.0 45.8 31.8 2.2

Tanker1a* 47.7 18.5 29.2 160.0 58.0 102.0 No Min No Min N/A 3.5

Container4 10.7 4.1 6.6 39.5 14.9 24.6 27.8 57.1 29.3 3.7

Tanker1b 6.3 3.2 3.1 No Peak No Peak N/A No Min No Min N/A N/A

Container5 4.1 3.1 1.0 38.0 7.1 30.9 37.4 45.4 8.0 30.9

Tanker2 12.0 6.7 5.3 32.4 10.1 22.3 8.7 38.6 29.9 4.2

Container6 12.0 8.6 3.4 41.6 15.8 25.8 23.7 44.8 21.1 7.6

CruiseShip1 10.8 6.6 4.2 52.3 11.7 40.6 26.0 46.4 20.4 9.7

Container7 9.0 4.7 4.3 13.2 10.5 2.7 23.0 35.1 12.1 0.6

* night-time chemistry altered by land breeze in Long Beach leading to high NOx & SO2 levels, but low O3

SO2 (ppbv) NOx (ppbv) O3 (ppbv)

70

3.4.4 Unique Plume Chemistry

Two types of ship plumes were observed in this study based on unique chemical

characteristics: one with a large number fraction of internally mixed organic carbon,

vanadium, and sulfate (OC-V-sulfate) particles and a second with a large number fraction

of elemental carbon (EC) or soot particles with mass spectral signatures characteristic of

fresh EC emissions [Moffet and Prather, 2009]. The average mass spectrum of the OC-

V-sulfate particle type is shown in Figure 3.3a, showing intense vanadium peaks at 51V+

and 67VO+ and organic carbon markers including 27C2H3+, 29C H2 5

+, 37C H3+, and

43C H O2 3+, as well as a strong bisulfate ion signal (97HSO4

-). Notably, a sulfuric acid

cluster peak (195H SO •HSO2 4 4-) was observed on ~80% of these OC-V-sulfate particles.

In these plumes, the OC-V-sulfate type represented 10-34% of particles sampled in the

100-500 nm size range.

71

Figure 3.3 Average negative and positive ion mass spectra for the OC-V-sulfate (a) and fresh soot (b) particle types.

72

The average mass spectrum of the soot (or elemental carbon) particle type is

shown in Figure 3.3b, characterized by both positive and negative carbon cluster ions

(e.g., 12C1+, 24C2

+,…, Cn+). These particles did not contain significant sulfate or nitrate

ion markers, likely due to limited time for atmospheric aging between the point of

emission and sampling (~68% of fresh soot particles had no sulfate or nitrate marker).

This signature is typical of fresh emissions and has been detected during source tests of

cars burning gasoline and trucks burning distillate (i.e. diesel) fuel [Sodeman, et al.,

2005;Toner, et al., 2006]. The characteristic mass spectrum contrasts most background

particle types observed at the Port of LA, including other elemental carbon particles,

which contained significant amounts of nitrate and/or sulfate, as discussed below.

Detailed information on mass spectral variability within the OC-V-sulfate and fresh soot

types is included below. While other particle types were present within these plumes, the

OC-V-sulfate and fresh soot particle types were observed as characteristic source markers

for the two different plume types. Plumes were labeled as OC-V-sulfate and/or fresh soot

when each particle type represented greater than 5% of particles sampled during the

plume event.

Residual fuels used in shipping contain much higher concentrations of vanadium

and other metals than distillate fuels [International Standards Organization, 2005].

Previous ATOFMS source studies of cars and trucks show <1% of particles mixed with

vanadium [Sodeman, et al., 2005] and have noted the correlation of vanadium and sulfur

in ambient single particles and attributed these particles to fuel oil combustion [Lake, et

al., 2004]. The distinct OC-V-sulfate and fresh soot plumes likely result from the

73

combustion of different fuels, with residual fuel oil producing the OC-V-sulfate plumes

and distillate fuel forming the fresh soot plumes. The soot plumes also had elevated

levels of Ca-ECOC particles. The Ca-ECOC particle type has been linked to distillate

(i.e. diesel) fuel in previous ATOFMS source characterization studies [Toner, et al.,

2006]; the mass spectrum is shown in Figure 3.6 The presence of Ca-ECOC particles

agrees with filter measurements in a marine distillate exhaust study, which found calcium

to be ~25% of the relative weighted emission factor for trace elements compared to <5%

in residual fuel combustion exhaust [Agrawal, et al., 2008b]. While the average ratio of

NOx to SO2 (ppbv/ppbv) during the study was 3.7, it was lower in the OC-V-sulfate

plumes (average 2.6, range: 1.4-4.2) and higher in the soot plumes (average 11.6, range:

0.6-30.9). This suggests a relative enrichment of SO2 in the V-plumes as would be

expected when burning residual fuels with high sulfur content (max 3.5 % (m/m))

[International Standards Organization, 2005]. A study comparing an auxiliary engine

burning distillate fuel and a main engine burning residual fuel found significantly higher

vanadium levels associated with the residual fuel combustion and higher levels of

calcium with the distillate fuel combustion, which agrees with this study as well as

previous diesel source testing studies using the UF-ATOFMS [Agrawal, et al.,

2008b;Toner, et al., 2006]. In addition, the amount of sulfate measured by ion

chromatography (IC) during the study by Cocker et al. 2008 does not change significantly

with ship engine load for residual or distillate fuels [Agrawal, et al., 2008b]. This

supports the assignment of OC-V-sulfate particles to residual fuel combustion plumes

and the lack of OC-V-sulfate to distillate fuel combustion plumes [Agrawal, et al.,

2008a].

74

3.4.5 Details on the OC-V-sulfate and Fresh Soot Particle Types

The digital color stack (DCS) for the OC-V-sulfate particle type is shown in

Figure 3.4. The x-axis is mass-to-charge and the y-axis shows the fraction of particles

within this type containing a specific peak. The color represents the fraction of particles

with different peak areas; a peak area > 500 (arbitrary units) was used to exclude noise in

the spectra. The ion peaks with the largest fraction and highest intensity in the positive

spectrum are the mass to charges corresponding to vanadium (51V+ and 67VO+) and

organic carbon (27C2H3+, 36C3

+, 37C3H+, 41C3H5+, 43C3H3O+, etc.). These peaks are on

nearly 100% of the particles in this type. In the negative spectrum the most intense peak

is sulfate (97HSO4-) and over 95% of these particles have a peak area above 20,000.

These particles also have sulfuric acid (195H2SO4HSO4-) on ~80% of the particles, which

is even more striking when the transmission efficiency of ~49% at m/z 200 for the co-

axial ATOFMS is considered [Pratt, et al., 2009b]. It should be noted that the intensity of

the sulfate peak led to considerable ringing, which is seen by the continuum of lower

intensity peaks between m/z -98 to -110.

75

Figure 3.4 A digital color stack for the OC-V-sulfate particle type showing the fraction of particles containing a specific peak on the y-axis versus mass-to-charge on the x-axis. The color represents the fraction containing a specific range of peak areas, with 500 being the lower peak area cutoff. The top portion of the figure represents the positive mass spectrum and the bottom portion represents the negative mass spectrum.

76

Figure 3.5 A digital color stack for the fresh soot particle type showing the fraction of particles containing a specific peak on the y-axis versus mass-to-charge on the x-axis. The color represents the fraction containing a specific range of peak areas, with 500 being the lower peak area cutoff. The top portion of the figure represents the positive mass spectrum and the bottom portion represents the negative mass spectrum.

77

The digital color stack for the fresh soot particle type is shown in Figure 3.5, with

the same thresholds and range used in Figure 3.4. For Figure 3.5, intense peaks are

observed at mass-to-charges corresponding to carbon clusters (12C+, 36C3+, 48C4

+, …,

144C12+) in the positive plot (top). The negative (bottom) plot also shows a similar pattern

with respect to the carbon clusters (24C2-, 36C3

-, 48C4-, …, 120C10

-). What is notable about

the fresh soot type is that < 10% of the particles have a sulfate peak and 0% have a

sulfuric acid peak, demonstrating the difference between the two particle types produced

in different plumes.

3.4.6 Description of the Ca-ECOC Particle Type

Figure 3.6 shows the average mass spectrum of the Ca-ECOC particle type, which

is characterized by an intense calcium ion peak (40Ca+) and elemental carbon clusters in

both the positive (12C1+, 24C2

+,…, Cn+) and negative spectra (12C1

-, 24C2-,…, Cn

-). The

significance of this particle type is that it is enhanced by number along with the fresh soot

particle type in distillate fuel plumes. Previous studies have shown higher calcium mass

fractions in filter measurements from the stacks of ships burning distillate fuel compared

to residual fuel [Agrawal, et al., 2008a;Agrawal, et al., 2008b].

78

Figure 3.6 Average mass spectrum of the Ca-ECOC particle type from a fresh soot plume.

79

3.4.7 Background Particle Types During Plumes

Figures 3.7 and 3.8 show average mass spectra for the background particle types

measured before and after the plumes from Container1b and Container4. The mass

spectra of aged soot particles (Figures 3.7a and 3.8a) were characterized by elemental

carbon cluster ions (12C1+, 24C2

+,…, Cn+) with less intense organic carbon markers

(27C2H3+, 29C2H5

+, 37C3H+, and 43C2H3O+) [Spencer and Prather, 2006] and secondary

markers for sulfate (97HSO4-) and nitrate (62NO3

-). The biomass burning particle type is

shown in Figures 3.7b and 3.7b and is characterized by a dominant potassium ion peak

(39K+) with less intense elemental carbon markers (12C1+, 24C2

+,…, Cn+) and organic

carbon markers 27C2H3+, 29C2H5

+, 37C3H+, and 43C2H3O+ [Spencer and Prather, 2006].

The mass spectra of the V-background particle type (Figures 3.7c and 3.7c) are

characterized by intense peaks at 51V+ and 67VO+ with less intense iron (56Fe+) and nickel

(58,60Ni+) ion peaks. Note the lack of peaks (other than detector crosstalk) in the negative

spectrum which is commonly observed after particles undergo aging and take up

significant amounts of water, leading to negative ion suppression [Neubauer, et al.,

1997]. These particles have been shown in other marine environments and are likely

highly aged [Ault, et al., 2009]. Although they share a strong vanadium signal, the

temporal trend of the V-background particles did not correlate with those of the freshly

emitted OC-V-sulfate particles.

80

Figure 3.7 Average mass spectra of background particle types before and after the plume from Container1a.

81

Figure 3.8 Average mass spectra of background particle types from before and after the plume from Container4. CT represents crosstalk on the detector.

82

3.4.8 Temporal Trends

Figure 3.9 shows the time series of the number of OC-V-sulfate (Figure 3.9a) and

fresh soot particles (Figure 3.9b) analyzed by the UF-ATOFMS for the first 5 days of the

study. After this period, Santa Ana winds disrupted normal wind patterns, making the

presence of clear ship plume impacts less apparent. Both time series are characterized by

spikes lasting 5-20 minutes; however, as shown in Figure 3.9c, OC-V-sulfate and fresh

soot spikes frequently occur at different times. Over 65% of the spikes in OC-V-sulfate

particles could be correlated to specific ships. Only 15% of the fresh soot spikes

corresponded to specific ships, which is expected given that fresh soot is a more

ubiquitous particle type associated with many distillate fuel combustion sources including

trucks and tug boats [Lack, et al., 2008]. Some OC-V-sulfate plumes contained

significant fresh soot, while others did not, suggesting that the varying amount of fresh

soot in residual fuel combustion plumes may be due to different plume dynamics, engine

type and condition, operating conditions, as well as background levels and other emission

sources.

83

Figure 3.9 Time series of the differential counting rate (number of ATOFMS counts per 2 minutes) for the OC-V-sulfate and fresh soot particle types. Peaks that could be correlated with ships are labeled; plumes with asterisks were only tentatively identified.

84

3.4.9 Size-resolved Chemistry

The chemical composition as a function of aerodynamic diameter for the urban

background is shown before and after a residual combustion plume (Figure 3.10a) and a

distillate combustion plume (Figure 3.10b). This background is defined as 10 minutes

before the plume was detected until 10 minutes after the plume disappeared with the

plume itself subtracted out. The particle types associated with the background urban

aerosol included aged soot, potassium mixed with elemental carbon, and a background

vanadium type, as described in detail above. The background-specific particle types did

not exhibit a temporal pattern similar to the OC-V-sulfate or fresh soot types, suggesting

that they were not associated with the freshly emitted ship plumes.

Figures 3.10c & 3.10d show the size-resolved chemical composition for specific

ship plumes representing an OC-V-sulfate plume (Container4) and a fresh soot plume

(Container1a), respectively. The OC-V-sulfate particles in the residual fuel combustion

plume increased 10-25 times in less than 2 minutes from a nearly negligible background

level before the plume arrived (Figure 3.10a) to accounting for 28% of 150-500 nm

particles (Figure 3.10c). While the UF-ATOFMS was unable to chemically characterize

<100 nm particles, measurements of residual fuel combustion exhaust particles using

transmission electron microscopy with energy dispersive x-ray spectroscopy have

detected distinct peaks of sulfur and vanadium within particles between 30-100 nm in

diameter [Murphy, et al., 2009]. Thus, it is likely that the OC-V-sulfate particle type also

represents a significant fraction of the ultrafine particles emitted in residual fuel

combustion exhaust, especially as ultrafine particle concentrations measured by the

SMPS increased during plumes.

85

For the soot plume (Container1a), a unique mode was observed between 100 –

200 nm composed of fresh soot and Ca-ECOC particles. This differs from the majority of

the study when the center of the particle size mode was ~ 270 nm due to the combination

of inlet transmission efficiencies and ambient concentrations [Pratt, et al., 2009b], also

shown in Figure 3.10a, 3.10b, and 3.10c. However, during the fresh soot plume, particles

with small aerodynamic diameters (100-200 nm) and large geometric diameters (resulting

in increased scattering signals), were observed. Previous ATOFMS studies measuring

particle optical properties have shown that these 100-200 nm particles were non-

spherical, fractal agglomerates, typical of fresh soot [Moffet and Prather, 2008]. Also,

previous tandem measurements using a differential mobility analyzer in line with an

ATOFMS showed fresh soot particles having small aerodynamic diameters relative to

their geometric diameters for fractal particles with a low effective density [Spencer, et al.,

2007].

86

Figure 3.10 Size-resolved number fractions of the different particle types for the background before and after a) OC-V-sulfate (Container4) and b) fresh soot (Container1a) plumes. (c,d) Size-resolved chemical fractions of these plumes.

87

3.4.10 Correlation Between Sulfate and Vanadium Particles

Increasing sulfate levels represent a major concern globally [Kim, et al., 2008]

and have been shown to influence ship tracks [De Bock, et al., 2000]. For each of the

residual fuel combustion plumes with high fractions of OC-V-sulfate particles, the

number fraction of each particle type containing sulfate or sulfuric acid peaks were

determined, as well as the average intensity of each peak, which is related to the mass of

the species present [Bhave, et al., 2002;Pratt, et al., 2009a]. The OC-V-sulfate particles

within the residual fuel combustion plumes were observed to contain high concentrations

of sulfate and sulfuric acid, particularly when compared to other particle types both in

plumes and during background conditions. A representative OC-V-sulfate plume from

the ship labeled Container4 is described; four other plumes are described later in the text.

Within the Container4 plume, high number fractions of OC-V-sulfate particles contained

sulfate (100%) and sulfuric acid (56%) (Figure 3.11). In comparison, only 14-40% by

number of the remaining particle types within the plume contained sulfate, and almost

none (0-1%) contained sulfuric acid. In addition, average absolute peak areas

corresponding to sulfate and sulfuric acid were up to ~760 times greater and 7,000 times

greater, respectively, for the OC-V-sulfate particle type compared to the other particle

types within the plume (Figure 3.11). The measured sulfate and sulfuric acid peak areas

are also larger than previously measured in ATOFMS source combustion studies,

including cars burning gasoline [Sodeman, et al., 2005], trucks burning diesel [Toner, et

al., 2006], and biomass burning [Silva, et al., 1999].

88

Figure 3.11 Average sulfate (red circles) and sulfuric acid (blue circles) absolute peak areas for major particle types during the plume of the vessel Container4; errors are shown by 95% confidence intervals. Number fractions of particles containing sulfate (red diamond) and sulfuric acid (blue diamond) are shown. The inset shows a zoomed-in version of particle types with lower peak areas than the OC-V-sulfate type.

89

Several mechanisms could lead to the high measured levels of sulfate in the OC-

V-sulfate particles in the plume: 1) homogeneous formation of sulfuric acid in the gas

phase followed by condensation onto particles upon reaction with gas phase ammonia, 2)

heterogeneous production of sulfate and sulfuric acid on the surface of particles, and 3)

aqueous phase production within water-containing particles. It should be noted that other

possibilities exist that could contribute to the enhanced sulfate/sulfuric acid concentration

on OC-V-sulfate particles. Further laboratory and source testing is needed to confirm

whether the proposed hypothesis is actually occurring and quantify the concentrations of

sulfate and sulfuric acid that might be produced by these pathways. Of these mechanisms,

the gas phase process is too slow to explain the high levels of sulfate and sulfuric acid

formed on the particles within minutes of emission [Finlayson-Pitts and Pitts, 2000].

This is supported by a study showing that at high relative humidity, aqueous processing is

the main oxidation pathway [McMurry and Wilson, 1983]. Also, if sulfuric acid was

produced in the gas phase, it would likely react with ammonia and condense on all

particle surfaces present, thus forming sulfate on all particle types, not selectively on one

particle type (OC-V-sulfate) as observed in this study. Ammonium was observed on 31%

of all particles over the course of the study, however only 13% of OC-V-sulfate particles

during residual fuel combustion plumes had ammonium present and the ammonium ion

signal was quite small compared to the sulfuric acid signal. This indicates that sulfate is

being introduced into these particles via a different pathway, and we hypothesize that

given the rapid formation of sulfate on a selective subset of particles the most likely

routes are aqueous phase oxidation of SO2 to sulfate and sulfuric acid or heterogeneous

reactions. The strong association between sulfate and sulfuric acid on OC-V-sulfate

90

particles further suggests the metal(s) present in the residual fuel could catalyze this

process [Barbaray, et al., 1978]. Vanadium, one of the most abundant trace metals in

bunker fuel [International Standards Organization, 2005], could be catalyzing the

oxidation process of SO2 in the presence of H2O and NO2 [Barbaray, et al., 1978].

Alternatively, vanadium could be acting as a proxy for iron or other catalytically active

metals to which the UF-ATOFMS is less sensitive, due to differences in ionization

energy (i.e. Fe - 7.90 eV vs. V - 6.75 eV) [Lide, 2009]. This possibility could be

important since Fe is one of the most effective catalysts of sulfur (IV) in aqueous solution

[Brandt and Vaneldik, 1995]. V2O5 was originally considered to be ineffective as an

atmospheric catalyst below 150 °C [Barbaray, et al., 1977], but subsequent work has

shown that in the presence of NO2 and adsorbed H2O, catalysis of SO2 by V2O5 can

occur down to ambient temperatures (25 °C) [Barbaray, et al., 1978]. More recent

research involving vanadium catalysis through laboratory and field measurements has

also shown the potential for vanadium catalysis in the sulfate production process [Gaston,

et al., 2010;Sahle-Demessie and Devulapelli, 2008]. The fact that plumes quickly cool to

ambient temperatures [Chosson, et al., 2008], combined with the high relative humidity

in the marine sampling location and the presence of NO2 (from the fast reaction between

O3 and NO) in the plumes, these atmospheric measurements show that catalytic sulfate

production could indeed be an important atmospheric process.

3.4.11 Sulfate and Sulfuric Acid in Additional Plumes

To provide greater detail on the relationship between sulfate and sulfuric acid

with different particle types during plumes, 4 additional plumes are shown in Figure 3.12.

Similarly to Figure 3.12, the average sulfate peak area and sulfuric acid peak areas are

91

shown on the left axes by particle type and the number fraction of particles within a type

for sulfate and sulfuric acid are shown on the right axis. For each of these 4 ship plumes

sulfate and sulfuric acid peak area are shown to be roughly an order of magnitude greater

on OC-V-sulfate particles in comparison to all other particle types observed. The fraction

of particles containing sulfate and sulfuric acid peaks are also considerably higher for the

OC-V-sulfate particle type. This consistency plume-to-plume supports our hypothesis

that sulfate formation may be catalyzed by OC-V-sulfate particles.

92

Figure 3.12 Plots of sulfate peak area (red circles), sulfuric acid peak area (blue circles), sulfate number fraction (red diamonds), and sulfuric acid number fraction (blue diamonds). The plumes shown are a) container2, b) container1b, c) container3, d) tanker1a.

93

3.4.12 Conclusions

Particulate emissions from ships are expected to increase ~5% globally by 2030.

The production of large amounts of sulfate will impact climate through both cloud and

radiative transfer processes [Eyring, et al., 2005]. Increasing ship emissions will increase

concentrations of submicron OC-V-sulfate particles [Agrawal, et al., 2008a], posing

deleterious consequences for human health [Campen, et al., 2001;Corbett, et al., 2007].

Herein, direct single-particle measurements of ship plume particle mixing-states provide

key insights into the chemistry of freshly emitted ship plumes. The ability to use

vanadium as a tracer for ship plumes through single-particle mass spectrometry in the

polluted environment of the Port of LA will strengthen efforts to accurately apportion

particulate matter to sources in other coastal environments. In addition, current emissions

inventories do not account for fuel type and potential conversion of SO2 by vanadium

and/or other metals (i.e. Fe and Mn) in the presence of H2O and NO2 (both observed in

the plumes). The elevated concentrations of vanadium (maximum allowable: 600 mg/kg)

present in the fuel and thus ship plumes could be leading to higher than estimated

atmospheric sulfate concentrations [International Standards Organization, 2005].

Incorporating fuel type could help explain the enhanced sulfate levels being measured in

places such as California that current emissions inventories cannot explain [2009]. Thus,

the enrichment of sulfate and sulfuric acid on OC-V-sulfate particles due to possible

catalytic aqueous phase reactions occurring in particles enriched with vanadium has

important implications for regulating anthropogenic sulfate levels in coastal

94

environments. The vanadium and sulfur content of fuels and potential impacts on sulfate

production should be taken into account in future inventories and air quality models.


The authors acknowledge Dr. Robert Moision and Melanie Zauscher for

assistance during the study. Prof. V. Ramanathan and Dr. Craig Corrigan are

acknowledged for providing an aethalometer. The Southern California Marine Institute,

specifically Dr. Richard Peiper and Carrie Wolfe, hosted the sampling. Funding was

provided by the California Air Resources Board (#04-336).

Chapter 3 is reproduced with permission from A. P. Ault, C. J. Gaston, Y. Wang,

G. Dominguez, M.H. Thiemens, and K.A. Prather, Characterization of the Single Particle

Mixing State of Individual Ship Plume Events Measured at the Port of Los Angeles,

Environmental Science & Technology, 44 (6), 1954-1961, 2010. Copyright 2010,

American Chemical Society.

95

3.6 References

California Air Resources Board Personal Communication, 2009. H. Agrawal, Q. G. J. Malloy, W. A. Welch, J. Wayne Miller and D. R. Cocker III, In-use

gaseous and particulate matter emissions from a modern ocean going container vessel, Atmospheric Environment, 42 (21), 5504, 2008a.


a Crude Oil Tanker at Sea, Environ. Sci. Technol., 2008b. A. P. Ault, M. J. Moore, H. Furutani and K. A. Prather, Impact of emissions from the Los

Angeles port region on San Diego air quality, Environmental Science & Technology, 43 (10), 3500-3506, 2009.

B. Barbaray, J. P. Contour and G. Mouvier, Sulfur-Dioxide Oxidation Over Atmospheric

Aerosol - X-Ray Photoelectron-Spectra Of Sulfur-Dioxide Adsorbed On V2O5 And Carbon, Environmental Science & Technology, 11 (4), 351-356, 1977.

B. Barbaray, J. P. Contour and G. Mouvier, Effects Of Nitrogen-Dioxide And Water-

Vapor On Oxidation Of Sulfur-Dioxide Over V2O5 Particles, Environmental Science & Technology, 12 (12), 1294-1297, 1978.

P. V. Bhave, J. O. Allen, B. D. Morrical, D. P. Fergenson, G. R. Cass and K. A. Prather,

A field-based approach for determining ATOFMS instrument sensitivities to ammonium and nitrate, Environmental Science & Technology, 36 (22), 4868-4879, 2002.

C. Brandt and R. Vaneldik, Transition-Metal-Catalyzed Oxidation Of Sulfur(Iv) Oxides -

Atmospheric-Relevant Processes And Mechanisms, Chem. Rev., 95 (1), 119-190, 1995.



F. Chosson, R. Paoli and B. Cuenot, Ship plume dispersion rates in convective boundary

layers for chemistry models, Atmos. Chem. Phys., 8 (16), 4841, 2008. J. J. Corbett, J. J. Winebrake, E. H. Green, P. Kasibhatla, V. Eyring and A. Lauer,

Mortality from ship emissions: A global assessment, Environmental Science & Technology, 41 (24), 8512-8518, 2007.

96

L. A. De Bock, P. E. Joos, K. J. Noone, R. A. Pockalny and R. E. Van Grieken, Single particle analysis of aerosols, observed in the marine boundary layer during the Monterey Area Ship Tracks Experiment (MAST), with respect to cloud droplet formation, Journal Of Atmospheric Chemistry, 37 (3), 299-329, 2000.

T. M. Exchange, The Marine Exchange, Ship Traffic Logs for the Ports of Los

Angeles/Long Beach, San Pedro, CA, 2007. V. Eyring, I. S. A. Isaksen, T. Berntsen, W. J. Collins, J. J. Corbett, O. Endresen, R. G.

Grainger, J. Moldanova, H. Schlager and D. S. Stevenson, Transport Impacts on Atmosphere and Climate: Shipping, Atmospheric Environment, In Press, Accepted Manuscript, 2009.

V. Eyring, H. W. Kohler, A. Lauer and B. Lemper, Emissions from international

shipping: 2. Impact of future technologies on scenarios until 2050, Journal Of Geophysical Research-Atmospheres, 110 (D17), 2005.

B. J. Finlayson-Pitts and J. N. Pitts, Chemistry of the Upper and Lower Atmosphere,

Academic Press, San Diego, CA, 2000. C. J. Gaston, K. A. Pratt, X. Y. Qin and K. A. Prather, Real-Time Detection and Mixing

State of Methanesulfonate in Single Particles at an Inland Urban Location during a Phytoplankton Bloom, Environmental Science & Technology, 44 (5), 1566-1572, 2010.

R. M. Healy, I. P. O'Connor, S. Hellebust, A. Arnaud, J. R. Sodeau and J. C. Wenger,

Characterisation of single particles from in-port ship emissions, Atmospheric Environment, 43 (40), 6408-6414, 2009.

International Standards Organization, Petroleum products -- Fuels (class F) --

Specifications of marine fuels, 22 pp., 2005. J. Isakson, T. A. Persson and E. S. Lindgren, Identification and assessment of ship

emissions and their effects in the harbour of Goeteborg, Sweden, Atmospheric Environment, 35 (21), 3659-3666, 2001.

D. Kim, C. Wang, A. M. L. Ekman, M. C. Barth and P. J. Rasch, Distribution and direct

radiative forcing of carbonaceous and sulfate aerosols in an interactive size-resolving aerosol-climate model, Journal Of Geophysical Research-Atmospheres, 113 (D16), 2008.

D. Lack, B. Lerner, C. Granier, T. Baynard, E. Lovejoy, P. Massoli, A. R. Ravishankara

and E. Williams, Light absorbing carbon emissions from commercial shipping, Geophysical Research Letters, 35 (13), 2008.

97

D. A. Lack, J. J. Corbett, T. Onasch, B. Lerner, P. Massoli, P. K. Quinn, T. S. Bates, D. S. Covert, D. Coffman, B. Sierau, S. Herndon, J. Allan, T. Baynard, E. Lovejoy, A. R. Ravishankara and E. Williams, Particulate emissions from commercial shipping: Chemical, physical, and optical properties, Journal of Geophysical Research, 114, 2009.

D. A. Lake, M. P. Tolocka, M. V. Johnston and A. S. Wexler, The character of single

particle sulfate in Baltimore, Atmos. Environ., 38 (31), 5311-5320, 2004. D. R. Lide, (ed.), CRC Handbook of Chemistry and Physics, CRC Press/Taylor and

Francis, Boca Raton, FL, 2009. P. H. McMurry and J. C. Wilson, Droplet Phase (Heterogeneous) And Gas-Phase

(Homogeneous) Contributions To Secondary Ambient Aerosol Formation As Functions Of Relative-Humidity, Journal Of Geophysical Research-Oceans And Atmospheres, 88 (NC9), 5101-5108, 1983.

R. C. Moffet and K. Prather, Aging of soot captured with in situ measurements:

Implications for aerosol climate forcing, Proceedings Of The National Academy Of Sciences Of The United States Of America, Submitted, 2008.

R. C. Moffet and K. A. Prather, In-situ measurements of the mixing state and optical

properties of soot with implications for radiative forcing estimates, 106 (29), 11872-11877, 2009.

S. M. Murphy, H. Agrawal, A. Sorooshian, L. T. Padro, H. Gates, S. Hersey, W. A.

Welch, H. Jung, J. W. Miller, D. R. Cocker, A. Nenes, H. H. Jonsson, R. C. Flagan and J. H. Seinfeld, Comprehensive Simultaneous Shipboard and Airborne Characterization of Exhaust from a Modern Container Ship at Sea, Environmental Science & Technology, 43 (13), 4626-4640, 2009.

K. R. Neubauer, M. V. Johnston and A. S. Wexler, On-line analysis of aqueous aerosols

by laser desorption ionization, International Journal Of Mass Spectrometry And Ion Processes, 163 (1-2), 29-37, 1997.

I. M. Organization, International Convention on Safety of Life at Sea (SOLAS) Chapter

V, 2000. K. A. Pratt, L. E. Hatch and K. A. Prather, Seasonal volatility dependence of ambient

particle phase amines, Environmental Science & Technology, 43 (14), 5276-5281, 2009a.


Furutani, M. Gonin, K. Fuhrer, Y. X. Su, S. Guazzotti and K. A. Prather,

98

Development and Characterization of an Aircraft Aerosol Time-of-Flight Mass Spectrometer, Analytical Chemistry, 81 (5), 1792-1800, 2009b.

M. S. Reinard, K. Adoua, J. M. Martini and M. V. Johnston, Source characterization and

identification by real-time single particle mass spectrometry, Atmospheric Environment, 41 (40), 9397-9409, 2007.

E. Sahle-Demessie and V. G. Devulapelli, Vapor phase oxidation of dimethyl sulfide

with ozone over V2O5/TiO2 catalyst, Appl. Catal. B-Environ., 84 (3-4), 408-419, 2008.



B. Silverman, Pacific Merchant Shipping Association v. James Goldstene, United States

Court of Appeals for the Ninth Circuit, No. 07-16695, 2008. P. Sinha, P. V. Hobbs, R. J. Yokelson, T. J. Christian, T. W. Kirchstetter and R.

Bruintjes, Emissions of trace gases and particles from two ships in the southern Atlantic Ocean, Atmospheric Environment, 37 (15), 2139-2148, 2003.





S. South Coast Air Quality Management District, Multiple Air Toxics Exposure Study

III,South Coast Air Quality Management District, 2008. M. T. Spencer and K. A. Prather, Using ATOFMS to determine OC/EC mass fractions in

particles, Aerosol Science And Technology, 40 (8), 585-594, 2006. M. T. Spencer, L. G. Shields and K. A. Prather, Simultaneous measurement of the

effective density and chemical composition of ambient aerosol particles, Environmental Science & Technology, 41 (4), 1303-1309, 2007.



99

S. M. Toner, D. A. Sodeman and K. A. Prather, Single particle characterization of ultrafine and accumulation mode particles from heavy duty diesel vehicles using aerosol time-of-flight mass spectrometry, Environmental Science & Technology, 40 (12), 3912-3921, 2006.

C. F. Wang, J. J. Corbett and J. J. Winebrake, Cost-effectiveness of reducing sulfur

emissions from ships, Environmental Science & Technology, 41 (24), 8233-8239, 2007.

4 Mobile Laboratory Observations of Temporal and Spatial

Variability within the Coastal Urban Aerosol

4.1 Synopsis

Understanding the spatial and temporal variation of particle mass and number

concentrations, size distributions, and chemical composition within urban areas is critical

to quantifying particle exposure levels and impacts on human health. However, it is

difficult to measure these variations with neighborhood level spatial resolution and high

time resolution (< 1 hr). This often leads to simplifications and assumptions when models

simulate the urban aerosol and an unrealistic representation. A mobile laboratory

equipped with an aerosol time-of-flight mass spectrometer (ATOFMS) and instruments to

measure PM2.5 mass concentrations, particle size distributions, and gas phase

concentrations was used to sample ambient air two different days around San Diego Bay

in California with similar meteorological and gas phase conditions, but low particle mass

and number concentrations on one day (Feb. 12th) and high concentrations on another

(Mar. 13th). Feb. 12th was identified as representing urban background conditions through

mass concentrations, number concentrations, and particle chemistry that were similar

throughout the San Diego Bay region. On Mar. 13th the urban aerosol transitioned as the

aged anthropogenic/secondary organic particles were transported inland and replaced by

fresh combustion and sea salt particles. The shift away from an aged aerosol was

observed consistently across the San Diego Bay region and was not observed to vary

spatially. Conversely, particle concentrations were observed to vary spatially identifying

downtown as the most concentrated source region observed around the bay with the

100

101

highest concentrations of fresh combustion particles; this spatial variation did not follow

discernible temporal trends. These results demonstrate that the high resolution temporal

and spatial patterns in particle mass and number concentrations around the San Diego

Bay are driven by the chemical composition of the urban aerosol. Further studies are

needed with high temporal and spatial resolution to determine how these changes affect

exposure levels and ultimately public health.

4.2 Introduction

Particulate matter (PM) has been shown to negatively impact human health over

short and long term periods of exposure [Pope, 2007;Poschl, 2005;Valavanidis, et al.,

2008]. Short term effects from exposure to elevated levels of particulate matter include

increased hospital admissions for asthma, cardiovascular disease, and chronic obstruction

pulmonary disease when PM with diameters ≤ 10 μm (PM10) increases by 10 μg/m3

[Valavanidis, et al., 2008]. At a series of hospitals in Utah, a recent study showed that a

10 μg/m3 increase in PM2.5 (particles with diameters < 2.5 μm) increased the risk of

ischemic heart disease events (e.g. heart attacks) by 4.5% in patients [Pope, et al., 2006].

Long term studies have shown connections between particulate matter and lung cancer,

cardiopulmonary mortality, and cardiovascular disease [Pope, 2007;Valavanidis, et al.,

2008].

In addition to mass concentrations, particle-health effects have been linked to

particle number concentration, size, hygroscopicity, and chemical composition [Nel,

2005;Pope, 2007;Valavanidis, et al., 2008]. Elevated number concentrations and ultrafine

(< 0.1 μm) number concentrations can cause oxidative stress in lung cells [Xia, et al.,

2006]. Particle size plays a key role in determining the location that inhaled particles

102

deposit in the tracheobronchial and pulmonary systems [Asgharian, 2004;Park and Wex,

2008;Yeh, et al., 1996]. Particle growth, when exposed to high relative humidity after

inhalation, impacts particle size and chemical composition and determines where

particles deposit in the body [Asgharian, 2004]. The chemical composition of particles

also determines particle solubility, which plays a role in the negative health effects from

atmospheric particles [Jalava, et al., 2008]. In addition, submicron particles are more

likely than supermicron particles to be inhaled and deposited into the tracheobronchial

and pulmonary regions of the body than the naso-oro-pharyngo-laryngeal regions

[Rostami, 2009]. Negative synergistic effects have been shown when specific chemical

species are present together such as oxidative stress and NFκB activation (iron and soot),

as well as cardiovascular and thermoregulatory effects (vanadium and nickel) [Campen,

et al., 2001;Zhou, et al., 2003]. While health effects have been linked to particle mass

concentrations, number concentrations, size, and chemical composition; studying the link

between particles and health is a relatively new and rapidly developing field. Due to the

considerable amount of uncertainty regarding particles-health impacts more detailed

measurements are necessary to develop effective regulations to mitigate the detrimental

impacts of particle on public health.

Understanding particle-health impacts is complicated by the fact that while

negative health effects have been linked to the different metrics, such mass and number

concentrations as discussed above, these metrics frequently do not track one another. To

determine the metric that correlates most strongly with decreases in human health, a

thorough understanding of how particle concentrations and chemistry vary temporally

and spatially is needed. This is supported by a recent study which found neighborhood to

103

neighborhood changes in PM2.5 exposure levels and subsequently life expectancy across

an urban area [Orru, et al., 2009]. Number concentrations often have even more

heterogeneous spatial and temporal patterns than mass concentrations; further

complicating attempts to understand particle-health relationships [Harrison and Jones,

2005;Moore, et al., 2009]. Recent studies have started to investigate changes in particle

number concentrations and sizes with high spatial and temporal resolution by operating

multiple sampling sites simultaneously and deploying mobile/portable laboratories. Mejia

et al. (2008) compared measurements at 8 sites in Brisbane, Australia and found

considerable variability in number concentrations and size distributions depending on

whether the vehicle fleet was primarily cars or trucks [Mejia, et al., 2008]. Krudysz et al.

(2009) and Moore et al. (2009) monitored 13 sites around the Ports of Los Angeles and

Long Beach and also found spatial variability in particle concentrations and size

distributions, especially among ultrafine particles [Krudysz, et al., 2009;Moore, et al.,

2009]. High spatial variation in particle number concentration was also observed by

Cyrys et al. (2008) in Augsburg, Germany [Cyrys, et al., 2008].

While recent studies have improved our understanding of changes in size

distributions and number concentrations, high time resolution measurements of the

spatial variation in chemical measurements has been lacking. This is because studies

monitoring multiple sites continuously have primarily monitored particle chemistry with

low time resolution filter techniques. Recently, many portable and/or mobile laboratories

have been developed which can monitor and move quickly between sampling locations or

sample while moving giving both high temporal and spatial resolution [Bukowiecki, et

al., 2002;Kolb, et al., 2004]. Many of these mobile laboratories have focused on fresh

104

vehicle emissions [Bukowiecki, et al., 2003;Canagaratna, et al., 2004;Zavala, et al.,

2009] or spatial patterns of specific species (i.e. hexavalent chromium) [Isakov, et al.,

2007]. Herein, this paper discusses measurements involving the development and

deployment of a mobile aerosol time-of-flight mass spectrometry (ATOFMS) laboratory.

These measurements of spatial and temporal variability with real-time particle size and

chemistry mark the first single particle mass spectrometry measurements at multiple sites

in a single day, providing standard bulk measurements (particle mass and number

concentrations), as well as the chemical composition and mixing state with secondary

species to explain changes to these bulk concentrations.

4.3 Experimental

4.3.1 Mobile ATOFMS Laboratory

A Pace American Class IV trailer (length 5.5 m; width 3.5 m; height 3 m;

maximum gross trailer weight 4,500 kg) was modified to house an array of gas phase and

particle instrumentation. The mobile ATOFMS laboratory was transported between sites

by a pickup truck holding two Honda EB11000 gasoline generators (9.5 kVA) supported

by a steel frame anchored to the truck bed. The generator exhaust passed through an

exhaust manifold and 20 m of 6 cm diameter steel tubing to transport the emissions

downwind of the sampling inlet. One generator was capable of powering the trailer at full

load for 3-4 hours under normal operating conditions on a single tank of gas. This study

focused on stationary sampling. At each site a 2.5 meter sampling line was attached to a

sampling manifold on the mobile ATOFMS laboratory roof (5.5 meters above the

ground). The manifold provided laminar sampling flow for 5-6 instruments in parallel

with similar residence times.

105

4.3.2 2009 San Diego Bay Measurements

The mobile ATOFMS laboratory sampled the urban aerosol at multiple locations

around San Diego Bay on both Feb. 12 and Mar. 13, 2009. Sampling times and locations

are provided in Figure 4.1 and Table 4.1. Table 4.1 also includes major sources within 1

km upwind of each site and the distance to the source. On Feb. 12th, sampling was

conducted at three locations: Silver Strand State Beach (SS), Cesar Chavez Park (CC) in

downtown San Diego, and Coronado Tidelands Park (Cor). These same three sites were

sampled on Mar. 13th, as well as Chula Vista Bayfront Park (CV) and Pepper Park (Pep)

in National City. Throughout this paper SS1, CC1, and Cor1 refer to the measurements at

the three sites on Feb. 12th and SS2, Cor2, CC2, CV2, and Pep2 reference the

measurements made on Mar. 13th.

4.3.3 Instrumentation

Standard aerosol instrumentation in the mobile ATOFMS laboratory included

PM2.5 mass concentrations from a beta attenuation monitor (BAM) (MetOne Model

BAM-1020) and black carbon mass concentrations from a seven wavelength

aethalometer (Model AE31, Magee Scientific). Aerosol size distributions were measured

between 0.011 - 0.604 μm using a scanning mobility particle sizer (SMPS) (Model

3936L, TSI Inc.) and 0.523 -10 μm using an aerodynamic particle sizer (APS) (Model

3321, TSI Inc.). Measurements of meteorology, gas phase concentrations, and mass

concentrations (PM2.5) were also incorporated from San Diego Air Pollution Control

106

Figure 4.1 Map of the San Diego metropolitan area around San Diego Bay with trailer sampling sites and air pollution control district (APCD) sites, as well as 200 meter HYSPLIT air mass back trajectories for (a) Feb. 12, 2009 and (b) Mar. 13, 2009.

107

Table 4.1 Sampling times, locations, and wind direction (WD) at sites around San Diego Bay are described. Sources within 1,000 meters of the sampling site and distances are listed.

Site Abbrev Date Start Stop Latitude Longitude WD Anth. Sources DistTime Time (°N) (°W) From WD (m)

Silver Strand SS1 2/12/2009 9:48 12:18 32° 37' 58" 117° 08' 29" 280 None N/ACesar Chavez Park CC1 2/12/2009 14:21 16:10 32° 41' 46" 117° 09' 01" 287 Industry and Port 300

Coronado Cor1 2/12/2009 17:26 19:09 32° 41' 28" 117° 09' 58" 360 Minor Road 250Silver Strand SS2 3/13/2009 8:29 10:20 32° 37' 58" 117° 08' 29" 259 None N/A

Coronado Cor2 3/13/2009 11:41 13:03 32° 41' 28" 117° 09' 58" 272 Major Road 75Cesar Chavez Park CC2 3/13/2009 13:57 15:02 32° 41' 46" 117° 09' 01" 274 Industry and Port 400Chula Vista Park CV2 3/13/2009 16:10 17:03 32° 37' 11" 117° 06' 11" 274 None N/A

Pepper Park Pep2 3/13/2009 18:35 20:02 32° 39' 02" 117° 06' 38" 222 None N/A

108

District (SD-APCD) sites in downtown San Diego (DT-APCD), Chula Vista (CV-

APCD), and at the San Diego International Airport (SDAP).

4.3.4 Aerosol Time-of-Flight Mass Spectrometry

An ATOFMS was operated inside the mobile laboratory measuring the size and

chemical composition of individual particles between 0.2-3.0 μm in real-time at each site.

The design and details of the ATOFMS have been described in detail previously [Gard,

et al., 1997]. Briefly, particles are introduced into the ATOFMS through a converging

nozzle into a differentially pumped vacuum chamber where the particles are accelerated

to a size-dependent velocity. The particles pass through two continuous wave lasers

(diode pumped Nd:YAG lasers operating at 532 nm) located 6 cm apart. Particle speed is

used to determine vacuum aerodynamic diameter by calibration with polystyrene latex

spheres of known size. The sized particles are desorbed and ionized in the mass

spectrometer source region by a 266 nm Q-switched Nd:YAG laser (1.2-1.4 mJ). Positive

and negative ions from the same single particle are detected using a dual-reflectron time-

of-flight mass spectrometer.

4.3.5 Single Particle Data Analysis

Single particle size and mass spectral information were analyzed with YAADA

1.2 (www.yaada.org), a data analysis toolkit for MATLAB 6.5.1 (The MathWorks, Inc.).

Particles were analyzed via two approaches: 1) searching mass spectral, aerodynamic

size, and temporal features and 2) clustering mass spectra using an Adaptive Resonance

Theory based neural network algorithm (ART-2a) at a vigilance factor of 0.8 [Song, et

al., 1999]. ART-2a combines particles into clusters based on the intensity of ion peaks in

109

individual mass spectra. General particle types are defined by the characteristic chemical

species or possible source to simplify the naming scheme; these labels do not reflect all

of the species present within a specific particle type, but reflect the most intense ion

peaks. Peak identifications within this paper correspond to the most probable ions for a

given m/z ratio.

4.3.6 Scaling Single Particle Measurements to Mass Concentrations

Number and mass concentrations were calculated for different particle types using

APS size-resolved number concentrations, a method which has been shown previously to

yield quantitative mass concentrations [Qin, et al., 2006]. Briefly, ATOFMS counts are

divided into APS size bins (0.523 – 2.5 μm) and scaling factors were calculated for each

site to account for ATOFMS transmission biases. Bins with less than 10 particles at a site

were excluded due to low statistics, which might lead to slight undercounting of PM1 and

PM2.5. After scaling to number concentration, mass was calculated using a representative

density for particles in an urban atmosphere that have taken up water and secondary

species, 1.5 g/cm3 [Moffet, et al., 2008]. Scaled ATOFMS mass concentrations have been

shown previously to correlate well with standard mass measurements and comparisons to

the BAM PM2.5 mass concentrations are included below [Qin, et al., 2006]. Previous

ATOFMS measurements in San Diego used scaled mass concentrations to determine the

impact of transported particles from the Los Angeles port region, demonstrating the

utility of the method locally [Ault, et al., 2009].

110


4.4.1 Temporal Patterns of Meteorological and Gas Phase Concentrations

To characterize the ambient conditions on Feb. 12th and Mar. 13th, the

meteorological and gas phase concentrations for each day are compared at downtown

APCD sites (Figure 4.2). The meteorological conditions on both Feb. 12th and Mar. 13th

had diurnal trends in temperature and relative humidity (RH). The wind direction (WD)

and wind speed (WS) on Feb. 12th were also diurnal, shifting from a light easterly land

breeze (~ 1 m/s) before sunrise to a stronger westerly sea breeze (~ 3 m/s) during the day

and shifting back to a light easterly land breeze after sunset. The wind data patterns on

Mar. 13th were similar during sampling (although at night there was not a shift to a land

breeze) and meteorological data from the Chula Vista APCD site was similar to the

downtown APCD sites on both days.

The gas phase concentrations observed on Feb. 12th and Mar. 13th were also

similar (Figure 4.2). An increase in NOx concentrations was observed before dawn from

vehicle traffic, followed by a decrease after sunrise as O3 concentrations increased. On

Feb. 12th, the O3 decreased after sunset and the wind shifted to a land breeze with NOx

increasing. The lack of a shift in wind direction on Mar. 13th was accompanied by high

O3 concentrations well after dark and sustained low NOx concentrations. On both days,

SO2 concentrations were low, and no obvious temporal trends were observed. Despite

similar trends in meteorological and gas phase data significant differences in particle

mass concentrations, number concentrations, and chemistry were observed around San

Diego Bay providing insight into sources and aging for two meteorologically similar

days.

111

Figure 4.2 Comparison of meteorological conditions (temperature-red, relative humidity-light blue, wind speed-grey, wind direction-black markers) and gas phase concentrations (O3-blue, NOx-green, SO2-gold) on Feb. 12th and Mar. 13th. PM2.5 from the BAM is shown as a black line.

112

Air mass back trajectories were calculated using HYSPLIT 4.9 for each site

during the sampling periods (38). Figure 4.1a shows the 200 meter air mass back

trajectory for the three sites sampled on Feb. 12th. All three back trajectories show a

strong oceanic influence with each air mass overland for less than an hour, which

corresponds with the wind direction in Figure 4.2a. The trajectory calculated for the

initial site on Mar. 13th was moved slowly over the San Diego area for 6 hours taking up

gases and particles from many local sources leading to a polluted air mass. A strong sea

breeze during the 4 remaining sites on Mar. 13th resulted in cleaner conditions than earlier

in the day. An interesting contrast can be drawn between the relatively clean day on Feb.

12th and the transition from polluted to clean conditions on Mar. 13th Taking an in-depth

view of Mar. 13th provides information about how quickly the particles in San Diego Bay

region can change in terms of chemistry and sources.

Particulate mass concentration is the metric used to evaluate particle pollution by

state and federal regulatory agencies. The BAM is a federally certified method and

APCD BAM measurements (Figure 4.2) exhibit different patterns on the two sampling

days. On Feb. 12th, which was preceded by 5 days of rain, a diurnal trend with low

concentrations during the day (1-3 μg/m3) was observed. The more polluted case of Mar.

13th (6-13 μg/m3) had a diurnal trend except for a peak in mass concentration between

08:00-10:00 PST. From the meteorology, gas phase, and BAM measurements, Feb. 12th

can be classified as urban background conditions for this coastal urban environment, and

Mar. 13th can be classified as a shift from regional pollution buildup to marine/fresh

emission conditions.

113

4.4.2 Variability of Particle Mass and Number Concentrations

To determine whether mass concentrations observed at the downtown site were

representative of the entire San Diego Bay region, two PM2.5 mass concentrations

measured at the sample sites (scaled ATOFMS and BAM) were compared with BAM

measurements at the downtown site (Figure 4.3a). On the urban background day (Feb.

12th) mass concentrations for the downtown BAM and the mobile site BAM are in

agreement, with values within 1 μg/m3 of one another. The ATOFMS mass

concentrations compare well to the mobile site and downtown BAM at two of the three

sites (CC1 and Cor1), while SS1 was an outlier due to an instrumental artifact. Number

concentrations (Figure 4.3b – orange bars) were also consistent (1865-4169 #/cm3), but

low for an urban environment (typically ~105 #/cm3) [Finlayson-Pitts and Pitts, 2000].

This shows homogeneous values throughout the day around the San Diego Bay region

during urban background conditions.

In contrast, on Mar. 13th (with regional buildup shifting to marine/fresh emission

conditions) the mass measurements downtown versus the mobile sites do not agree

throughout the day (Figure 4.3a), though the two independent mass measurements

(ATOFMS and BAM) at the mobile sites are within error of one another at two of the

three sites with reliable measurements of both (SS2 and CV2). At mobile sites further

from downtown a significant difference in mass concentration between downtown and

the mobile sites is observed, while at mobile sites close to downtown the mass

concentrations are similar. This indicates that concentrations are not homogeneous

around the bay, as they were during background conditions. The higher concentrations

observed downtown demonstrate spatial variation in mass concentrations and show that

114

Figure 4.3 (a) PM2.5 mass concentration data at each site for the: Scaled ATOFMS mass (blue bars), mobile site BAM (red bars), and DT-APCD BAM (green bars). (b) Number concentration by site from summed SMPS scans (gold bars).

115

downtown is likely a source region. When viewed by instruments that give greater detail

on aerosol properties these relatively small differences in BAM concentrations result in

very different in particle number concentrations, size, and chemistry to which residents

are exposed.

The number concentrations on Mar. 13th show both temporal and spatial variation

around the San Diego Bay. Temporally, a steady decrease from the high concentrations

during local stagnant conditions to lower concentrations as the sea breeze strengthens is

overlaid with a spatial pattern of higher number concentrations with increasing proximity

to downtown. Taken together the mass and number concentrations suggest that

downtown San Diego is the largest source region influencing San Diego Bay and,

although point sources do exist at other points around the bay, these did not disrupt the

trends observed at the mobile sites during sampling. The relative importance of the

temporal trend versus spatial trend with respect to what is being inhaled will be

investigated both from a bulk perspective by size distribution measurements and from a

single particle perspective by analyzing particle size, sources, and secondary chemistry.

4.4.3 Variation in Aerosol Size Distributions

Size distributions were analyzed at the mobile lab sites for both Feb. 12th (Figure

4.4a) and Mar. 13th (Figure 4.4b). At SS1, a coastal site that had a strong sea breeze

during sampling, a higher concentration of larger particles (i.e. sea salt) was observed.

CC1, an urban site close to many sources, had more Aitken mode particles (0.010-0.030

μm) and fewer particles > 0.030 μm than the coastal site. The APS size distributions for

CC1 and Cor1 are nearly identical as would be expected for two sites ~1.5 km apart on a

day when mass concentrations are similar around the bay. Overall, the SMPS and APS

116

Figure 4.4 Size distribution measurements (SMPS and APS) for Feb. 12th and Mar. 13th as measured at the mobile sampling sites. The red traces correspond to Silver Strand (SS), green traces to Cesar Chavez Park (CC), yellow to Coronado (Cor), blue to Chula Vista (CV), and purple to Pepper Park (Pep).

117

scans from Feb. 12th agree with the mass and number concentrations observed and are as

expected for a day with urban background conditions.

The modes and shapes of the size distributions on Mar. 13th support the finding of

a temporal trend of decreasing aerosol concentrations as a sea breeze built in and a spatial

trend of higher concentration near downtown. The larger number of ultrafine particles

earlier in the day (SS2, Cor2, and CC2) and low concentrations of ultrafine particles later

in the day support the temporal trend discussed above. Additional evidence supporting

Feb. 12th as a background time period and of local pollution buildup on Mar. 13th is that

the downtown site (CC) on Feb. 12th has a smaller mode (0.018 μm) than Mar. 13th

(0.055 μm), which indicates a small number of fresh particles at the source region on Feb.

12th and a larger number of particles that have undergone some atmospheric processing

on Mar. 13th. The spatial trend of downtown as a source of aerosols is supported by

higher numbers of the smallest (< 0.015 μm) particles at the downtown site (CC2) and a

nearby site (Cor2) then at the site sampled earlier in the day (SS2) and later in the day

(CV2 and Pep2). Additionally, CC2 has a greater variation in number concentrations than

any of the other sites suggesting that downtown is a fresh and dynamic source of

aerosols.

An important aspect of Figure 4.4 is that the size distributions on Feb. 12th in the

overlap region between the SMPS and APS (0.523-0.604 μm) are within instrumental

variation at those sizes, but there is a larger disconnect in the overlap region between the

SMPS and APS measurements on Mar. 13th that instrumental variation cannot explain.

This is likely due to the fact that the SMPS measures electrical mobility diameter (dm)

and the APS measures aerodynamic diameter (da). These two parameters are related to

118

one another by the following equation [DeCarlo, et al., 2004] where ρ0 is the unit density

(1 g/cm3), ρp is the particle density, and χ is the dynamic shape factor:

0

pa

m

d

d

(0.1)

As can be seen from the equation, as shape factor approaches 1 (spherical) and

density approaches the unit density, aerodynamic and electrical mobility diameter

converge. This indicates that since the distributions align on Feb. 12th, particles were on

average spherical and had taken up water decreasing density. It has been shown

previously by Spencer et al. that when atmospheric liquid water content increases,

effective density approaches unity [Spencer, et al., 2007]. By comparison, on Mar. 13th

there is a multiple order of magnitude difference in number concentration in the overlap

region indicating different types of particles that have non-unity density or shape factor,

as mobility and aerodynamic diameter give different distributions. These differences

show that particles are likely more dense by the direction of the shift between SMPS and

APS distribution. This indicates that particles on Mar. 13th have not taken up as much

water as on Feb. 12th, but likely taken up higher density secondary species. The higher

particle density on a polluted than on a day with urban background conditions can be

used to infer information about the particles, such as particle solubility or hygroscopicity,

which are related to human health. To directly gain detailed information on particle

composition measurements of single particle size and chemistry were made and are

discussed below.

119

4.4.4 Differing Aerosol Sources Determined by Particle Chemistry

Single particle measurements at each site using ATOFMS were used to determine

the extent of diurnal and spatial patterns influenced by the types of particles observed

based on sources and the degree to which they had been atmospherically processed (i.e.

aged). The main particle types observed at the different sampling sites included: sea salt,

dust, elemental carbon mixed with organic carbon (ECOC), organic carbon (OC), and

potassium-combustion (K-Combustion) particles. These particle types have been

described for the San Diego area previously [Ault, et al., 2009;Guazzotti, et al., 2001].

Briefly, sea salt particles were identified by 23Na+, 62Na2O+, 63Na2OH+, and 81,83Na2Cl+

and dust particles were identified by peaks corresponding to 23Na+, 27Al+, 39K+, 56Fe+, -

76SiO3-, and -77HSiO3

-. Organic carbon (OC) particles had high intensity peaks at

27C2H3+, 29C2H5

+, 37C3H+, and 43C2H3O+, as well as other carbon fragments, and is

indicative of anthropogenic secondary organic aerosol. The average mass spectra of the

most abundant OC cluster during Mar. 13th compared with the average mass spectra from

the most abundant cluster during the Study of Organic Aerosols in Riverside (SOAR) had

a dot product of 0.9409 [Qin, et al., in prep]. The organic carbon aerosol in Riverside, CA

during SOAR has been identified as an aged secondary organic carbon by thermodenuder

studies and the similarity between the spectra suggests that the OC particles on Mar. 13th

were secondary anthropogenic organic aerosol as well [Pratt and Prather, 2009b].

Elemental carbon particles mixed organic carbon (ECOC) were characterized by peaks at

(12C1+, 24C2

+,…, Cn+) with less intense organic carbon markers (27C2H3

+, 29C2H5+,

37C3H+, and 43C2H3O+); and K-Combustion particles identified by a strong potassium

peak (39K+) and minor organic carbon ion markers. Many K-Combustion particles are

120

from biomass burning, but other sources exist in the San Diego Bay region with a similar

signature, which is why we use the broader label. The mixing state of different particle

types with markers for clean (chloride) and aged (nitrate and sulfate) particles will be

discussed below.

The relative mass fractions for submicron and supermicron particle types from

difference sources are shown Figures 4.5a and 4.5b respectively, with the black markers

representing the total ATOFMS mass concentration at each site for that size range. The

very low concentrations of submicron non-sea salt particle types (less than 10% at each

site) support the earlier finding of very clean urban background conditions. On the more

polluted Mar. 13th between 25-50% of the submicron particle mass is carbonaceous

particles (OC, ECOC, and K-combustion) from anthropogenic sources. The OC mass

decreases over the course of the day from 10% of submicron mass to negligible levels as

the sea breeze pushes the secondary OC particles that formed overnight across the San

Diego region inland replacing them with fresh emissions sea salt particles. The combined

fraction of mass from ECOC and K-Combustion particles is substantial throughout the

day (14-32%), indicating that these particle types that are the result of consistent local

sources and not the diurnal process that formed the OC particles. The supermicron mass

concentrations and mass fractions (Figure 4.5b) are remarkably consistent during the 8

time periods over 2 different days. This confirms that most variation in particle number

and chemistry spatially and temporally were driven by changes in the submicron size

range [Finlayson-Pitts and Pitts, 2000].

121

Figure 4.5 Chemically-resolved relative mass fractions of particle types for (a) submicron (0.523 – 1 μm) and (b) supermicron (1-2.5 μm) particles at each site. The black diamonds represent the average ATOFMS mass concentration by site.

122

4.4.5 Differences in Particle Aging and Mixing State

The degree of aging of the different particle types provides another means of

determining the relative importance of temporal and spatial trends on the properties of the

aerosol and specifically the variation observed on the polluted Mar. 13th in comparison to

the background conditions observed on Feb. 12th. Mass spectral markers for chloride (-

35Cl-), nitrate (62NO3-), and sulfate (97HSO4

-) were used to provide information as to the

degree and form of particle aging. Chloride has been shown as a good marker for

unreacted sea salt particles, while nitrate has been shown as a marker for aging on many

particle types [Noble and Prather, 1997;Pratt and Prather, 2009a;Sullivan, et al., 2007].

Sulfate is identified by the 97HSO4- peak [Whiteaker and Prather, 2003], except for sea

salt particles where there is interference from the 23Na37Cl2- at m/z -97. Figure 4.6a shows

the average nitrate and chloride peak areas at each site for sea salt particles, while Figure

4.6b and 4.6c show nitrate and sulfate for K-Combustion and ECOC particles, the most

abundant submicron particle types, respectively. The clean Feb. 12th has higher chloride

peak area for sea salt at all 3 sites, while more polluted conditions on Mar. 13th have a

nitrate peak area that starts larger and decreases to roughly even with chloride during the

day (Figure 4.6a). This decrease in the amount of nitrate on sea salt particles follows a

temporal trend with aged particles being transported inland by the sea breeze and

replaced by fresh particles that become more abundant. There were not enough K-

Combustion and ECOC particles with negative mass spectra to analyze peak areas during

the clean Feb. 12th. On the more polluted Mar. 13th, nitrate area decreases during the day

as it did on sea salt, indicating that the combustion particles are fresher (Figure 4.6b and

123

Figure 4.6 Peak areas related to aging for (a) Sea salt particles, (b) K-Combustion particles, and (c) ECOC particles. For sea salt particles, chloride (blue – 35Cl-) and nitrate (green – 62NO3

-) are plotted. For K-Combustion and ECOC particles, nitrate and sulfate (purple – m/z 97HSO4

-) are plotted. The ratio of the peak areas for nitrate and sulfate (d) are given for K-Combustion and ECOC particle types.

124

6c). For K-Combustion, sulfate peak area becomes larger throughout the day on Mar.

13th. This is likely because the particles are fresher and had less time to react

heterogeneously and/or have ammonium nitrate and water condense (which can suppress

sulfate signal)[Neubauer, et al., 1998]. Figure 4.6d shows the ratio of the sulfate peak

area and nitrate peak area at each site during the day. The ratio increases throughout the

day for both particle types showing relatively less secondary nitrate as the aged aerosol is

transported inland and replaced by more fresh particles. The observations from the sea

salt and combustion particles (ECOC and K-Combustion) demonstrate that the secondary

marker nitrate follows a temporal pattern and occurs across the region without much

spatial variation.

4.4.6 Variations in Chemistry and Implications for Models and Regulations

The measurements described above show that even on days with relatively similar

meteorological conditions there can be considerable differences in particle mass and

number concentrations and subsequently differences in human exposure levels to

particles. These differences are driven by changes in particle sources and aging

determined by chemical mixing state within a marine-influenced urban atmosphere. On

Feb. 12th, similar low mass concentrations, low number concentrations, and particle

chemistry were observed around the San Diego Bay, providing a baseline of the urban

background for comparison to more polluted time periods. During polluted time periods,

transport of aged particles inland and replacement by fresher emissions was observed

homogeneously across the region driven sea breeze cleanout during the day of particles

that had been chemically processed overnight. Changes in the fraction of aerosols

spatially were used to identify downtown San Diego as the most concentrated source

125

region for the fresh submicron particles (mostly combustion) observed. Thus, by

understanding the changes to sources and aging from a chemical composition perspective

we can describe why variations in mass and number concentrations were observed.

The initial deployment of the mobile ATOFMS laboratory and single particle

measurements from this platform demonstrate the importance of understanding spatial

and temporal changes driven by particle chemistry as they relate to sources and aging

processes. Describing the urban aerosol from a single particle mixing state perspective

will be a key to improving future models that will analyze the links between aerosol

composition/sources, human exposure to these aerosols, and their broader impact on

public health. When crafting future regulations, the vast variation possible within similar

PM2.5 mass concentrations and the chemistry causing these changes needs to be

considered. Future high time resolution methods and measurements of spatial variation

across urban environments will need to be incorporated into more advanced models to

help bridge the causal gap between air pollution and human health.


Andrew Ault acknowledges the Unified Port of San Diego Environmental

Services Department for funding sampling with a graduate research grant and the is

grateful for a Department of Energy Global Change Education Program graduate research

environmental fellowship.. Joseph Mayer is thanked for his efforts with adapting the

trailer and installing the generators. Additional funding was provided by the California

Air Resources Board under contract 04-336. The authors gratefully acknowledge the

NOAA Air Resources Laboratory for the provision of the HYSPLIT transport model and

READY website used in this publication.

126

Chapter 4 is in preparation for submission to the Atmospheric Environment: A. P.

Ault, J. M. Creamean, and K.A. Prather, Mobile Laboratory Observations of Temporal

and Spatial Variability within the Coastal Urban Aerosol.

127

4.6 References

B. Asgharian, A model of deposition of hygroscopic particles in the human lung, Aerosol Science And Technology, 38 (9), 938-947, 2004.

A. P. Ault, M. J. Moore, H. Furutani and K. A. Prather, Impact of emissions from the Los


N. Bukowiecki, J. Dommen, A. S. H. Prevot, R. Richter, E. Weingartner and U.

Baltensperger, 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, 2002.

N. Bukowiecki, J. Dommen, A. S. H. Prevot, E. Weingartner and U. Baltensperger, Fine

and ultrafine particles in the Zurich (Switzerland) area measured with a mobile laboratory: an assessment of the seasonal and regional variation throughout a year, Atmospheric Chemistry And Physics, 3, 1477-1494, 2003.



M. R. Canagaratna, J. T. Jayne, D. A. Ghertner, S. Herndon, Q. Shi, J. L. Jimenez, P. J.

Silva, P. Williams, T. Lanni, F. Drewnick, K. L. Demerjian, C. E. Kolb and D. R. Worsnop, Chase studies of particulate emissions from in-use New York City vehicles, Aerosol Sci. Technol., 38 (6), 555-573, 2004.

J. Cyrys, M. Pitz, J. Heinrich, H. E. Wichmann and A. Peters, Spatial and temporal

variation of particle number concentration in Augsburg, Germany, Sci. Total Environ., 401 (1-3), 168-175, 2008.

P. F. DeCarlo, J. G. Slowik, D. R. Worsnop, P. Davidovits and J. L. Jimenez, Particle

morphology and density characterization by combined mobility and aerodynamic diameter measurements. Part 1: Theory, Aerosol Science And Technology, 38 (12), 1185-1205, 2004.

B. J. Finlayson-Pitts and J. N. Pitts, Chemistry of the Upper and Lower Atmosphere,

Academic Press, San Diego, CA, 2000. E. Gard, J. E. Mayer, B. D. Morrical, T. Dienes, D. P. Fergenson and K. A. Prather, Real-


128


measurements of the chemical composition of size-resolved particles during a Santa Ana wind episode, California USA, Atmospheric Environment, 35 (19), 3229-3240, 2001.

R. N. Harrison and A. M. Jones, Multisite study of particle number concentrations in

urban air, Environ. Sci. Technol., 39 (16), 6063-6070, 2005. V. Isakov, J. S. Touma and A. Khlystov, A method of assessing air toxics concentrations

in urban areas using mobile platform measurements, Journal Of The Air & Waste Management Association, 57 (11), 1286-1295, 2007.

P. I. Jalava, R. O. Salonen, A. S. Pennanen, M. S. Happo, P. Penttinen, A. I. Halinen, M.

Sillanpaa, R. Hillamo and M. R. Hirvonen, Effects of solubility of urban air fine and coarse particles on cytotoxic and inflammatory responses in RAW 264.7 macrophage cell line, 229 (2), 146-160, 2008.

C. E. Kolb, S. C. Herndon, B. McManus, J. H. Shorter, M. S. Zahniser, D. D. Nelson, J.

T. Jayne, M. R. Canagaratna and D. R. Worsnop, Mobile laboratory with rapid response instruments for real-time measurements of urban and regional trace gas and particulate distributions and emission source characteristics, Environmental Science & Technology, 38 (21), 5694-5703, 2004.

M. Krudysz, K. Moore, M. Geller, C. Sioutas and J. Froines, Intra-community spatial

variability of particulate matter size distributions in Southern California/Los Angeles, Atmos. Chem. Phys., 9 (3), 1061-1075, 2009.

J. F. Mejia, L. Morawska and K. Mengersen, Spatial variation in particle number size

distributions in a large metropolitan area, Atmos. Chem. Phys., 8 (5), 1127-1138, 2008.



K. Moore, M. Krudysz, P. Pakbin, N. Hudda and C. Sioutas, Intra-Community

Variability in Total Particle Number Concentrations in the San Pedro Harbor Area (Los Angeles, California), Aerosol Sci. Technol., 43 (6), 587-603, 2009.

A. Nel, Air pollution-related illness: Effects of particles, Science, 308 (5723), 804-806,

2005.

129

K. R. Neubauer, M. V. Johnston and A. S. Wexler, Humidity effects on the mass spectra of single aerosol particles, Atmospheric Environment, 32 (14-15), 2521-2529, 1998.

C. A. Noble and K. A. Prather, Real-time single particle monitoring of a relative increase

in marine aerosol concentration during winter rainstorms, Geophysical Research Letters, 24 (22), 2753-2756, 1997.

H. Orru, E. Teinemaa, T. Lai, T. Tamm, M. Kaasik, V. Kimmel, K. Kangur, E. Merisalu

and B. Forsberg, Health impact assessment of particulate pollution in Tallinn using fine spatial resolution and modeling techniques, Environmental Health, 8, 2009.

S. S. Park and A. S. Wex, Size-dependent deposition of particles in the human lung at

steady-state breathing, Journal Of Aerosol Science, 39 (3), 266-276, 2008. C. A. Pope, Mortality effects of longer term exposures to fine particulate air pollution:

Review of recent epidemiological evidence, Inhalation Toxicology, 19, 33-38, 2007.

C. A. Pope, J. B. Muhlestein, H. T. May, D. G. Renlund, J. L. Anderson and B. D. Horne,

Ischemic heart disease events triggered by short-term exposure to fine particulate air pollution, Circulation, 114 (23), 2443-2448, 2006.


effects, Angewandte Chemie-International Edition, 44 (46), 7520-7540, 2005. K. A. Pratt and K. A. Prather, Real-Time, Single-Particle Volatility, Size, and Chemical

Composition Measurements of Aged Urban Aerosols, Environmental Science & Technology, 43 (21), 8276-8282, 2009a.

K. A. Pratt and K. A. Prather, Real-time, single-particle volatility, size, and chemical

composition measurements of aged urban aerosols., Environmental Science & Technology, In Preparation, 2009b.

X. Qin, K. A. Pratt, L. G. Shields, S. M. Toner and K. A. Prather, Seasonal Comparisons

of the Single Particle Mixing State in Riverside, CA duing the SOAR 2005 Campaign Aerosol Science & Technology, in prep.



A. A. Rostami, Computational Modeling of Aerosol Deposition in Respiratory Tract: A

Review, Inhal. Toxicol., 21 (4), 262-290, 2009.

130



M. T. Spencer, L. G. Shields and K. A. Prather, Simultaneous measurement of the

effective density and chemical composition of ambient aerosol particles, Environmental Science & Technology, 41 (4), 1303-1309, 2007.


the atmospheric processing of Asian mineral dust, Atmospheric Chemistry And Physics, 7, 1213-1236, 2007.

A. Valavanidis, K. Fiotakis and T. Vlachogianni, Airborne Particulate Matter and Human

Health: Toxicological Assessment and Importance of Size and Composition of Particles for Oxidative Damage and Carcinogenic Mechanisms, Journal of Environmental Science and Health Part C-Environmental Carcinogenesis & Ecotoxicology Reviews, 26 (4), 339-362, 2008.

J. R. Whiteaker and K. A. Prather, Hydroxymethanesulfonate as a tracer for fog

processing of individual aerosol particles, Atmospheric Environment, 37 (8), 1033-1043, 2003.

T. Xia, M. Kovochich, J. Brant, M. Hotze, J. Sempf, T. Oberley, C. Sioutas, J. I. Yeh, M.

R. Wiesner and A. E. Nel, Comparison of the abilities of ambient and manufactured nanoparticles to induce cellular toxicity according to an oxidative stress paradigm, 6 (8), 1794-1807, 2006.

H. C. Yeh, R. G. Cuddihy, R. F. Phalen and I. Y. Chang, Comparisons of calculated

respiratory tract deposition of particles based on the proposed NCRP model and the new ICRP66 model, Aerosol Science and Technology, 25 (2), 134-140, 1996.

M. Zavala, S. C. Herndon, E. C. Wood, J. T. Jayne, D. D. Nelson, A. M. Trimborn, E.

Dunlea, W. B. Knighton, A. Mendoza, D. T. Allen, C. E. Kolb, M. J. Molina and L. T. Molina, Comparison of emissions from on-road sources using a mobile laboratory under various driving and operational sampling modes, Atmos. Chem. Phys., 9 (1), 1-14, 2009.

Y. M. Zhou, C. Y. Zhong, I. M. Kennedy, V. J. Leppert and K. E. Pinkerton, Oxidative

stress and NF kappa B activation in the lungs of rats: a synergistic interaction between soot and iron particles, Toxicology And Applied Pharmacology, 190 (2), 157-169, 2003.

5 Potential Changes to California Orographic Precipitation

Due to Transported Asian Dust

5.1 Synopsis

Aerosols impact the microphysical properties of clouds by serving as cloud

condensation nuclei (CCN) and ice nuclei (IN). By modifying cloud properties, aerosols

have the potential to alter the location and intensity of precipitation [Rosenfeld, et al.,

2008], but determining the magnitude and reproducibility of aerosol-induced changes

remains a significant challenge to experimentalists and modelers [Stevens and Feingold,

2009]. Efforts to characterize these changes are of vital importance as policymakers enact

regulations that will impact water management in regions where water is a limited

resource. Models and measurements suggest dust aerosols, which can be transported

around the globe on timescales of days [Uno, et al., 2009], can serve as effective

IN[DeMott, et al., 2003;VanCuren and Cahill, 2002] and impact mixed phase

clouds[Uno, et al., 2009;Yumimoto, et al., 2009]. Herein, results from the CalWater Early

Start field campaign combine detailed chemistry of the insoluble particle residues in

precipitation samples and back trajectory analysis with highly spatio-temporally resolved

precipitation and related meteorological observations at the surface and aloft. Two large

storms are compared with similar meteorology (e.g. 72 hour integrated horizontal water

vapor transport) that together produced 23% of the annual precipitation and 38% of the

maximum snowpack at California’s Central Sierra Snow Lab (CSSL). One storm with

high levels of Asian dust in the precipitation produced 1.4 times as much precipitation

and 1.6 times the increase in snow pack at CSSL. These uniquely comprehensive

131

132

measurements suggest that Asian dust, transported across the Pacific, became

incorporated into the upper altitudes of the precipitation-producing clouds over the Sierra

Nevada, and ultimately may have impacted precipitation amounts.

5.2 Introduction

Recent advances in observational and diagnostic methods for both aerosols and

orographic precipitation processes have created the opportunity to make further advances

on the challenging problem of quantifying aerosol-cloud-precipitation interactions. The

Sierra Nevada has seen a reduction of the snowpack in recent years, which has been

attributed to impacts of anthropogenic pollution on clouds and orographic precipitation

[Rosenfeld, et al., 2008]. In an effort to better understand the impacts of aerosols on

orographic precipitation in the Sierra Nevada [Rosenfeld, et al., 2008], the CalWater

campaign was designed to study of aerosols-meteorology-precipitation. The primary

measurement site at Sugar Pine Dam (SPD-1066 meters MSL) is ideally situated to

observe orographic precipitation associated with synoptic scale weather systems affecting

the west coast of the United States (Figure 5.1). Before investigating the potential role of

aerosols, it is critical to compare the two storms meteorologically.

133

Figure 5.1 The primary sampling site (Sugar Pine Dam – SPD, 39.129°N, 120.801°W, 1066 m ASL) and sites where supporting data were collected (Colfax, Bodega Bay – BBY, Sloughhouse – SHS, and Central Sierra Snow Lab - CSSL) are marked on a terrain base map of California and map inset. The different instruments used at the SPD site are shown in the accompanying photograph.

134

5.3 Experimental

5.3.1 Meteorological Methods

The meteorological features described above were detected by a unique network

of experimental meteorological observations aloft and at the surface, with at least hourly

time resolution. Composites of Special Sensor Microwave Imager(SSM/I) data from

polar orbiting satellites were used to measure integrated water vapor [Schluessel and

Emery, 1990] over the ocean. Atmospheric rivers were identified by IWV >2 cm IWV,

length >2000 km, and width <1000 km, criteria that have been previously established at

sea level [Neiman, et al., 2008]. A modified IWV threshold of >1.58 cm was used at

Colfax due to the elevation at the site (~ 644 meters). A 915-MHz wind profiler and

surface meteorological instrumentation at the Colfax, Sloughouse, and Bodega Bay sites

were analyzed to deduce mesoscale meteorological conditions in the vicinity of Sugar

Pine and are discussed in detail below. Over land, continuous 30-minute-resolution

measurements of IWV in the full atmospheric column above a dual-frequency, global

positioning system (GPS) receiver at Colfax, California were retrieved with high

accuracy (~1 mm) using the apparent delays in the arrival of radio signals transmitted by

a constellation of GPS satellites, as well as collocated surface temperature and pressure

measurements (Mattioli et al. 2007 and references therein[Mattioli, et al., 2007]). At

midlatitudes, the observing domain above a GPS receiver corresponds to an inverted cone

with a radius of ~11 km at 500 hPa.

At SPD the S-band profiling radar (S-PROF), a fixed dish antenna, was operated

at 2875 MHz and directed vertically to study the backscatter of energy from

135

hydrometeors and cloud droplets via Rayleigh scattering and to monitor the radar

brightband melting level (which approximates the freezing level) [White, et al., 2003]. A

Particle Size and Velocity (PARSIVEL) disdrometer measured the duration and

amplitude of the decrease in voltage of a continuous laser on a photodiode as a drop

passes through the beam to determine the drop’s fall speed and size, respectively.

Disdrometer data was used to determine drop size distributions at the surface of the

sampling site.

5.3.2 Precipitation Sampling

Precipitation was collected for Periods 1-8, snowfall was collected and melted for

Period 9, and mixed phase precipitation was collected and melted during Period 10.

Precipitation samples were analyzed chemically using a previously described method

[Holecek, et al., 2007]; the insoluble residues in the precipitation samples were

resuspended using a Collison nebulizer, dried using two silica gel diffusion driers, and

sampled by an aerosol time-of-flight mass spectrometer (ATOFMS) [Gard, et al., 1997].

The ATOFMS provides size and dual polarity mass spectra for individual particles

between 0.2-3.0 micrometers. Particles were classified into different types based on

characteristic peaks in the positive and negative mass spectra. Representative mass

spectra, as well as further details on the classification and labeling scheme are provided

below.

5.3.3 FLEXPART Model

The Lagrangian particle dispersion model FLEXPART [Stohl, et al., 2005]

version 8.1 was run with NOAA Global Forecast System 0.5° x 0.5° global data at 26

levels using SPD as the point source for 7-day backward air mass trajectory calculations

136

using the cluster analysis feature. Analyses at 0000, 0600, 1200, and 1800 UTC and

forecasts at 0300, 0900, 1500, and 2100 UTC were used. Calculations with releases of

400-500 particles were conducted every 3 hours during each storm from 500 m from 500

– 10000 meters, as well as at the mid-point in time for each precipitation sample.

Releases were also conducted at the bright band height and cloud echo top height as

determined from the S-PROF radar. Walker et al. (2009) have developed a high-

resolution (1-km) dust source database for Southwest Asia and East Asia [Walker, et al.,

2009]. The dust source database is based on topography, land use, atlases and ten years

of Dust Enhancement Product (DEP) imagery derived from MODIS. The MODIS DEP

imagery allows Walker et al. 2009 to identify the location of the small erodible area at the

head of each plume[Walker, et al., 2009]. This dataset is unique as it retains the 1-km

resolution of the satellite radiance data used to derive the dust source database. However,

for this paper, the data have been averaged to 81 km horizontal resolution.


5.4.1 Atmospheric Rivers Impacting California

The large contribution of the two extratropical cyclones described in this paper to

the annual rainfall in the Sierra Nevada was related to atmospheric river (AR) conditions

observed during both storms [Neiman, et al., 2008]. ARs transport water vapor into

California and modulate orographic precipitation. ARs lead to twice the daily average

precipitation of other storms in California during winter [Neiman, et al., 2008] and in

extreme cases lead to severe flooding [Ralph, et al., 2006], however they also increase

snowpack and fill reservoirs critical to the region’s water supply [Neiman, et al., 2008].

AR conditions for Storm 1 (February 22-24, 2009) and Storm 2 (March 1-4, 2009) are

137

evident from well-defined plumes of integrated water vapor (IWV) in special sensor

microwave imager (SSM/I) satellite imagery offshore (Figs. 5.2a and 5.2b) and in AR

observatory data at Bodega Bay on the coast, at Sloughhouse in the central valley, and at

Colfax in the Sierra foothills (Fig. 5.2). Though SPD is 250 km inland, AR conditions

(IWV >1.58 cm at 1066 meters) at the sampling site were confirmed using data from

nearby Colfax with wind profiling and global positioning system meteorology data (Figs.

5.2c and 5.3). AR conditions accounted for >90% of precipitation during both storms at

SPD. Overall, not only did both storms have similar orographic forcing by upslope winds

and similar synoptic-scale and local-scale IWV, but also remarkably similar horizontal

water vapor “bulk” fluxes. Nonetheless, the storms had significantly different

precipitation amounts due to other factors, possibly aerosol-cloud interactions.

138

Figure 5.2 SSM/I satellite images of IWV for a) Storm 1 on the afternoon of 22 February 2009, and b) Storm 2 on the afternoon of 1 March 2009. Both are composites of two SSM/I satellite overpasses during the descending phase of the orbits (North to South) between 1200 and 2359 UTC. The atmospheric rivers are best represented by the associated regions of IWV> 2 cm (green-to-yellow-to-orange), the threshold for atmospheric rivers at sea level. Gaps of satellite coverage are shown as grey areas. c) Time series of IWV (from Colfax; 13 km west of SPD) and precipitation (from SPD) spanning both storms.

139

Table 5.1 Atmospheric river characteristics for Storms 1 & 2 during the CalWater Early Start field campaign.

140

The term atmospheric river refers to a corridor of water vapor transported in the

lower level jet of an extratropical cyclone [Ralph, et al., 2004;Zhu and Newell, 1998].

These conditions account for > 90% of poleward water vapor transport in 10% of zonal

circumferences at midlatitudes [Neiman, et al., 2008;Zhu and Newell, 1998]. The

determination of atmospheric river conditions during the two extratropical cyclones was

based on a previously described methodology [Neiman, et al., 2008;Neiman, et al.,

2009;Ralph, et al., 2004]. Atmospheric rivers are typically long (>2000 km), narrow

(<1000 km) corridors of strong water vapor transport. A distinguishing characteristic of

atmospheric river conditions is the presence of large water vapor contents, as indicated

using a threshold for integrated water vapor (IWV), which is normally 2 cm, for

elevations near sea level [Ralph, et al., 2004]. The IWV threshold used in this paper for

data nearest the Sugar Pine Dam (SPD) field site (located at 1066 m MSL) was adjusted

to 1.58 cm based on analysis of atmospheric profile data from a rawinsonde launched

near sea level at Oakland, California, during Storm 2. The Oakland data indicated that

20% of the total IWV was contained in the lower atmosphere below the 644 m MSL

elevation of the Global Positioning System (GPS) receiver used to measure IWV at

Colfax, California. This correlates with previously measured IWV values during

dropsonde measurements over the Pacific Ocean [Ralph, et al., 2005], where roughly

23% was contained below 644 m MSL. The atmospheric rivers observed during

CalWater Early Start were of modest amplitude offshore and at the coast, with similar

IWV maxima (2.7 cm vs. 3.0 cm). Maximum vertically integrated water vapor transports

(IVT) were calculated from North American Regional Reanalysis (NARR) data and

found to be similar (625 kg/(s*m) vs. 675 kg/(s*m)) [Neiman, et al., 2008;Ralph, et al.,

141

2004]. Storm-total bulk water vapor flux was calculated from BBY ARO data over 72

hours for Storm 1 (starting at 0000 UTC 22 Feb 2009) and again for Storm 2 (starting at

0000 UTC 1 Mat 2009). These time windows were chosen since they encompass the

entirety of the precipitation observed in each Storm, and are of the same duration. The

72-h integrated bulk water vapor flux was nearly identical for the two storms (Table 5.1),

and yet the coastal mountain site (Cazadero) near BBY observed much more rain in the

72 h containing Storm 2 (22.9 cm) than in Storm 1 (10.2 cm).

Key tropospheric observational data were collected at Colfax, California (CFC),

which is located only ~13 km southwest of the chemistry sampling site at Sugar Pine.

Instrumentation at Colfax included a 915-MHz wind profiler for measuring hourly wind

profiles between ~0.1 and 4.0 km above ground with 100-m vertical resolution (see

Carter et al. 1995 for instrumentation details), a GPS receiver for recording IWV at 30-

min intervals, and a meteorological tower and tipping bucket for measuring standard

surface parameters every 2 min. Unless noted otherwise, the discussion here refers to the

analysis shown in Fig. 5.3 for the period 1 – 4 March 2009. This analysis combines

along-front isotachs (directed from 215°) and thermal-wind derived axes of warm and

cold advection below 4.5 km mean sea level (MSL), radar brightband melting levels,

hourly precipitation rates, and surface equivalent potential temperatures. More than 116

mm of precipitation fell at CFC during the three-day period.

142

Figure 5.3 Wind profiler and surface observations from Colfax, California (CFC) on 1-4 March 2009. (a) Time-height section of hourly wind profiles (wind flags = 25 m s-1, barbs = 5 m s-1, half-barbs = 2.5 m s-1), along-front isotachs (m s-1; black contours; directed from 215°), brightband melting level heights (black dots), and axes of thermal-wind derived (i.e., geostrophic) warm and cold advection (red and blue dashed lines, respectively; derivation technique described previously [Neiman and Shapiro, 1989]). Every other range gate is plotted, as is the passage of fronts near the surface (thin vertical dashed lines). (b) Time series of surface equivalent potential temperature (K) and hourly precipitation rate (mm h-1; 10-min averages). The total accumulation for the plotted time series is provided. Thin vertical dashed lines are as in panel (a).

143

The flow at CFC generally contained a southerly component in response to the

offshore position of an extratropical cyclone, although the surface flow was often

decoupled from that observed aloft. A warm front and two cold fronts migrated across

CFC. Geostrophic warm advection associated with the warm front descended from 4 km

MSL at 1000 UTC 1 March 2009 to 1 km MSL at 0230 UTC 2 March. The melting level

jumped by 300 m in response to the descending warm advection at 2200 UTC 1 March

and ~2 km MSL. The flow veered with time across the front, from southerly to

southwesterly above ~2 km MSL and from southeasterly to southerly below that altitude.

Steady precipitation accompanied the warm frontal descent between 1600 UTC 1 March

and 0200 UTC 2 March, after which the surface equivalent potential temperature

increased steadily. Strong southerly flow and descending bands of geostrophic warm

advection in the cyclone warm sector persisted from 0230 UTC 2 March until the first

cold frontal passage at 2100 UTC 2 March. During the warm-frontal descent and within

most of the warm sector, values of IWV exceeded the 1.58-cm threshold for atmospheric-

river conditions at the altitude of Colfax. The first cold front was characterized by an

abrupt wind shift from southerly to southwesterly below 2 km MSL, a decrease in

altitude of the melting level, a deep layer of geostrophic cold advection, and a pair of

rapid 4-K decreases in equivalent potential temperature at the surface. In addition,

significant precipitation rates flanked this cold-frontal passage, including an

instantaneous spike of 30 mm h-1 that marked a narrow cold-frontal rainband, which

forced strong upward air motion – creating the heavy precipitation. A second cold front

passed CFC at ~1830 UTC 3 March, during which the wind shifted from strong southerly

to weaker southwesterly below ~1.5 km MSL, the melting level dropped in altitude to ~ 1

144

km MSL, geostrophic cold advection was observed between 1.5 and 2.5 km MSL, and

the equivalent potential temperature decreased sharply by 5 K. Enhanced precipitation

straddled this cold-frontal passage.

5.4.2 Meteorology and Precipitation Properties for Storms 1 and 2

The meteorology and precipitation properties for both storms are shown in Fig.

5.4a-e (Storm 1) and 5.4f-j (Storm 2) including: a/f) precipitation residue chemistry, b/g)

rain rate and frontal passages, c/h) reflectivity vertical profiles and transport heights, d/i)

precipitation type, and e/j) drop size distributions at the surface. During both storms

precipitation was initially characterized by mid-to-upper-tropospheric upward motion and

associated "seeder" clouds leading to bright band (BB) rain [Martner, et al., 2008], which

shifted to a higher fraction of non-bright band (NBB) rain [Martner, et al., 2008] from

shallow clouds after warm frontal passage. Storm 1 ends with a high fraction of NBB rain

and decreased rain rates. Storm 2 has three notable differences from Storm 1: rain rate

increases after both cold fronts, precipitation type late in the storm shifts away from NBB

rain to BB, convective rain, and snow, and drop size distribution shifts to larger drops.

Many factors may have played a role in the differences between the storms, including

cold frontal forcing, freezing level changes, atmospheric instability, as well as a change

in aerosols incorporated in the clouds. Most non-aerosol-related precipitation processes

were driven by mesoscale meteorological processes that create upward air motion (and

condensation) that vary over time scales of hours and distances of tens of km and were

resolved by CalWater data. Below, we explore a third factor that may have played a role,

the entrainment of Asian dust at cloud top above the freezing level.

145

Figure 5.4 Time series of chemical and meteorological measurements during Storms 1 and 2. The fractions of particle types with different chemical compositions are shown for each period in Storm 1 (Fig. 5.4a) and Storm 2 (Fig. 5.4f). Period numbers with white text font denote periods with dust transport. The average precipitation rate during each sample period and the warm and cold frontal passages are shown for Storm 1 (Fig. 5.4b) and Storm 2 (Fig. 5.4g). A time-height cross section of radar reflectivity from the S-PROF radar is shown for Storm 1 (Fig. 5.4c) and Storm 2 (Fig. 5.4h), with the orange line circling altitudes with back trajectories from Asian dust source regions and the underlying yellow-green line circling back trajectories from Asia west of longitude 120E. The fraction of each type of precipitation (bright band (BB), non-bright band (NBB), convective (Conv.), Snow, and Insufficient (Insuf.) for each sample period (where snow indicates that the bright band is below the lowest radar range gate) is shown for Storm 1 (Fig. 5.4d) and Storm 2 (Fig. 5.4i). The surface drop size distribution (DSD) with size on the y-axis, number density (N(D) - pseudo-color background) and mean diameter (black line) are shown for Storm 1 (Fig. 5.4e) and Storm 2 (Fig. 5.4j).

146

5.4.3 Transported Asian Dust in Precipitation Samples

Aerosols that act as IN (i.e. unprocessed mineral dust) can be lofted by

extratropical cyclones into the upper troposphere [Stohl, et al., 2002] and transported

intercontinentally [Eguchi, et al., 2009]. This dust-enriched air can descend and become

entrained at cloud top heights, where even dust-free ambient air entrainment can increase

ice concentrations [Hobbs and Rangno, 1985]. The added presence of dust in entrained

ambient air has been modeled to induce an early initiation of the ice phase, increasing

riming rates, and precipitation efficiency for orographic clouds [Muhlbauer and

Lohmann, 2009]. The possibility that entrained dust may affect precipitation is supported

by earlier findings from cloud seeding experiments which show that descending dry air

over a shallow moist zone increases potential instability providing a mechanism for

strong evaporative cooling, ice nucleation [Hobbs and Rangno, 1985], and precipitation

enhancement [Rauber, et al., 1988].

To investigate the chemistry of aerosols incorporated into clouds and

precipitation, multiple precipitation samples were collected. Aerosol time-of-flight mass

spectrometry (ATOFMS) was used to determine the size and chemical composition of

individual resuspended insoluble residues [Gard, et al., 1997]. During Storm 1 and the

beginning of Storm 2 (Figs. 5.4a/5.4f), the most abundant residue observed in the

precipitation samples was an oxidized organic carbon particle type comprised of aromatic

organic species. These species most likely came from CCN active biomass burning

particles [Holecek, et al., 2007;Petters, et al., 2009] and are likely composed of water

soluble humic-like substances that can represent large fractions of organic carbon in

precipitation [Graber and Rudich, 2006]. During Storm 2, a dramatic shift occurred in

147

residue chemistry from a high number fraction of organic carbon residues to a high

fraction of mineral dust residues. Along with fractional increases, by the end of Storm 2

the precipitation had roughly an order of magnitude higher dust residue counts per

sampling hour than was observed for either Storm 1 or the beginning of Storm 2. These

mineral dust residues are characterized by aluminosilicate ions (59AlO2-, 60SiO2

-, 76SiO3-)

and inorganic metal ions (27Al+, 39K+, 48Ti+), likely phyllosilicates [Gallavardin, et al.,

2008] (OC/dust spectra in Fig. 5.5), and differences were observed in the dust mass

spectra between Storms 1 and 2. Storm 2 dust has tracer combinations that have been

observed in Asian soil samples [Jeong, 2008] and spectra are similar to Asian dust ice

residues measured by ATOFMS during aircraft-observations of mixed phase clouds

[Pratt, et al., 2009]. Ion patterns for Storm 1 particles were more representative of local

dust. Simultaneous ground-based measurements show very low concentrations of dust

particles were present at the end of Storm 2 (Table 5.2), suggesting scavenging was not

the source of residues, but instead that the dust residues were serving as IN.

Information on the different precipitation samples collected during Storms 1 & 2

is given in Table 5.2. Sample length varied based on site conditions and the fact that

collection was not an automated process. Future studies will incorporate standardized and

automated sample collection techniques to allow greater information on the precipitation

sample properties to be determined. During Storm 2 the number of precipitation residues

is observed to decrease considerably over the course of the storm, even if sample length

is normalized. This is due to the incorporation of anthropogenic aerosols both in-cloud

and during scavenging as drops fall towards the site and is typical of precipitation at the

beginning of storms [Pruppacher and Klett, 1997]. Precipitation type was classified into

148

4 primary types with half hour time resolution using methods developed that use S-Prof

data: bright band (BB), non-bright band (NBB), convective (Conv), and snow. During

some periods not enough data was collected to determine the type of precipitation and are

labeled as insufficient. Fractions of precipitation type for each sample are included as a

numerical reference due to the difficulty in determining these values from the visual

display in Figure 5.4. In addition the average and range of bright band heights and echo

top heights determined from the S-Prof radar are included.

149

Table 5.2 Precipitation sampling period information for the CalWater Early Start Campaign.

Precip Weather Precip NumConc Precip BB NBB Conv Insuf. Snow BB Ht** Echo Top

Period System Start (UTC) Stop (UTC) Residues (#/cm3) (mm)* (mm) (mm) (mm) (mm) (mm) (km‡) (km‡)

1 Storm1 02/22 19:30 02/23 18:45 403 180 50.3 31.5 17.8 0.0 1.0 0.0 1.2 3.92 Storm1 02/23 18:45 02/24 19:20 299 300 33.8 9.9 21.3 0.0 2.5 0.0 1.0 3.03 N/A 02/26 00:00 02/26 19:45 504 N/A 14.0 2.8 7.1 0.0 2.0 0.0 0.6 2.74 Storm2 03/01 16:00 03/02 01:30 6579 750† 22.6 0.8 0 0.0 0.0 0.0 1.1 6.45 Storm2 03/02 01:30 03/02 04:30 506 130 1.3 0 0 0.0 0.3 0.0 1.1 6.86 Storm2 03/02 05:20 03/02 20:20 751 280 19.1 0 7.4 0.0 1.5 0.0 1.2 3.37 Storm2 03/02 20:20 03/03 01:45 253 510 17.5 0 10.4 0.0 0.3 0.0 0.9 4.68 Storm2 03/03 05:20 03/03 18:20 549 410 27.7 13.7 2.5 2.5 0.0 7.9 0.4 3.69 Storm2 03/03 18:45 03/04 01:00 253 210 31.5 0 0 0.0 0.0 31.5 N/A 3.5

10 Storm2 03/04 01:00 03/04 12:00 82 260 40.9 0 0 0.0 0.5 40.6 N/A 3.4* HTB Precip Gauge to nearest 1/2 hour** For 30-min periods when >5 BB heights were detected† concentration from second half of sample only‡ heights in AGL

Sample Collection

150

Figure 5.5 ATOFMS mass spectra of a) organic carbon and b) mineral dust from precipitation samples.

151

Average mass spectra from Period 6 for the two most abundant residues observed

in the precipitation samples are shown in Figure 5.5. One was an organic carbon particle

type characterized by oxidized organic markers (43C2H3O+/43CHNO+ and 45COOH+) and

aromatics (77C6H5+ and 91C7H7

+) and the other a mineral dust particle type characterized

by aluminosilicate peaks (76SiO3-,

77HSiO3-,

59AlO2-, 60SiO2

-) in the negative ions and

inorganic metal peaks in the positive ions (27Al+, 39K+, 56Fe+, 48Ti+, 64TiO+, 23Na+, and

24Mg+). Period 6 was chosen as it represents a period with roughly equal numbers of

organic carbon and dust particles. Given the prevalence of the organic carbon particle

type and a relative lack of spectra corresponding to the standard ATOFMS biomass

burning signature, it is likely that these spectra are of the water soluble organic species

from biomass burning particles that have dissolved in the precipitation solution and are

resuspended independent of the potassium cores often found in biomass burning particles.

The water soluble oxidized organic carbon particle type has been shown previously in

ATOFMS analysis of precipitation samples [Holecek, et al., 2007] and has been

supported by recent single particle x-ray microscopy showing organic carbon mixed with

inorganic cores in urban outflow [Moffet, et al., 2010]. The dust spectrum shown in

Figure 5.5b has many characteristic markers of dust observed in previous ATOFMS

measurements [Sullivan, et al., 2007]. The combination of these markers indicates that

these dust particles are likely phyllosilicates, such as ilite or kaolinite [Gallavardin, et al.,

2008].

152

5.4.4 Back Trajectory Analysis

The Asian continent represents a large global source of IN due to its increasingly

severe dust storms [Solomon, et al., 2007]. Studies have shown that transported Asian

dust becomes incorporated into ice clouds over North America [Pratt, et al., 2009;Sassen,

2002]. For mixed phase clouds, enhancement of ice concentrations is predicted to occur

primarily at cloud top [Hobbs and Rangno, 1985] by glaciating supercooled droplets or

acting as IN in water sub-saturated and ice saturated temperature regimes [Muhlbauer

and Lohmann, 2009]. To investigate the source of the dust residues observed in the

precipitation samples, the Lagrangian model FLEXPART [Stohl, et al., 2005] was used to

calculate back trajectories by releasing particles at altitudes above SPD and calculating

their trajectories backward in time through gridded-meteorological-fields. S-PROF

precipitation profiler data allowed these back trajectories to be initiated from altitudes

that corresponded to observed precipitation radar echo top heights (physical cloud-top

height is above the radar echo-top height). Fig. 5.6 shows 7-day average backward

trajectory calculations for particles released at the average precipitation echo-top height

at SPD for the mid-time bin of each precipitation sample during both storms. The

precipitation echo-top heights are within the range of previously reported heights (2 – 10

km) for dust transport [Eguchi, et al., 2009]. For Storm 1 and the beginning of Storm 2,

backward trajectories do not show significant trans-Pacific dust transport. However, later

in Storm 2 trajectories shift to periods of long range transport occurring over 5-7 days

(Periods 6-10), consistent with previously reported timescales [Holzer, et al., 2005].

153

Figure 5.6 FLEXPART average backward trajectories calculated from the midpoint in time of each precipitation sampling period with particles released at the average cloud echo top height of that period. a) The above ground level (AGL) altitude for each trajectory starting at the site (0 hours) and markers representing each day backward in time from right to left. b) The back trajectories on a latitude (y-axis) and longitude (x-axis) grid with markers each day. Dust source regions are marked in brown as determined from the Navy Aerosol Analysis Prediction System (NAAPS). The observing periods in both panels have the same color coding.

154

To obtain a greater understanding of the transport conditions, a series of back

trajectories were calculated every 3 hours during each storm in 500 meter increments.

Each back trajectory was checked to see whether it passed over Asia or the climatological

source region for Asian dust below 6 km AGL within the previous 7 days. Back

trajectories originating in Asia are superimposed on S-PROF radar images in Figs.

5.4c/5.4h. Backward trajectories specifically crossing dust source regions in Asia below

6 km AGL are also shown. Minimal overlap occurs between trajectories and clouds

during Storm 1 and the beginning of Storm 2. However, during the latter half of Storm 2,

the overlap between the precipitating clouds and the trajectories indicative of dust

transport implies that mineral dust was most likely incorporated well above the melting

level in the ice phase portion of precipitating mixed phase clouds, which is confirmed by

the residue chemistry (Fig. 5.4f). Additionally, transport conditions did not occur below

cloud, which is consistent with the absence of dust in ambient sampling at the surface,

and eliminates scavenging as a source of dust in the precipitation samples. While

uncertainties exist in the back trajectories, the concurrence of three sources of

information (trajectories, S-PROF radar observations, and measurements of insoluble

residues within the precipitation) provides strong evidence of the incorporation of Asian

dust in the clouds and precipitation.

155

Figure 5.7 Schematic summary of Storm 2 case study. a) Transport path from the Asian climatological dust source region (brown), across the Pacific, to the Sierra Nevada Mountains of western North America. The path is derived from back trajectories calculated using FLEXPART ending near cloud top over the SPD observing site. Transport occurred between 2- 8 km MSL, over a roughly 7 day period between 26 February and 4 March 2009. b) Time-height cross-section of observed and derived conditions at the SPD observing site over roughly 72 hours, showing the overlap of observed clouds and precipitation with the times and altitudes for which back trajectories indicate Asian dust is likely present aloft. ATOFMS diagnostics documented Asian dust in the surface precipitation during this same time period. Precipitating clouds (green shaded) and the radar brightband melting level (black dotted line) are derived from S-Prof radar observations. Fronts and atmospheric river conditions are diagnosed from wind profiler and GPS-met observations at the nearby Colfax site. The overall pattern summarized in (b) is similar to the conditions that Rauber et al. (1988) identified as having the greatest potential for aerosol cloud seeding [Rauber, et al., 1988].

156

5.4.5 Conclusions

Herein, the first results are shown detailing the incorporation of transported Asian

dust into orographic precipitation during a synoptic scale weather system impacting

North America. A proposed mechanism of how dust might impact precipitation (Fig. 5.7)

shows a) long range dust transport and b) incorporation of dust at cloud top over regions

identified by earlier cloud seeding experiments as susceptible to precipitation

enhancement by aerosol entrainment [Rauber, et al., 1988]. This figure reflects the

intertwined nature of dynamics, transport conditions, cloud properties, and aerosols in the

atmosphere, leading to simultaneous changes in cloud and precipitation processes.

Previously, when a change occurred in the rate of precipitation, one had to hypothesize

which factors caused the change. Herein, a unique combination of aerosol-precipitation

measurements with highly spatio-temporally resolved meteorology methods allowed for

measurement of each aspect of the complex aerosol-cloud-precipitation interplay.

Different cloud properties and precipitation amounts were observed between two major

storms, one of which had a dramatic shift in aerosol sources and chemistry, impacting the

precipitating clouds due to the transport of Asian dust to cloud-top height over the Sierra

Nevada. We hypothesize that the observed shift to Asian dust affected the formation and

rate of precipitation within these rapidly changing meteorological conditions, however

much work remains to document these potential impacts definitively. This study provides

an initial glimpse into the role of different aerosol sources in precipitation processes and

sets the stage for future long term studies focused on aerosol, cloud, and meteorological

measurements. These will be explored in detail during subsequent CalWater field seasons

which will capture many more precipitation events. The susceptibility of California’s

157

water budget to any shifts in precipitation demonstrates the need to verify and quantify

the relationship between aerosols, clouds, and precipitation.


Funding was provided by the California Energy Commission under contract

UCOP/CIEE C-09-07. A.P. Ault has been funded in part by a U.S. Department of

Energy Global Change Education Program Graduate Research Environmental

Fellowship. J. Mayer, M. Zauscher, and M. Moore provided assistance with UCSD/SIO

equipment set-up. The deployment of the NOAA and UCSD/SIO equipment at the Sugar

Pine site involved many field staff, particularly C. King (NOAA/ESRL/PSD). C. Johnson

(UCSD) assisted with FLEXPART model calculations. Ms. Annette Walker and Dr.

Ming Liu of Naval Research Laboratories provided the dust source region information

from the Navy Aerosol Analysis and Prediction System (NAAPS).

Chapter 5 is preparation for submission to Nature Geoscoence. A. P. Ault, C. R.

Williams, A. B. White, P. J. Neiman, J. M. Creamean, C. J. Gaston, F. M. Ralph, and

K.A. Prather, Observations of Transported Asian Dust in California Precipitation during

CalWater.

158

5.6 References

P. J. DeMott, K. Sassen, M. R. Poellot, D. Baumgardner, D. C. Rogers, S. D. Brooks, A. J. Prenni and S. M. Kreidenweis, African dust aerosols as atmospheric ice nuclei, Geophys. Res. Lett., 30 (14), 2003.

K. Eguchi, I. Uno, K. Yumimoto, T. Takemura, A. Shimizu, N. Sugimoto and Z. Liu,

Trans-pacific dust transport: integrated analysis of NASA/CALIPSO and a global aerosol transport model, Atmospheric Chemistry And Physics, 9 (9), 3137-3145, 2009.

S. Gallavardin, U. Lohmann and D. Cziczo, Analysis and differentiation of mineral dust

by single particle laser mass spectrometry, International Journal Of Mass Spectrometry, 274 (1-3), 56-63, 2008.



E. R. Graber and Y. Rudich, Atmospheric HULIS: How humic-like are they? A

comprehensive and critical review, Atmospheric Chemistry And Physics, 6, 729-753, 2006.

P. V. Hobbs and A. L. Rangno, Ice Particle Concentrations in Clouds, Journal Of The

Atmospheric Sciences, 42 (23), 2523-2549, 1985. J. C. Holecek, M. T. Spencer and K. A. Prather, Analysis of rainwater samples:

Comparison of single particle residues with ambient particle chemistry from the northeast Pacific and Indian oceans, Journal Of Geophysical Research-Atmospheres, 112 (D22), 2007.

M. Holzer, T. M. Hall and R. B. Stull, Seasonality and weather-driven variability of

transpacific transport, J. Geophys. Res.-Atmos., 110 (D23), 2005. G. Y. Jeong, Bulk and single-particle mineralogy of Asian dust and a comparison with its

source soils, J. Geophys. Res.-Atmos., 113 (D2), 2008. B. E. Martner, S. E. Yuter, A. B. White, S. Y. Matrosov, D. E. Kingsmill and F. M.

Ralph, Raindrop size distributions and rain characteristics in California coastal rainfall for periods with and without a radar bright band, J. Hydrometeorol., 9 (3), 408-425, 2008.

V. Mattioli, E. R. Westwater, D. Cimini, J. C. Liljegren, B. M. Lesht, S. I. Gutman and F.

J. Schmidlin, Analysis of radiosonde and ground-based remotely sensed PWV

159

data from the 2004 North Slope of Alaska Arctic Winter Radiometric Experiment, Journal Of Atmospheric And Oceanic Technology, 24 (3), 415-431, 2007.

R. C. Moffet, T. R. Henn, A. V. Tivanski, R. J. Hopkins, Y. Desyaterik, A. L. D.

Kilcoyne, T. Tyliszczak, J. Fast, J. Barnard, V. Shutthanandan, S. S. Cliff, K. D. Perry, A. Laskin and M. K. Gilles, Microscopic characterization of carbonaceous aerosol particle aging in the outflow from Mexico City, Atmospheric Chemistry And Physics, 10 (3), 961-976, 2010.

A. Muhlbauer and U. Lohmann, Sensitivity Studies of Aerosol-Cloud Interactions in

Mixed-Phase Orographic Precipitation, Journal Of The Atmospheric Sciences, 66 (9), 2517-2538, 2009.

P. J. Neiman, F. M. Ralph, G. A. Wick, J. D. Lundquist and M. D. Dettinger,

Meteorological characteristics and overland precipitation impacts of atmospheric rivers affecting the West Coast of North America based on eight years of SSM/I satellite observations, Journal of Hydrometeorology, 9 (1), 22-47, 2008.

P. J. Neiman and M. A. Shapiro, Retrieving Horizontal Temperature Gradients and

Advections from Single-Station Wind Profiler Observations, Weather And Forecasting, 4 (2), 222-233, 1989.

P. J. Neiman, A. B. White, F. M. Ralph, D. J. Gottas and S. I. Gutman, A water vapour

flux tool for precipitation forecasting, Proc. Institution of Civil Engineers – Water Management, 162 (2), 83-94, 2009.

M. D. Petters, C. M. Carrico, S. M. Kreidenweis, A. J. Prenni, P. J. DeMott, J. L. Collett

and H. Moosmuller, Cloud condensation nucleation activity of biomass burning aerosol, J. Geophys. Res.-Atmos., 114, 2009.


Twohy, A. J. Prenni and K. A. Prather, In situ detection of biological particles in cloud ice-crystals, Nature Geoscience, 2 (6), 397-400, 2009.

H. R. Pruppacher and J. D. Klett, Microphysics of Clouds and Precipitation, 954 pp.,

Kluwer Academic Publishers, Norwell, Massachusetts, U.S.A, 1997. F. M. Ralph, P. J. Neiman and R. Rotunno, Dropsonde observations in low-level jets over

the northeastern Pacific Ocean from CALJET-1998 and PACJET-2001: Mean vertical-profile and atmospheric-river characteristics, Mon. Weather Rev., 133 (4), 889-910, 2005.

F. M. Ralph, P. J. Neiman and G. A. Wick, Satellite and CALJET aircraft observations of

atmospheric rivers over the eastern north pacific ocean during the winter of 1997/98, Monthly Weather Review, 132 (7), 1721-1745, 2004.

160

F. M. Ralph, P. J. Neiman, G. A. Wick, S. I. Gutman, M. D. Dettinger, D. R. Cayan and

A. B. White, Flooding on California's Russian River: Role of atmospheric rivers, Geophysical Research Letters, 33 (13), 2006.

R. M. Rauber, R. D. Elliott, J. O. Rhea, A. W. Huggins and D. W. Reynolds, A

Diagnostic Technique For Targeting During Airborne Seeding Experiments In Wintertime Storms Over the Sierra Nevada, Journal Of Applied Meteorology, 27 (7), 811-828, 1988.

D. Rosenfeld, W. L. Woodley, D. Axisa, E. Freud, J. G. Hudson and A. Givati, Aircraft

measurements of the impacts of pollution aerosols on clouds and precipitation over the Sierra Nevada, J. Geophys. Res.-Atmos., 113 (D15), 2008.

K. Sassen, Indirect climate forcing over the western US from Asian dust storms,

Geophys. Res. Lett., 29 (10), 2002. P. Schluessel and W. J. Emery, Atmospheric Water-Vapor Over Oceans From SSM/I

Measurements, International Journal of Remote Sensing, 11 (5), 753-766, 1990. S. Solomon, D. Qin, M. Manning, Z. Chen, M. Marquis, K. B. Averyt, M. Tignor and H.


B. Stevens and G. Feingold, Untangling aerosol effects on clouds and precipitation in a

buffered system, Nature, 461 (7264), 607-613, 2009. A. Stohl, S. Eckhardt, C. Forster, P. James and N. Spichtinger, On the pathways and

timescales of intercontinental air pollution transport, J. Geophys. Res.-Atmos., 107 (D23), 2002.

A. Stohl, C. Forster, A. Frank, P. Seibert and G. Wotawa, Technical note: The

Lagrangian particle dispersion model FLEXPART version 6.2, Atmos. Chem. Phys., 5, 2461-2474, 2005.





161

R. A. VanCuren and T. A. Cahill, Asian aerosols in North America: Frequency and concentration of fine dust, J. Geophys. Res.-Atmos., 107 (D24), 2002.

A. L. Walker, M. Liu, S. D. Miller, K. A. Richardson and D. L. Westphal, Development

of a dust source database for mesoscale forecasting in southwest Asia, J. Geophys. Res.-Atmos., 114, 2009.

A. B. White, P. J. Neiman, F. M. Ralph, D. E. Kingsmill and P. O. G. Persson, Coastal

orographic rainfall processes observed by radar during the California land-falling jets experiment, J. Hydrometeorol., 4 (2), 264-282, 2003.

K. Yumimoto, K. Eguchi, I. Uno, T. Takemura, Z. Liu, A. Shimizu and N. Sugimoto, An

elevated large-scale dust veil from the Taklimakan Desert: Intercontinental transport and three-dimensional structure as captured by CALIPSO and regional and global models, Atmos. Chem. Phys., 9 (21), 8545-8558, 2009.

Y. Zhu and R. E. Newell, A proposed algorithm for moisture fluxes from atmospheric

rivers, Monthly Weather Review, 126 (3), 725-735, 1998.

6 Characterization of Asian Outflow at Gosan, Korea by

Single Particle Mass Spectrometry

6.1 Synopsis

Frequent intercontinental transport of mineral dust and anthropogenic pollution

from Asia impact the transfer of solar radiation through the atmosphere, alter cloud

properties, and affect aerosol concentrations and composition in North America and

beyond. Models currently use simplified schemes to account for the aging of these

outflow aerosols as they are transported globally, but the current parameterizations do not

account for the complex processes associated with particle aging. To improve our

understanding of these processes, real-time single-particle measurements of the size and

chemistry of atmospheric particles were made at the Gosan supersite on Jeju Island,

Korea. This study provided insight into the chemical mixing state of outflow particles

after ~1 day of transport over the ocean from China and Korea. The main contributors to

high concentrations of submicron particles were carbonaceous combustion particles,

while supermicron particles were dominated by sea salt and dust during transport periods.

Internal mixing with secondary species was dependent on primary particle chemistry and

size, wherein nitrate was primarily mixed with supermicron dust and sea salt and

ammonium sulfate was primarily found internally mixed with submicron carbonaceous

combustion particles. While considerable complexity exists in the aging of these outflow

particles, incorporating even simple empirically-based parameterizations of the impact of

aging on particle mixing states may improve model results significantly beyond those

obtained with the current time-based methods.

162

163

6.2 Introduction

Significant air pollution outflow from Asia has been observed for several decades

[Duce, et al., 1980;Hobbs, et al., 1971]. Aerosols transported from Asia have been shown

to play a major role in impacting climate directly by scattering and absorbing solar

radiation [Ramanathan and Carmichael, 2008] and indirectly by altering cloud properties

[Feng and Ramanathan, 2010;Sassen, 2002]. Downwind of Asia, the outflow can

significantly contribute to PM2.5 (particulate matter <2.5 μm) mass concentrations on the

west coast of North America [VanCuren and Cahill, 2002;Yu, et al., 2008]. Atmospheric

particles from Asia have even been shown to circumnavigate the globe under the proper

conditions [Uno, et al., 2009]. This transport is most prevalent during intense spring dust

storms [Holzer, et al., 2005] and can occur both in the boundary layer as well as the free

troposphere [Hara, et al., 2009] after particles have been injected at high altitudes by

warm conveyor belts [Eckhardt, et al., 2004] or frontal and postfrontal convection [Yu, et

al., 2008]. These Asian outflow plumes can range from primarily dust [Eguchi, et al.,

2009] to primarily anthropogenic sources [Yu, et al., 2008], including mixtures,

depending on meteorology and air mass history [Eguchi, et al., 2009;Hara, et al., 2009].

Simultaneous plumes have also been observed at different altitudes with one plume

containing primarily dust at one altitude and a mixture of dust and pollution in another

plume at a different altitude [Eguchi, et al., 2009]. The impacts of the Asian outflow on

climate are predicted increase in the future [Carmichael, et al., 2009] as the intensity of

dust storms likely increase [Solomon, et al., 2007] and population growth leads to

expanded anthropogenic emissions in the region [Carmichael, et al., 2009].

164

The rural super-site at Gosan on Jeju Island in Korea provides an ideal location to

observe the properties of particles following 1-2 days of transport from the east coast of

the Asian continent [Sahu, et al., 2009]. At Gosan, changes in radiative forcing have been

measured due to increased scattering and absorption during dust storms [Kim, et al.,

2006b;Kim, et al., 2005;Nakajima, et al., 2007]. In addition to aerosol loading, relative

humidity has been also shown to alter radiative forcing [Yoon and Kim, 2006], motivating

studies measuring the hygroscopicity and cloud condensation nucleating properties of

aerosols at Gosan [Kim, et al., 2006a;Kuwata, et al., 2008;Yum, et al., 2007]. Most

aerosol chemistry studies at Gosan have used bulk techniques, including photoacoustic

measurements of black carbon [Kim, et al., 2006b;Sahu, et al., 2009], ion

chromatography [Kim, et al., 2006b;Sahu, et al., 2009;Topping, et al., 2004], and aerosol

mass spectrometry [Topping, et al., 2004]. The many particle types with different

chemical composition observed at the Gosan site have been studied in small numbers of

particles through offline X-ray microscopy single particle analysis techniques [Ro, et al.,

2001]. Elevated concentrations of secondary species (eg. sulfate, nitrate) indicative of

atmospheric aging have been measured at Gosan and other island sites off the east coast

of Asia [Carmichael, et al., 1996;Chen, et al., 1997;Nishikawa, et al., 1991] with lower

concentrations of secondary species measured during time periods of marine influence, as

expected [Kim, et al., 2002]. Due to the large climate impact of Asian outflow aerosol

numerous campaigns have examined Asian outflow [Ramanathan, et al., 2001] [Russo, et

al., 2003] [Bates, et al., 2004] [Ramana and Ramanathan, 2006] [Nakajima, et al., 2007]

[Dunlea, et al., 2009] [Stith, et al., 2009], including many studies at the Gosan super-site.

165

However, for all that has been learned about the Asian outflow one significant

challenge that remains is accurately modeling how outflow particles evolve during

transport. In fact, on a broader level effects accurately incorporating particle mixing state

into climate models has been identified as one of the main steps needed to improve global

climate models [Ghan and Schwartz, 2007]. This is because significant uncertainties

remain in our understanding of particle aging processes and timescales [Riemer, et al.,

2010]. For example, with black carbon understanding aging processes has important

implications for the timescales on which hydrophobic particles will have taken up enough

hydrophillic secondary species to act as cloud condensation nuclei (CCN). In global

climate models this aging is frequently incorporated as a single aging timescale (τ) used

to move particles from the fresh/hydrophobic classification to the aged/hydrophillic

classificant where they can act as CCN [Croft, et al., 2005]. How this aging

parameterization is formulated is critical to the final model results [Croft, et al.,

2005;Koch, 2001] and using a single aging parameter has been shown to oversimplify

aging processes leading to incorrect assessments of the global black carbon burden

[Riemer, et al., 2003]. More thorough treatments of aging processes from a single particle

perspective are being developed [Riemer, et al., 2010;Riemer, et al., 2009], which have

shown urban plume aging timescales to vary from less than an hour to multiple days

depending on conditions. Considerable work remains to provide measurements that can

constrain and improve the treatment of aging in these models, with the ultimate goal of

providing a more complex yet computationally feasible method for use in global climate

models. The PACDEX ground sampling discussed below described measurements of the

single particle size and chemical composition of Asian outflow particles in spring at the

166

Gosan supersite. Particles at the site have been transported over the ocean in air masses

from different source regions for 1-2 days providing an ideal setup to directly investigate

single particle mixing and aging processes in Asian outflow.

6.3 Experimental

6.3.1 Location and Instrumentation

This study was conducted as part of the Pacific Dust Experiment (PACDEX) at

the Gosan super-site (33.29° N 126.16° E) in Korea from April 12-May 15, 2007. The

sampling site was on the western coast of Jeju Island 20 meters from the ocean at an

elevation of 35 meters with few major local sources and primarily onshore winds. The

external sampling line was ~0.3 meters in diameter and extended 40 meters above the

research station. Instruments connected to a sampling manifold with 10 parallel ports

included a condensation particle counter (CPC, 1 L/min), aerodynamic particle sizer

(APS, 5 L/min), aethalometer (4 L/min), nephelometer (17 L/min), aerosol time-of-flight

mass spectrometer (ATOFMS, 1 L/min), and a pump to pull the remaining flow, which

was adjusted as needed to maintain a consistent flow of 28.3 L/min through the PM10

cyclone (URG). The CPC (TSI Inc., Model 3010) measured particle number

concentration for particles >0.010 μm, and the APS (TSI Inc., Model 3321) provided

aerosol size distributions from 0.523-10 μm. The aethalometer (Magee Scientific, Model

AE31) measured black carbon (BC) mass concentrations. The scanning mobility particle

sizer measurements (SMPS, TSI Inc., Model 3034) were made for 0.010-0.487 μm

diameter particles on a separate sampling line. PM10 mass concentrations were obtained

from the co-located Korean Meteorological Association (KMA) site.

167

6.3.2 Single Particle Mass Spectrometry

An ATOFMS, described in detail previously [Gard, et al., 1997], measured the

size and chemical composition of individual particles in real-time. Briefly, particles are

introduced into the ATOFMS through a converging nozzle inlet into a differentially

pumped vacuum chamber, during which particles are accelerated to a terminal velocity.

The particles pass through two continuous wave lasers (diode pumped Nd:YAG lasers

operating at 532 nm) located 6 cm apart. Particle velocity is used to determine vacuum

aerodynamic diameter by calibration with polystyrene latex spheres of known size. The

sized particles are desorbed and ionized by a 266 nm Q-switched Nd:YAG laser (1.2-1.4

mJ). Positive and negative ions from the each individual particle are detected using a

dual-reflectron time-of-flight mass spectrometer.

Particle size and mass spectral data were imported into MatLab 6.5.1 (The

Mathworks Inc.) and analyzed utilizing YAADA 1.2 (www.yaada.org), a toolkit

developed for ATOFMS data analysis. Particles were grouped via two methods: 1)

searches of mass spectral, aerodynamic size, and temporal information and 2) clustering

via an adaptive resonance theory based neural network algorithm (ART-2a) at a vigilance

factor of 0.8 [Song, et al., 1999]. General particle types are defined by characteristic

chemical species; these labels do not reflect all of the species present within a specific

particle type. Peak identifications correspond to the most probable ions for a given m/z

ratio.

ATOFMS counts were scaled to number and mass concentrations using APS size-

resolved number concentrations, a method shown previously to yield quantitative mass

concentrations [Ault, et al., 2009;Qin, et al., 2006]. Briefly, ATOFMS counts were

168

binned by diameter for the APS size bins (0.5-2.5 m), and scaling factors were

determined for each hour and size bin to account for ATOFMS transmission biases [Qin,

et al., 2006]. After scaling the number concentrations, previous ATOFMS measurements

of particle density were used to convert from number to mass concentrations [Moffet and

Prather, 2009;Moffet, et al., 2008]. Bins of scaled concentrations were summed to

provide PM1 (PM <1 m), PM2.5 (PM<2.5 m), and PM1-2.5 (PM 1.0-2.5 m in

diameter). Scaled ATOFMS PM2.5 mass concentrations have been shown previously to

be well correlated with standard mass measurements, including the beta attenuation

monitor (BAM) and micro-orifice uniform deposit impactor (MOUDI) [Qin, et al., 2006].

ATOFMS scaled PM1 concentrations have been shown to track mass concentrations [Liu,

et al., 2000;Qin, 2007].


6.4.1 Air Mass Characteristics and Meteorology

Air mass back trajectories, calculated using HYSPLIT (Figure 6.1), for the Gosan

sampling site over the 5 week study agree with measured meteorological data (Figure

6.2). The most commonly observed air masses back trajectories were grouped into five

categories (China, Korea, Sea of Japan, Marine, and Local/Transition) (Figure 6.1). The

amount of time each air mass type was observed at the sampling site and the time the air

masses were over the ocean before reaching the site is shown in Table 6.1. The

trajectories start at 500 meters above the sample site and go 72 hours backward (each

marker represents 24 hours) and are representative of other heights in the boundary layer

based on comparisons at 200 and 1000 meters. Back trajectories were defined as those

that passed over the East China Sea and eastern China; these “China” trajectories had an

169

average transport time over the ocean of 20 hours from China to the sampling site, similar

to previously calculated transport times [Sahu, et al., 2009]. Air masses passing over

Northeastern China, the Yellow Sea and/or Korean peninsula before arriving at Gosan

(labeled Yellow Sea/Korea) traveled an average of 28 hours over the ocean before

reaching the sampling site. Marine air masses were defined as those that did not pass over

land during the 72 hour back trajectories. “Sea of Japan” air masses traveled east over

Manchuria in northeastern China before looping southwest over the Sea of Japan to the

sampling site (avoiding the Korean peninsula); these air masses traveled an average of 44

hours over the ocean. The most prevalent air mass type was from the Yellow Sea/Korea

region (32% of sampling time), followed by China (25%), Marine (11%), and the Sea of

Japan air mass (5%). Local/transition periods, defined as those affected by local sources

or air masses transitioning between two source regions, account for the remaining 27% of

sampling time. Air mass back trajectories previously observed at the Gosan site have

matched the 4 primary air mass types described above [Carmichael, et al., 1997;Kim, et

al., 2007;Kim, et al., 2006a;Kim, et al., 2005;Ro, et al., 2001;Wong, et al., 2007].

To determine which of these air mass types contribute to long range transport, 5

day forward trajectories were calculated at 500 meters for each of time periods defined

above (Figure 6.3).Of the four air mass types, China and Marine were indicative of long

range transport and Yellow Sea/Korea and Sea of Japan did not represent long range

transport conditions. The China and Marine air masses were lofted (1000-9000 meters) 1-

2 days east of the sampling site and reach the west coast of North America in less than

five days. In contrast, forward trajectories for the Korea and Sea of Japan air masses do

not travel beyond eastern Asia. Thus, understanding the differences in particle

170

concentrations and chemical mixing state between China and the Marine air masses is

critical to understanding the Asian outflow that reaches North American and beyond.

171

Figure 6.1 500 m, 72 hr HYSPLIT air mass back trajectories every 2 hrs (with 24 hour markers) are shown for representative time periods during the five main air mass categories. Air masses passing over eastern China and the East China Sea (April 25, 2007 15:00 – April 27, 2007 12:00) are marked in red, air masses passing over the Yellow Sea/Korea (April 17, 2007 12:00 – April 19, 2007 04:00), marine based air masses (April 20, 2007 19:00 – April 22, 2007 01:00), air masses passing over the Yellow Sea and NE China (April 17, 2007 12:00 – April 18, 2007 10:00), air masses passing over the Sea of Japan and Manchuria in NE China (April 23, 2007 8:00 – April 24, 2007 20:00), and local and/or transitioning back trajectories (April 16, 2007 6:00 – 18:00).

172

Figure 6.2 Meteorological data including: a) Relative humidity (blue) and temperature (green), b) Wind speed (red) and wind direction (black), and c) the Hysplit periods during the study.

173

Table 6.1 The fraction of sampling time represented by each air mass type and the average time (standard deviation) of the trajectory over ocean for the representative air mass categories shown in Figure 6.1.

Hysplit Air Mass Sampling Fraction of Time Over Back Trajectory Hours Sampling Time Ocean (Hrs)

Eastern China 196 0.25 20 ± 5YellowSea/Korea 252 0.32 28 ± 9

Marine 83 0.11 > 72Sea of Japan/Manchuria 40 0.05 44 ± 9

Transition/Local 210 0.27 N/A

174

Figure 6.3 5 day (120 hour) 500 meter forward trajectories from the sampling site color coded by the back trajectories observed at the start of each forward trajectory.

175

6.4.2 Aerosol Size Distributions and Bulk Mass Concentrations

Temporal profiles of PM10 (particulate matter <10 μm) mass, BC mass, and size-

resolved particle number concentrations are shown in Figure 6.4 with the time periods of

the different air mass categories highlighted. Elevated particle number and mass

concentrations corresponded to time periods with transport from China. Average PM10

mass concentrations for each different air mass are shown in Figure 6.5a. The China time

period was characterized by the highest average PM10 mass concentration (77 μg/m3),

followed by the Sea of Japan (52 μg/m3), Korea (43 μg/m3), and Marine (24 μg/m3) air

mass periods. As expected, the Marine air masses possessed the lowest average PM10

mass concentrations due to the lower influences of anthropogenic and dust sources.

During the China air mass periods, PM10 mass concentrations were nearly twice those of

the Korea time period, despite having only spent on average of only 8 fewer hours over

the ocean. This suggests that the China source region is characterized by significantly

higher PM10 concentrations than the Korea source region. This is supported by emissions

inventories that show Eastern China has higher anthropogenic emissions than Korea or

Manchuria [Bond, et al., 2004]. A specific source region that the China back trajectories

pass directly over Shanghai, the 10th largest city in the world. Wet and dry deposition

may explain part of the difference between Eastern China and Korea, but this is likely not

a major process since Sea of Japan air masses have higher mass concentrations than the

Korea time period and having been transported for 16 hours longer over the ocean. .

Overall, time over the ocean was not a strong predictor of mass concentration.

176

Average particle size distributions for each air mass category are shown in Figure

6.5b. The China time period possessed the largest mode (0.111 μm) in the accumulation

size range with higher 0.083-1.04 μm number concentrations compared to other air mass

periods. Suggestive of less atmospheric processing, the accumulation mode was shifted to

smaller diameters for the Sea of Japan time periods. In addition, the Sea of Japan air

masses were characterized distinct supermicron (>1 μm) mode. In comparison, the Korea

time periods were characterized by an average accumulation mode shifted to larger

diameters (more aged) than the Sea of Japan time periods, but yet shifted to smaller

diameters (less aged) compared to the China time periods. The Marine air masses had the

lowest concentrations in the accumulation mode, but somewhat high supermicron

concentrations, due to sea salt. To examine the different particle sources and atmospheric

processes influencing each air mass, the chemical composition of the particles was

examined, as detailed below.

177

Figure 6.4 Time series of bulk mass measurements and size distributions. a) PM10 and BC mass concentrations, b) APS and SMPS size distributions plotted as size vs. time with color indicating concentration, c) air mass categories versus time (China – red, Korea - yellow, Marine – blue, Sea of Japan – green).

178

Figure 6.5 Average a) PM10 mass concentrations and b) size-resolved number concentrations (SMPS, 0.010-0.487 μm; APS, 0.523-2.5 μm) for each of the air mass categories (China – red, Korea – yellow, Marine – blue, and Sea of Japan – green) with the standard deviation shown as the error bar.

179

Figure 6.6 Size- and chemically-resolved mass distributions (dM/dlogDa, 0.523 – 2.5 μm) for the different air mass categories: a) China, b) Korea, c) Marine, d) Sea of Japan.

180

Figure 6.7 Number fraction vs. size (0.200-3.000 μm) for the observed ATOFMS particle types for each of the air mass categories: a) China, b) Korea, c) Marine, d) Sea of Japan. Overlaid are fractions of nitrate (62NO3

-) – blue , sulfate (HSO4-) - black, and

ammonium (18NH4+) – green, as a function of size.

181

6.4.3 Aerosol Chemistry

Chemically resolved mass and relative number size distributions are shown in Figures 6.6

and 6.7, respectively, for the different air mass categories. The characteristic mass

spectral signatures of the particle types (fresh elemental carbon (Fresh EC), elemental

carbon mixed with organic carbon (ECOC), organic carbon (OC), biomass burning, fresh

sea salt, aged sea salt, and dust) are similar to those observed during previous single

particle studies in the region [Ro, et al., 2005;Ro, et al., 2001] and are described in Figure

6.8. As an example the a large fraction of the dust observed in this study consisted of

aluminosilicates, which was also observed by previous single particle studies of Asian

outflow [Ro, et al., 2005]. While dust and sea salt were observed in the submicron size

range, they were primarily observed in the supermicron size range. Conversely, the

carbonaceous particle types were primarily observed in the submicron. Both of these are

consistent with their respective formation mechanisms [Seinfeld and Pandis, 2006]. The

large submicron (<1 μm) mass concentration mode during the China air masses is

consistent with the aged submicron number concentration mode discussed above and is

primarily composed of ECOC, biomass burning, and dust particles. The supermicron

particles were primarily dust with a small fraction of aged sea salt and other minor

particle types. Despite possessing lower average submicron mass concentrations than the

China period, the Yellow Sea/Korea air masses were characterized by similar particle

chemical composition. Separating the air masses that passed over Korea, the Yellow Sea,

northeastern China from those passing over only northeastern China and the Yellow Sea

do not significantly change the shape or chemical composition of the size distributions.

As expected, the Marine air masses are characterized by primarily fresh (non-aged) sea

182

salt particles in the supermicron mode, while the submicron mode has high mass

concentrations of ECOC and biomass burning particles, likely due to a local fire during

one marine time period. Aside from that local fire, the Marine time period observations

the bulk ion concentration trends from previous Gosan observations when winds are from

the south or southeast [Kim, et al., 2002], as observed during the Marine time periods in

this study. The highest average mass concentration of supermicron The size-resolved

number fraction of individual particles internally mixed with nitrate (62NO3-), sulfate

(97HSO4-), or ammonium (18NH4

+), secondary markers indicative of atmospheric aging,

are also shown in Figure 6.7. For all air masses, the majority of the supermicron particles

(59-74%) were mixed with nitrate, with the number fraction decreasing in the submicron

size range. In contrast, ammonium was internally mixed with few supermicron particles

with the most extreme cases being 2% by number for the Marine air masses and 12% for

the China air masses. This suggests that the majority of the observed nitrate was not in

the form of ammonium nitrate. Rather, it was likely sodium nitrate on sea salt particles

that have reacted with nitric acid or similar heterogeneous processing on dust particles.

For each of the land based air masses, nitrate was mixed with >50% of the submicron

particles; whereas during the Marine time periods, <25% were mixed with nitrate,

showing the lower degree of atmospheric aging. For all air masses, >90% of the

submicron particles were internally mixed with sulfate. This fraction decreases

precipitously in the supermicron size range for all air masses. The most extreme decrease

occurs for the Marine air masses, for which sulfate was internally mixed with 8% of all

supermicron particles; whereas, sulfate was mixed with 27% of all supermicron particles

during the China time periods. This was largely related having more carbonaceous

183

particles (i.e. ECOC and Biomass) in the supermicron during the China time period that

are highly mixed with sulfate (> 80%), but sea salt and dust during the China period are

also mixed with sulfate in higher fractions (19%) than during the Marine period (3%).

Ammonium is primarily internally mixed with ~25% submicron particles during all air

mass time periods, likely in the form of ammonium sulfate based on the ammonium

distribution by size. These results indicate the degree of aging by air mass type and give

an indication of the form of secondary species on particles as a function of size. These

results suggest that the fraction of particles mixed secondary species are distributed based

on size, but that during more polluted time periods a greater fraction of sulfate, nitrate,

and ammonium are observed outside their preferred size range.

184

Figure 6.8 Positive and negative mass spectra for the main particle types: a) Fresh EC, b) ECOC, c) OC, d) Biomass Burning, e) Fresh SS, f) Aged SS, g) Dust

185

Figure 6.9 Matrix showing the fraction of different particle types (Fresh EC, ECOC, OC, biomass burning, sea salt, and dust) containing peaks indicative of secondary aging (35Cl-, 46NO2

-, 62NO3-, 125HNO3NO3

-, 80SO3-, 97HSO4

-, 195H2SO4HSO4-, 18NH4

+, 43C2H3O+).

186

6.4.4 Aerosol Chemical Mixing States

To investigate the internal mixing of secondary species with different particle

types, the number fractions of each particle type mixed with various ion markers are

shown in a matrix in Figure 6.9. Only particles with both positive and negative mass

spectra were considered in the calculations of these fractions. The secondary ion markers

examined included: chloride (35Cl-), nitrate (46NO2-, 62NO3

-), nitric acid (125HNO3NO3-),

sulfate (80SO3- and 97SO4

-), sulfuric acid (195H2SO4HSO4-), ammonium (18NH4

+), and

oxidized organic carbon (43C2H3O+). For particle types that were observed primarily in

the submicron size range (Fresh EC, ECOC, OC, and biomass), the most frequently

observed secondary species is sulfate, while for particle types primarily in the

supermicron (sea salt and dust) the most abundant secondary species are nitrate markers.

Though the fraction of nitrate is highest in the supermicron particle type it is also

observed on roughly half of submicron particle types. Sulfate is observed on a low

fraction of supermicron particle types (dust and sea salt), though more so during polluted

time periods as discussed above. Ammonium is observed most frequently on submicron

particle types, specifically ECOC, OC, and biomass burning. Oxidized organic carbon is

observed primarily on submicron carbonaceous particle types and in very low fractions

on the dust and sea salt particle types (< 10%). These trends were consistent across all air

mass categories, as shown in Figure 6.10. A second important piece of information from

this figure involves which secondary markers are observed on the same particle types.

For the primary submicron particle types (ECOC and biomass burning) ~15% of particles

are mixed with sulfuric acid, indicating that the sulfate on that fraction of particle likely

from sulfuric acid. Similarly, ~10% of submicron particle types are mixed with nitric

187

acid. A large fraction of the remainder of the sulfate and nitrate on submicron particles is

likely in the form of ammonium sulfate and ammonium nitrate based on the ~50% of

ECOC and biomass burning particles are mixed with ammonium. In the supermicron the

low fraction of ammonium and nitric acid indicates that most of the nitrate is in the form

of sodium nitrate or nitrate formed from other heterogeneous processes.

188

Figure 6.10 Matrices showing the fraction of particles from each particle type mixed with selected secondary species during 4 time periods.

189

Figure 6.11 Ternary plots showing the relative intensities of ammonium, nitrate, and sulfate ion signals for individual particles for the a) ECOC, b) biomass burning, c) sea salt, and d) dust particle types. The color of each marker indicates its air mass time period (China – red, Korea – yellow, Sea of Japan – green, Marine – blue). ~1000 representative particles are shown on each plot for each time period.

190

As ion signals are related to the mass of a species with the individual particles [Gross, et

al., 2000;Pratt, et al., 2009], the relative amounts of sulfate, nitrate, and ammonium on

the different particle types may be examined. For the ternary plots shown in Figure 6.11

each point represents one particle and the closer to a corner the particle is, the closer that

peak is to 100% of the sum of the three peaks plotted. The two submicron particle types

(Figure 6.11a – ECOC and Figure 6.11b – Biomass) are centered on the sulfate corner of

the plot extending slightly more towards the ammonium corner than the nitrate corner.

For both of these particle types the marine time period has the strongest sulfate signal,

with the China the most mixed with ammonium and nitrate. The Korea and Sea of Japan

time periods are similar to the China time period, but with slightly less nitrate and

ammonium. The two supermicron particle types (Figure 6.11c – Sea Salt and Figure

6.11d – dust) are centered around the nitrate corner. Dust during the Sea of Japan and

Marine time periods is mixed mostly between ammonium and nitrate, while a greater

fraction is mixed with sulfate during the China (and to a lesser extent Korea) time

periods. This is consistent with Figure 6.7 showing higher fractions of supermicron

particles mixed with sulfate during the most polluted (China) time period. Sea salt is

mixed most strongly with nitrate, especially since there is an interfering sodium chloride

peak at m/z -97 that shifts some marine points towards the sulfate corner during the

marine period. This analysis by particle type shows that the sulfate and nitrate trends are

consistent across the sub and supermicron size ranges, with China the slightly more aged

than the Korea and Sea of Japan air masses and the marine air masses the least aged.

The relative contributions of nitrate, sulfate, and ammonium by both particle type

and size are shown in Figure 6.12. A clear trend was observed, wherein the larger

191

particles were predominantly internally mixed with nitrate with the smaller particles

internally mixed with sulfate, similar to that observed in Figure 6.7. The larger particles

also have a subset mixed with ammonium, but these particles have very low levels of

sulfate, suggesting that ammonium sulfate is not present on these particles. In contrast, as

the smaller particles age, ammonium nitrate condenses on the individual particles. These

trends are consistent across each air mass category (not shown). The prevalence of nitrate

on supermicron particles and sulfate on submicron particles processes reinforces the

importance of particle size on secondary processing across all particle types as has been

shown for dust previously [Sullivan, et al., 2007].

192

Figure 6.12 Ternary plot showing the relative intensities of ammonium, nitrate, and sulfate for individual particles for ECOC (circle), biomass (triangle), sea salt (cross), and dust (x) particles during all time periods. The color of the marker indicates the size of the particle from red (smallest) to purple (largest). ~16,000 representative particles are shown.

193

6.5 Conclusions

The results presented herein demonstrate amount and form of secondary aging

that occurs to Asian outflow particles during different air masses sampled at the Gosan

supersite. For the two air masses that are most likely to be involved in long range

transport (China and Marine), distinct differences between the particles in these time

periods were observed. The air masses originating over China were characterized by the

highest observed particle mass and number concentrations, particularly for the submicron

size range. Despite the shortest transport time to the site (and hence shortest amount of

time for atmospheric processing), the China air masses were the most aged with respect

to the number fraction mixed with nitrate, sulfate, and ammonium, as well as the amount

of ammonium nitrate on the submicron particles. In fact, all of the land-based air masses

(China, Korea, and Sea of Japan) were substantially mixed with secondary species,

reflecting the exposure of most aerosols in the region to high concentrations of gaseous

pollutants. Also, the high fractions and intensities of nitrate and sulfate in comparison to

ammonium show that there is likely not enough base to neutralize acidic gases including

nitric and sulfuric acid involved in the aging process, suggesting that the Asian outflow

sampled is likely quite acidic. This is all in contrast to the Marine period which, while

somewhat aged, was far less so than the land based time periods.

The differences in the chemistry and mixing state of sub and supermicron particles

indicate the limitations of using a single time based aging parameter to represent aging in

global models. This is related both the sources of the particles, i.e. aging reactions on

biomass burning particles are quite different than on sea salt particles. Also the

194

prominence of sulfate on submicron particles and nitrate on supermicron particles

indicates different aging processes are taking place that proceed on different time scales.

These differences impact how well particles take up water and subsequently influence

whether they can act as cloud nuclei and how they scatter solar radiation. To improve

models parameterizations both the differences in sub and supermicron particle mixing

state need to be incorporated in a more complex manner, while dealing with

computational constraints. Efforts to explore parameterizations between single particle

measurements and single particle modeling efforts will be an important first step to

addressing these current model limitations.


Andrew Ault is grateful for a Department of Energy Global Change Education

Program graduate research environmental fellowship. This work was supported by the

Atmospheric Brown Cloud project funded by the United Nations Environmental

Programme and the National Oceanic and Atmospheric Administation (NOAA). We

thank Prof. Yoon at Seoul National University and his group for their hospitality and

assistance during the campaign. Prof. Yoon and Dr. Sang Woo-Kim provided the SMPS

data. We appreciate Craig Corrigan for technical support prior to and during the

campaign. The Korean Meteorological Administration is acknowledged for

meteorological and PM10 data at the Gosan site. The authors gratefully acknowledge the

NOAA Air Resources Laboratory for the provision of the HYSPLIT transport model and

READY website (http://www.arl.noaa.gov/ready.html) used in this publication.

195

Chapter 6 is in preparation for submission to Atmospheric Environment: A. P.

Ault, A. Zhu, V. Ramanathan, and K.A. Prather, Characterization of Asian Outflow at

Gosan, Korea by Single Particle Mass Spectrometry.

196

6.7 References

A. P. Ault, M. J. Moore, H. Furutani and K. A. Prather, Impact of emissions from the Los Angeles port region on San Diego air quality, Environmental Science & Technology, 43 (10), 3500-3506, 2009.

T. S. Bates, P. K. Quinn, D. J. Coffman, D. S. Covert, T. L. Miller, J. E. Johnson, G. R.

Carmichael, I. Uno, S. A. Guazzotti, D. A. Sodeman, K. A. Prather, M. Rivera, L. M. Russell and J. T. Merrill, Marine boundary layer dust and pollutant transport associated with the passage of a frontal system over eastern Asia, Journal Of Geophysical Research-Atmospheres, 109 (D19), 2004.

T. C. Bond, D. G. Streets, K. F. Yarber, S. M. Nelson, J. H. Woo and Z. Klimont, A

technology-based global inventory of black and organic carbon emissions from combustion, Journal Of Geophysical Research-Atmospheres, 109 (D14), 2004.

G. R. Carmichael, B. Adhikary, S. Kulkarni, A. D'Allura, Y. H. Tang, D. Streets, Q.

Zhang, T. C. Bond, V. Ramanathan, A. Jamroensan and P. Marrapu, Asian Aerosols: Current and Year 2030 Distributions and Implications to Human Health and Regional Climate Change, Environmental Science & Technology, 43 (15), 5811-5817, 2009.

G. R. Carmichael, M. S. Hong, H. Ueda, L. L. Chen, K. Murano, J. K. Park, H. G. Lee,

Y. Kim, C. Kang and S. Shim, Aerosol composition at Cheju Island, Korea, Journal Of Geophysical Research-Atmospheres, 102 (D5), 6047-6061, 1997.

G. R. Carmichael, Y. Zhang, L. L. Chen, M. S. Hong and H. Ueda, Seasonal variation of

aerosol composition at Cheju Island, Korea, Atmospheric Environment, 30 (13), 2407-2416, 1996.

L. L. Chen, G. R. Carmichael, M. S. Hong, H. Ueda, S. Shim, C. H. Song, Y. P. Kim, R.

Arimoto, J. Prospero, D. Savoie, K. Murano, J. K. Park, H. G. Lee and C. Kang, Influence of continental outflow events on the aerosol composition at Cheju Island, South Korea, Journal Of Geophysical Research-Atmospheres, 102 (D23), 28551-28574, 1997.

B. Croft, U. Lohmann and K. von Salzen, Black carbon ageing in the Canadian Centre for

Climate modelling and analysis atmospheric general circulation model, Atmospheric Chemistry And Physics, 5, 1931-1949, 2005.

R. A. Duce, C. K. Unni, B. J. Ray, J. M. Prospero and J. T. Merrill, Long-Range

Atmospheric Transport Of Soil Dust From Asia To The Tropical North Pacific - Temporal Variability, Science, 209 (4464), 1522-1524, 1980.

197

E. J. Dunlea, P. F. DeCarlo, A. C. Aiken, J. R. Kimmel, R. E. Peltier, R. J. Weber, J. Tomlinson, D. R. Collins, Y. Shinozuka, C. S. McNaughton, S. G. Howell, A. D. Clarke, L. K. Emmons, E. C. Apel, G. G. Pfister, A. van Donkelaar, R. V. Martin, D. B. Millet, C. L. Heald and J. L. Jimenez, Evolution of Asian aerosols during transpacific transport in INTEX-B, Atmospheric Chemistry Physics, 9 (19), 7257-7287, 2009.

S. Eckhardt, A. Stohl, H. Wernli, P. James, C. Forster and N. Spichtinger, A 15-year

climatology of warm conveyor belts, J. Clim., 17 (1), 218-237, 2004. K. Eguchi, I. Uno, K. Yumimoto, T. Takemura, A. Shimizu, N. Sugimoto and Z. Liu,

Trans-pacific dust transport: integrated analysis of NASA/CALIPSO and a global aerosol transport model, Atmospheric Chemistry And Physics, 9 (9), 3137-3145, 2009.

Y. Feng and V. Ramanathan, Investigation of aerosol-cloud interactions using a chemical

transport model constrained by satellite observations, Tellus Series B-Chemical And Physical Meteorology, 62 (1), 69-86, 2010.



S. J. Ghan and S. E. Schwartz, Aerosol properties and processes - A path from field and

laboratory measurements to global climate models, Bulletin Of The American Meteorological Society, 88 (7), 1059-+, 2007.

D. S. Gross, M. E. Galli, P. J. Silva and K. A. Prather, Relative sensitivity factors for

alkali metal and ammonium cations in single particle aerosol time-of-flight mass spectra, Analytical Chemistry, 72 (2), 416-422, 2000.

Y. Hara, K. Yumimoto, I. Uno, A. Shimizu, N. Sugimoto, Z. Liu and D. M. Winker,

Asian dust outflow in the PBL and free atmosphere retrieved by NASA CALIPSO and an assimilated dust transport model, Atmospheric Chemistry And Physics, 9 (4), 1227-1239, 2009.

P. V. Hobbs, G. C. Bluhm and T. Ohtake, Transport Of Ice Nuclei Over North Pacific

Ocean, Transactions - American Geophysical Union, 52 (5), 429-&, 1971. M. Holzer, T. M. Hall and R. B. Stull, Seasonality and weather-driven variability of

transpacific transport, J. Geophys. Res.-Atmos., 110 (D23), 2005. J. Kim, C. H. Jung, B. C. Choi, S. N. Oh, F. J. Brechtel, S. C. Yoon and S. W. Kim,

Number size distribution of atmospheric aerosols during ACE-Asia dust and precipitation events, Atmospheric Environment, 41 (23), 4841-4855, 2007.

198

J. Kim, S. C. Yoon, A. Jefferson and S. W. Kim, Aerosol hygroscopic properties during

Asian dust, pollution, and biomass burning episodes at Gosan, Korea in April 2001, Atmospheric Environment, 40 (8), 1550-1560, 2006a.

J. Kim, S. C. Yoon, S. W. Kim, F. Brechtel, A. Jefferson, E. G. Dutton, K. N. Bower, S.

Cliff and J. J. Schauer, Chemical apportionment of shortwave direct aerosol radiative forcing at the Gosan super-site, Korea during ACE-Asia, Atmos. Environ., 40 (35), 6718-6729, 2006b.

K. H. Kim, M. H. Lee, G. W. Lee, Y. P. Kim, Y. H. Youn and J. M. Oh, Observations of

aerosol-bound ionic compositions at Cheju Island, Korea, Chemosphere, 48 (3), 317-327, 2002.

S. W. Kim, S. C. Yoon, A. Jefferson, J. A. Ogren, E. G. Dutton, J. G. Won, Y. S. Ghim,

I. B. Lee and J. S. Han, Aerosol optical, chemical and physical properties at Gosan, Korea during Asian dust and pollution episodes in 2001, Atmospheric Environment, 39 (1), 39-50, 2005.

D. Koch, Transport and direct radiative forcing of carbonaceous and sulfate aerosols in

the GISS GCM, Journal Of Geophysical Research-Atmospheres, 106 (D17), 20311-20332, 2001.

M. Kuwata, Y. Kondo, Y. Miyazaki, Y. Komazaki, J. H. Kim, S. S. Yum, H. Tanimoto

and H. Matsueda, Cloud condensation nuclei activity at Jeju Island, Korea in spring 2005, Atmospheric Chemistry And Physics, 8 (11), 2933-2948, 2008.




properties of soot with implications for radiative forcing estimates, 106 (29), 11872-11877, 2009.



T. Nakajima, S. C. Yoon, V. Ramanathan, G. Y. Shi, T. Takemura, A. Higurashi, T.

Takamura, K. Aoki, B. J. Sohn, S. W. Kim, H. Tsuruta, N. Sugimoto, A. Shimizu, H. Tanimoto, Y. Sawa, N. H. Lin, C. T. Lee, D. Goto and N. Schutgens, Overview of the Atmospheric Brown Cloud East Asian Regional Experiment

199

2005 and a study of the aerosol direct radiative forcing in east Asia, Journal Of Geophysical Research-Atmospheres, 112 (D24), 2007.

M. Nishikawa, S. Kanamori, N. Kanamori and T. Mizoguchi, Kosa Aerosol As Eolian

Carrier Of Anthropogenic Material, Science Of The Total Environment, 107, 13-27, 1991.

K. A. Pratt, L. E. Hatch and K. A. Prather, Seasonal volatility dependence of ambient

particle phase amines, Environmental Science & Technology, 43 (14), 5276-5281, 2009.

X. Y. Qin, Characterization of Ambient Aerosol Composition and Formation

Mechanisms and Development of Quantification Methodologies Utilizing ATOFMS, Doctoral Thesis thesis, University of California San Diego, La Jolla, 2007.



M. V. Ramana and V. Ramanathan, Abrupt transition from natural to anthropogenic

aerosol radiative forcing: Observations at the ABC-Maldives Climate Observatory, Journal Of Geophysical Research-Atmospheres, 111 (D20), 2006.

V. Ramanathan and G. Carmichael, Global and regional climate changes due to black

carbon, Nature Geoscience, 1 (4), 221-227, 2008. V. Ramanathan, P. J. Crutzen, J. Lelieveld, A. P. Mitra, D. Althausen, J. Anderson, M. O.

Andreae, W. Cantrell, G. R. Cass, C. E. Chung, A. D. Clarke, J. A. Coakley, W. D. Collins, W. C. Conant, F. Dulac, J. Heintzenberg, A. J. Heymsfield, B. Holben, S. Howell, J. Hudson, A. Jayaraman, J. T. Kiehl, T. N. Krishnamurti, D. Lubin, G. McFarquhar, T. Novakov, J. A. Ogren, I. A. Podgorny, K. Prather, K. Priestley, J. M. Prospero, P. K. Quinn, K. Rajeev, P. Rasch, S. Rupert, R. Sadourny, S. K. Satheesh, G. E. Shaw, P. Sheridan and F. P. J. Valero, Indian Ocean Experiment: An integrated analysis of the climate forcing and effects of the great Indo-Asian haze, Journal Of Geophysical Research-Atmospheres, 106 (D22), 28371-28398, 2001.

N. Riemer, H. Vogel, B. Vogel and F. Fiedler, Modeling aerosols on the mesoscale-

gamma: Treatment of soot aerosol and its radiative effects, Journal Of Geophysical Research-Atmospheres, 108 (D19), 2003.

N. Riemer, M. West, R. Zaveri and R. Easter, Estimating black carbon aging time-scales

with a particle-resolved aerosol model, Journal Of Aerosol Science, 41 (1), 143-158, 2010.

200

N. Riemer, M. West, R. A. Zaveri and R. C. Easter, Simulating the evolution of soot

mixing state with a particle resolved aerosol model, Journal Of Geophysical Research-Atmospheres, in press, 2009.

C. U. Ro, H. Hwang, Y. Chun and R. Van Grieken, Single-particle characterization of

four "Asian Dust" samples collected in Korea, using low-Z particle electron probe X-ray microanalysis, Environmental Science & Technology, 39 (6), 1409-1419, 2005.

C. U. Ro, K. Y. Oh, H. Kim, Y. P. Kim, C. B. Lee, K. H. Kim, C. H. Kang, J. Osan, J. De

Hoog, A. Worobiec and R. Van Grieken, Single-particle analysis of aerosols at Cheju Island, Korea, using low-Z electron probe X-ray microanalysis: A direct proof of nitrate formation from sea salts, Environmental Science & Technology, 35 (22), 4487-4494, 2001.

R. S. Russo, R. W. Talbot, J. E. Dibb, E. Scheuer, G. Seid, C. E. Jordan, H. E. Fuelberg,

G. W. Sachse, M. A. Avery, S. A. Vay, D. R. Blake, N. J. Blake, E. Atlas, A. Fried, S. T. Sandholm, D. Tan, H. B. Singh, J. Snow and B. G. Heikes, Chemical composition of Asian continental outflow over the western Pacific: Results from Transport and Chemical Evolution over the Pacific (TRACE-P), Journal Of Geophysical Research-Atmospheres, 108 (D20), 2003.

L. K. Sahu, Y. Kondo, Y. Miyazaki, M. Kuwata, M. Koike, N. Takegawa, H. Tanimoto,

H. Matsueda, S. C. Yoon and Y. J. Kim, Anthropogenic aerosols observed in Asian continental outflow at Jeju Island, Korea, in spring 2005, Journal Of Geophysical Research-Atmospheres, 114, 2009.

K. Sassen, Indirect climate forcing over the western US from Asian dust storms,

Geophys. Res. Lett., 29 (10), 2002. J. H. Seinfeld and S. N. Pandis, Atmospheric Chemistry and Physics, Wiley Interscience,

Hoboken, NJ, 2006. S. Solomon, D. Qin, M. Manning, Z. Chen, M. Marquis, K. B. Averyt, M. Tignor and H.




201

J. L. Stith, V. Ramanathan, W. A. Cooper, G. C. Roberts, P. J. DeMott, G. Carmichael, C. D. Hatch, B. Adhikary, C. H. Twohy, D. C. Rogers, D. Baumgardner, A. J. Prenni, T. Campos, R. Gao, J. Anderson and Y. Feng, An overview of aircraft observations from the Pacific Dust Experiment campaign, Journal Of Geophysical Research-Atmospheres, 114, 2009.



D. Topping, H. Coe, G. McFiggans, R. Burgess, J. Allan, M. R. Alfarra, K. Bower, T. W.

Choularton, S. Decesari and M. C. Facchini, Aerosol chemical characteristics from sampling conducted on the Island of Jeju, Korea during ACE Asia, Atmospheric Environment, 38 (14), 2111-2123, 2004.



R. A. VanCuren and T. A. Cahill, Asian aerosols in North America: Frequency and

concentration of fine dust, J. Geophys. Res.-Atmos., 107 (D24), 2002. H. L. A. Wong, T. Wang, A. Ding, D. R. Blake and J. C. Nam, Impact of Asian

continental outflow on the concentrations of O-3, CO, NMHCs and halocarbons on Jeju Island, South Korea during March 2005, Atmospheric Environment, 41 (14), 2933-2944, 2007.

S. C. Yoon and J. Kim, Influences of relative humidity on aerosol optical properties and

aerosol radiative forcing during ACE-Asia, Atmospheric Environment, 40 (23), 4328-4338, 2006.

H. B. Yu, L. A. Remer, M. Chin, H. S. Bian, R. G. Kleidman and T. Diehl, A satellite-

based assessment of transpacific transport of pollution aerosol, Journal Of Geophysical Research-Atmospheres, 113 (D14), 2008.

S. S. Yum, G. Roberts, J. H. Kim, K. Y. Song and D. Y. Kim, Submicron aerosol size

distributions and cloud condensation nuclei concentrations measured at Gosan, Korea, during the Atmospheric brown clouds - East Asian Regional Experiment 2005, J. Geophys. Res.-Atmos., 112 (D22), 2007.

7 An Intercontinental Overview of Size-Resolved Chemical

Mixing State by Single Particle Mass Spectrometry

7.1 Synopsis

The radiative forcings associated direct and indirect effects of atmospheric

aerosols represent the one of the greatest sources of uncertainty in predicting future

climate change. Of particular important is the representation of aerosol mixing state with

global climate models, as estimates of radiative forcings vary dramatically with different

mixing state assumptions. We present an intercontinental overview of aerosol time-of-

flight mass spectrometry measurements, providing single-particle size, chemistry, and

mixing state. We focus on the six main observed classes of particles (fresh elemental

carbon, aged elemental carbon mixed with organic carbon, organic carbon, biomass

burning, sea salt, and dust) that represent 94% of particles in the accumulation and coarse

modes (0.1 – 3.0 μm). Differences in chemical mixing state for these particles across

aged urban, fresh urban, marine, and remote continental locations are described with

emphasis on the mixing of secondary species (nitrate, sulfate, ammonium, and secondary

organic carbon). In particular, sulfate was primarily found to be internally mixed with

carbonaceous particle types, even in clean marine environments, and nitrate was often

mixed with sea salt and dust. The observed particle mixing states are not in agreement

with recent depictions by the Intergovernmental Panel on Climate Change reports and

have important ramifications on the estimates of aerosol radiative forcings.

202

203

7.2 Introduction

Atmospheric aerosols have the ability to alter climate on a global scale by

scattering and absorbing solar radiation (direct effect) and altering the properties of

clouds by serving as cloud condensation and ice nuclei (indirect effect) [Forster, et al.,

2007]. On average, aerosols are predicted to have a net cooling effect globally [Forster,

et al., 2007] with direct aerosol radiative forcing estimated at -0.3 ± 0.2 W/m2 [Myhre,

2009]. Despite these significant impacts, the overall scientific understanding of the

impact of aerosol effects on climate is still low [Forster, et al., 2007]. This is related in

part to the complexity of aerosols and the variability in their properties. Primary aerosol

particles originate from a variety of both natural and anthropogenic sources, including

biomass burning, incomplete fossil fuel combustion, and the wind-driven suspension of

soil and sea spray [Poschl, 2005]. Depending on particle size, chemical composition, and

local meteorology, lifetimes of atmospheric particles range from seconds to

approximately 1 month, allowing transportation on a hemispheric scale [Raes, et al.,

2000;Williams, et al., 2002]. During transport, atmospheric particles undergo physical

and chemical transformations (atmospheric aging), due to coagulation, heterogeneous

reactions of particles with trace gases, gas-particle partitioning of semi-volatile species,

and aqueous-phase processing, leading to changes in particle size, structure, and chemical

composition [Poschl, 2005]. Due to the evolving physical and chemical properties of the

distribution of aerosols, quantifying aerosol impacts on radiative forcing and clouds

represents a challenging task [Forster, et al., 2007].

Particle mixing state is defined as the distribution of different chemical species

within single particles, resulting in differing chemistry between individual particles.

204

Generally, this is considered in the context of secondary species that have become mixed

with a primary emission particle (e.g. dust internally mixed with nitrate or soot internally

mixed with sulfate). The mixing state of a particle determines its morphology, reactivity,

hygroscopicity, and optical and toxicological properties [Fuzzi, et al., 2006;Prather, et

al., 2008;Schlesinger, et al., 2006]. As an important example, as externally mixed soot

particles age, they can convert to internally mixed particles with a soot core surrounded

by sulfate, ammonium, organics, nitrate, and water [Moffet and Prather, 2009]. The

radiative forcing of soot particles is significantly altered by mixing with sulfuric acid and

sulfate species, resulting in increased absorption of solar radiation [Jacobson,

2001;Moffet and Prather, 2009], as well as increased particle hygroscopicity [Zuberi, et

al., 2005]. However, global climate models (GCMs), frequently represent black carbon

and sulfate as externally mixed, with black carbon as absorbing particles and sulfate as

purely scattering particles [Forster, et al., 2007]. This leads to lower estimates of

radiative forcing from black carbon (0.2-0.4 W/m2) [Highwood and Kinnersley,

2006;Koch, et al., 2007] than predictions from models that account for particle

processing in the atmosphere (0.6-1.2 W/m2) [Chung and Seinfeld, 2002;Jacobson,

2001;Ramanathan and Carmichael, 2008]. The representation of black carbon particles,

in particular, is important as they are estimated to have the second largest global warming

potential of any atmospheric constituent besides CO2 [Jacobson, 2001;Moffet and

Prather, 2009].

The mixing of particles with secondary species (e.g. oxidized organic, sulfate,

nitrate, ammonium) impacts the ability of individual particles to act as cloud

condensation nuclei (CCN) and ice nuclei (IN) [Cantrell and Heymsfield,

205

2005;McFiggans, et al., 2006]. For example, when reacted with nitric acid, calcium dust

becomes more hygroscopic and a more effective CCN; in contrast, when reacted with

oxalic acid, calcium dust becomes less soluble and a less effective CCN [Sullivan, et al.,

2009a]. CCN closure studies have shown that size-resolved mixing state is often needed

for accurate predictions of CCN concentrations [Cubison, et al., 2008]. Further, in

laboratory studies, coatings of sulfuric acid and ammonium sulfate on mineral dust were

observed to decrease the potential for heterogeneous ice nucleation [Cziczo, et al.,

2009a]. Changes in aerosols due to sulfate coatings impacting mixed-phase clouds are

predicted to significantly impact indirect aerosol radiative forcing estimates [Lohmann

and Hoose, 2009]. Thus, understanding the size-resolved chemical mixing state of

particles, including both the primary sources of particles in the atmosphere and the

impacts of atmospheric aging, is critical to accurately representing particles in global

models that predict future climate impacts.

One of the greatest challenges for modeling atmospheric aerosols is representing

particle mixing state accurately, particularly due to changes in mixing state over the

particle lifetime [Ghan and Schwartz, 2007]. As such GCMs frequently parameterize

aerosols in a simplistic fashion that is not representative of their state in the atmosphere.

In the 2007 Intergovernmental Panel on Climate Change (IPCC) report [Solomon, et al.,

2007], the major aerosol types were considered to be sulfate, fossil fuel organic carbon,

fossil fuel black carbon, biomass burning, nitrate, and mineral dust. Most of the aerosol

classes are defined as external mixtures, with the exception being the biomass burning

aerosols, defined as an internal mixture of organic carbon, black carbon, and inorganics,

including nitrate and sulfate [Solomon, et al., 2007]. There are currently many efforts to

206

improve the incorporation and treatment of aerosol properties into climate models [Ghan

and Schwartz, 2007]. However, many assumptions and parameterizations in large scale

climate models are still lacking, such as the frequent use of a single aging timescale to

convert fresh hydrophobic soot particles to aged hydrophilic particles internally mixed

with secondary species, for example [Croft, et al., 2005;Koch, 2001]. This can lead to

incorrect assessments of the black carbon burdens [Riemer, et al., 2003]. Efforts are

being made to improve particle mixing state in aerosol models by incorporating more

detailed mixing state[Bauer, et al., 2008] and treating particles individual [Riemer, et al.,

2010;Riemer, et al., 2009]. Ambient measurements of size-resolved chemical mixing

state are necessary to provide empirical evidence to properly constrain, improve, and

evaluate the representation of particles in these models [Ghan and Schwartz, 2007].

Single-particle mass spectrometry provides the ability to measure the size and

chemistry of individual particles in real-time, as reviewed recently [Pratt and Prather,

2010b]. Laser desorption/ionization allows for analysis of both the non-refractory (e.g.

organics, ammonium nitrate) and refractory (e.g. mineral dust, soot) components of

atmospheric aerosols. The aerosol time-of-flight mass spectrometer (ATOFMS) [Prather,

et al., 1994] utilizes a dual-polarity mass spectrometer [Gard, et al., 1997], which is used

to simultaneously detect both the positive and negative ions resulting from a single

particle. This provides the ability to identify the primary particle types/source, such as

biomass burning, organic carbon, or elemental carbon, and examine its mixing with

secondary species, such as ammonium, nitrate, and/or sulfate. In the years since the

development of the ATOFMS, it has been deployed in numerous field campaigns across

the globe [Noble and Prather, 2000;Pratt and Prather, 2010b;Sullivan and Prather,

207

2005]. Recent aircraft-based ATOFMS measurements showed significant variations in

single particle mixing state with altitude [Pratt and Prather, 2010a].

Herein, we provide an overview of ground-based ATOFMS measurements of

single particle mixing state from a nearly global perspective. The size-resolved chemical

composition of particles determined from studies in North America, Europe, Asia, and

Africa are described and grouped according to similarities in location and particle

chemistry. The mixing of general particle classes (fresh elemental carbon, organic

carbon, elemental carbon mixed with organic carbon, biomass burning, sea salt, dust, and

other) with specific secondary species (oxidized organic carbon, ammonium, nitrate, and

sulfate) from each site is then discussed. This overview incorporates ATOFMS data from

22 measurement campaigns, wherein large numbers of individual 0.1-3.0 μm particles

(~25,000 – 3,400,000 per study) were chemically analyzed in real-time with high size

resolution. While additional single-particle chemical information (e.g. mixing with

dicarboxylic acids [Sullivan and Prather, 2007], oligomers [Denkenberger, et al., 2007],

or metals [Ault, et al., 2010b;Moffet, et al., 2008b] is obtained, we present a simplified

view of the major components and secondary species contributing to particle mixing state

in an effort to provide a reference for modelers incorporating single particle processes

into small scale aerosol modules and GCMs.

7.3 Experimental

7.3.1 Aerosol Time-of-Flight Mass Spectrometry (ATOFMS)

The design and details of the ATOFMS have been described in detail elsewhere

[Gard, et al., 1997]. Briefly, particles are introduced into the ATOFMS through a

converging nozzle or aerodynamic lens system into a differentially pumped vacuum

208

chamber where they are collimated into a beam and accelerated to terminal velocities.

The particles then pass through two continuous wave lasers (diode pumped Nd:YAG

lasers operating at 532 nm) located 6 cm apart. The velocity of each particle is used to

determine the vacuum aerodynamic diameter (dva) from calibration with polystyrene

latex spheres of known diameter, shape, and density. The sized particles are then

desorbed and ionized using a Nd:YAG laser (266nm, ~1.0-1.5 mJ/pulse) that is fired

when the particle is calculated to enter the center of the mass spectrometer ion source

region. Positive and negative ions from each individual particle are detected using a dual-

polarity, reflectron time-of-flight mass spectrometer. As detailed in Table 7.1, ATOFMS

instruments during 18 of the 22 the studies possessed a converging nozzle inlet with a

size range from ~0.1 – 3.0 μm [Gard, et al., 1997]. The other 4 studies used an

aerodynamic lens inlet with a size range of 0.1 – 1.7 μm [Dall'Osto, et al., 2008].

Ultrafine ATOFMS [Su, et al., 2004] and aircraft ATOFMS [Pratt, et al., 2009c]

easur

n;

although for small communities in large urban areas the urban area is referenced. Also

m ements are not discussed.

Information on the 22 studies examined herein is listed in Table 1 including:

study location, study abbreviation, study name, dates, ATOFMS inlet type, number of

particles chemically analyzed, and previous publications. The 22 study locations are

illustrated on a nearly global map (Figure 7.1a), as well as zoomed in on the United

States of America (Figure 7.1b), colored according to their empirically determined

categories. All studies included multiple weeks of surface-based ATOFMS

measurements, resulting in ~25,000 – 3,400,000 individual particles chemically analyzed

per study. Within this manuscript, each study is referred to by its geographic locatio

209

Table 7.1 List of the ATOFMS studies included in this comparison with information on the study location, dates, instrument type (nozzle or lens), number of chemically-analyzed particles, and literature associated with the different studies.

Study Location Abbreviation Study Name Lat Lon Dates Inlet Particles Papers PublishedGrand Canyon, AZ, USA Grand Grand Canyon Aerosol 36.05 -112.10 Jun.-Jul. '98 Nozzle 25777

Canyon and Visibility StudyAldine (Houston), TX, USA TexAQS Texas Air Quality Study 29.90 -95.33 Aug.-Sept. '00 Nozzle 371537

New York, NY, USA PMTACS PM 2.5 Technology Assessment 40.74 -73.82 Jul.-Aug. '01 Nozzle 98266and Characterization Study

Fresno, CA, USA CRPAQS California Regional Particulate 36.75 -119.75 Nov. '00-Feb '01 Nozzle 673558 [Qin and Prather, 2006]Air Quality Study [Qin et al., 2006]

Mace Head, Ireland NAMBLEX North American Marine 53.32 -9.9 Jul.-Sept. '02 Nozzle 143524 [Dall'Osto et al., 2004]Boundary Layer Experiment

Yellowstone, WY, USA Yellowstone Mercury Roadshow-Yellowstone 44.74 -110.49 Sept. '03 Nozzle 60421 [Hall et al., 2006]Mt. Horeb, WI, USA Mt. Horeb Mercury Roadshow-Mt. Horeb 43.00 -89.65 Aug. '04 Nozzle 92453

E. St. Louis (St. Louis), IL, USA St. Louis Mercury Roadshow-East St. Louis 38.61 -90.16 Dec. '03-Feb. '04 Nozzle 1722813 [Snyder et al., 2009]Trinidad Head, CA, USA CIFEX Cloud Indirect Forcing Experiment 41.05 -124.15 Apr. '04 Nozzle 315324 [Holecek et al., 2007]Hanimaadhoo, Maldives APMEX Atmospheric Brown Cloud 6.78 73.18 Oct.-Nov. '04 Nozzle 326245 [Spencer et al., 2008]

Post Monsoonal ExperimentRiverside, CA, USA SOAR Study of Organic Aerosols 33.97 -117.32 Jul.-Aug. '05 Nozzle 1080675 [Moffet et al., 2008]

in Riverside [Gaston et al., 2010]Cape Verde Islands, Africa DODO Dust Outflow and Deposition Cruise Cruise Jan.-Feb. '06 Lens 167771

to the OceanMexico City, Mexico MILAGRO Megacity Initiative: Local And 19.49 -99.15 Mar. '06 Nozzle 901552 [Moffet et al., 2008a,b,c]

Global Research Observations [Moffet et al., 2009]North Atlantic MAP Marine Aerosol Production Cruise Cruise Jun.-Jul. '06 Lens 65336

La Jolla (San Diego), CA, USA SIO Pier Scripps Oeaonography 32.87 -117.26 Aug.-Sept. '06 Nozzle 3346035 [Ault et al., 2009]Pier Study - 2006

London, UK REPARTEE Regents Park and Tower 51.5 -0.12 Oct. '04 Lens 127832 [Dall'Osto et al., 2009]Environmental Experiment

Gosan, Jeju, Korea PACDEX Pacific Dust Experiment 33.29 126.18 Apr.-May. '07 Nozzle 2967725 Dearborn (Detroit), MI, USA Detroit Detroit, MI 42.25 83.20 Aug. '07 Nozzle 345665

Los Angeles, CA, USA LABMS Los Angeles Mobile Study 33.74 -118.27 Nov. '07 Nozzle 1014946 [Ault et al., 2010]Bologna, Italy SPC San Pietro Capofiume 44.53 11.3 Jul. '08 Lens 47559

Atlanta, GA, USA AMIGAS August Mini-Intensive Gas 33.78 -84.42 Aug.-Sept. '08 Nozzle 245791 and Aerosol Sampling

Sugar Pine Dam, CA, USA Calwater Calwater Early Start 39.13 -120.80 Feb.-Mar. '09 Nozzle 76806

210

Figure 7.1 a) Global map showing the ATOFMS studies analyzed for this comparison, b) United States map of ATOFMS studies used in this comparison. The site classification of each study is shown as the color of the marker with the field study name also noted.

211

the two ship-based data sets are referred to by their sampling region (Cape Verde for the

Dust DODO and North Atlantic for MAP).

7.3.2 Particle Classification

Data analysis of the single particle mass spectra was completed on each study

data set. 18 of the studies were analyzed using YAADA, a data analysis toolkit for

MATLAB (The MathWorks, Inc.). Dual-polarity mass spectra were clustered using an

Adaptive Resonance Theory based neural network algorithm (ART-2a) at a vigilance

factor of 0.8 [Song, et al., 1999]. To classify a higher fraction of particles, submicron (<1

μm) and supermicron (1-3 μm) diameter particles were analyzed separately for 7 of the

studies. For the 4 studies not analyzed using YAADA (Yellowstone, St. Louis, Mt.

Horeb, and Detroit), the graphical user interface based program ENCHILADA was used,

and clustering was carried out using the K-means clustering algorithm [Gross, et al.,

2010]. Both ART-2a and the K-means algorithm combine particles into clusters based on

the intensity of ion peaks in individual mass spectra and have similar accuracy in

clustering [Rebotier and Prather, 2007]. On average > 85% of all particles from each

field study were classified and used for this analysis. After the initial clustering analysis,

clusters were combined into six general particle classes (fresh elemental carbon (Fresh

EC), aged elemental carbon mixed with organic carbon (ECOC), organic carbon (OC),

biomass burning (Biomass), sea salt (Sea Salt), and dust (Dust)) based on previous field

and source apportionment studies and using the criteria discussed below. Particles that

did not fit into the defined classes were placed in the “other” category. These particle

type classifications do not account for all of the mass spectral information obtained from

each particle, and many prior publications go into great depth regarding the presence of

212

other chemical species, such as hydroxymethanesulfonate [Whiteaker and Prather,

2003a], methanesulfonate [Gaston, et al., 2010], alkylamines [Angelino, et al.,

2001;Pratt, et al., 2009b], and specific metals [Ault, et al., 2010b;Moffet, et al., 2008b],

for example. Ammonium sulfate has been observed on a limited number of studies and

methods have been developed to identify its presence [Spencer, et al., 2008;Wenzel, et

al., 2003]. However, for the stated goal of providing information in an accessible manner

to modelers, general particle types and associated secondary species were chosen as the

focus of this comparison.

The single particle mass spectral signatures serve as fingerprints of particle

sources [Bhave, et al., 2001;Pratt and Prather, 2009;Toner, et al., 2008]. Previous

aerosol source studies have investigated the ATOFMS signatures of car emissions

[Gross, et al., 2000;Silva and Prather, 1997;Sodeman, et al., 2005], truck emissions

[Ferge, et al., 2006;Shields, et al., 2007;Toner, et al., 2006], ship emissions [Ault, et al.,

2010b;Healy, et al., 2009], biomass burning [Silva, et al., 1999], fireworks [Liu, et al.,

1997], grass mowing [Drewnick, et al., 2008], soil dust [Beddows and Telle, 2005;Silva,

et al., 2000], coal dust [Liu, et al., 2003], pesticides [Whiteaker and Prather, 2003a],

bacteria [Pratt, et al., 2009a], welding [Su, et al., 2005], brake dust [Beddows and Telle,

2005], and steel factories [Dall'Osto, et al., 2008]. Several laboratory studies have also

examined single particle mass spectral properties [Angelino, et al., 2001;Ferge, et al.,

2006;Gross, et al., 2006;Silva, et al., 2000;Silva and Prather, 2000;Spencer and Prather,

2006;Spencer, et al., 2006;Sullivan, et al., 2009b;Whiteaker and Prather,

2003a;Zimmermann, et al., 2003] and changes due to aging [Sullivan, et al., 2009b].

Numerous field studies in marine regions [Furutani, et al., 2008;Guazzotti, et al.,

213

2001a;Guazzotti, et al., 2001b;Spencer, et al., 2008] and regions impacted by dust

[Guazzotti, et al., 2001b;Sullivan, et al., 2007] have also provided source information.

Additionally, field studies have identified different source characteristics and how aging

changes them [Hughes, et al., 1999;Liu, et al., 2003;Moffet, et al., 2008a;Pratt and

Prather, 2009;Qin and Prather, 2006;Toner, 2007;Whiteaker and Prather,

2003b;Whiteaker, et al., 2002]. Together these ATOFMS mass spectral signatures

represent a robust framework that can be used to identify particle sources globally.

Chemical species desorbed and ionized by a laser have different affinities for

producing positive and negative ions [Murphy, 2007]. In general, positive ion mass

spectra can be linked to the primary source signature, and negative ion mass spectra can

provide information on mixing with secondary species [Pratt and Prather, 2009;Pratt

and Prather, 2010a]. As noted above, the six most prevalent general particle classes

globally are discussed in this paper: Fresh EC, ECOC, OC, Biomass, Sea Salt, and Dust.

These classifications refer to either the main chemical component by mass that was

present or the general source. Figure 7.2 shows representative spectra for each of these

particle classes; defining ion markers are noted in Table 7.2. Elemental carbon particles

(Fresh EC) are characterized by dominant carbon clusters in both the positive and

negative ion mass spectra (e.g., 12C1+/-,

24C2+/-, …, Cn

±) (Figure 7.2a) [Moffet and

Prather, 2009;Spencer and Prather, 2006]. Elemental carbon particles mixed with

organic carbon (ECOC) have large carbon cluster peaks in the positive ion mass spectrum

(Figure 7.2b), but mostly at low carbon numbers (i.e. n < 6), and a number of organic

carbon marker peaks, including 27C2H3+, C29

2H5+, C37

3H , and C+ 432H3O [Moffet and +

214

Prather, 2009;Spencer and Prather, 2006]. The ECOC negative ion mass spectra do not

contain

Figure 7.2 Representative average positive and negative ion mass spectra are shown for each of the particle classes: a) Fresh EC, b) ECOC, c) OC, d) Biomass, e) Sea Salt, and f) Dust.

215

Table 7.2 Defining characteristics of the six main particle types, including representative ion mass-to-charge (m/z) values and literature that the classification is based on.

216

significant carbon cluster ions and frequently have intense secondary markers of sulfate

and nitrate, as discussed below. Organic carbon (OC) particles are characterized by peaks

for 37C3H+ and 43C2H3O+, which are more intense than 12C+ and 36C3+ (the markers for

elemental carbon), as well as many other OC peaks including 27C2H3+ and 29C2H5

+

(Figure 7.2c) [Silva and Prather, 2000]. It should be noted that OC-dominant particles

can be formed through nucleation or accumulate significant OC through gas-particle

partitioning. ECOC is produced through the aging of fresh elemental carbon/soot through

the addition of OC [Moffet and Prather, 2009;Spencer and Prather, 2006]. For these OC

particles, OC, in addition to inorganic secondary species, is the primary source of

particulate mass [Spencer and Prather, 2006]. Often EC, ECOC, and OC particles are

from vehicular combustion sources [Pratt and Prather, 2009]; however, other sources

exist, detailed in the above-mentioned studies. Biomass burning particles are

characterized by a intense potassium peak (39K+), as well as many less intense peaks for

elemental carbon (12C+ and 36C3+) and organic carbon (27C2H3

+, 29C2H5+, 37C3H+, and

43C2H3O+) in the positive ions (Figure 7.2d) [Silva, et al., 1999]. Sea salt (SS) particles

are characterized by an intense sodium peak (23Na+) and markers for sodium chloride

(81,83Na2Cl+) (Figure 7.2e). Dust particles are characterized by positive ion inorganic

peaks (23Na+, 27Al+, 39K+, 40Ca+, 48Ti+, and/or 56Fe+) and their inorganic oxides (55KO+,

56CaO+, 64TiO+,70Al2O+, and/or 72FeO+) (Figure 7.2f). The presence of these inorganic

ions as well as often aluminum and silicon oxides (59AlO2-, 60SiO2

-, 76AlO2(OH)-, 76SiO3-

, and 77HSiO3-) in the negative ion mass spectra was the one method for identifying dust

particles as the dust mass spectra vary considerably particle to particle due to their

217

complex chemistry, morphology, and high absorption of 266 nm radiation [Sullivan, et

al., 2007].

To examine mixing with secondary species, searches were conducted on

representative ion markers (m/z) with the criterion that the peak area had to be at least 1%

relative area; in this way, the fraction of particles within each general particle class

containing each secondary species were identified. Secondary species investigated

include: oxidized organic carbon (43C2H3O+), ammonium (18NH4+), nitrate (46NO2

- and

62NO3-), and sulfate (80SO3

- and 97HSO4-).


ATOFMS studies were grouped into three location types: urban (Riverside, CA,

USA; St. Louis, MO, USA; Houston, TX, USA, Atlanta, GA, USA; Fresno, CA, USA;

Mexico City, Mexico; Detroit, MI, USA; London, United Kingdom; San Diego, CA,

USA; Los Angeles, CA, USA; Bologna, Italy; New York, NY, USA), marine (Maldives,

Maldives; Gosan, Korea; Cape Verde, Africa; Trinidad Head, CA, USA; North Atlantic,

Ireland; Mace Head, Ireland), and remote continental (Mt. Horeb, WI, USA; Sugar Pine

Dam, CA, USA; Grand Canyon, AZ, USA; Yellowstone, WY, USA). Urban sites are

referred to by the metropolitan area for simplicity, even if the sampling site was in an

adjacent community. The exact sampling site is listed in Table 1, with the metropolitan

area in parentheses. Given the large volume of urban sites (12 total), they were

subdivided into two groups: urban sites influenced by significant photochemistry,

oxidation, and aging (abbreviated Aged Urban) (Riverside, St. Louis, Houston, Atlanta,

Fresno, Mexico City, Detroit, and London) and urban sites without significant

photochemistry, oxidation, and aging (abbreviated Fresh Urban) (San Diego, Los

218

Angeles, Bologna, and New York). Studies have often been conducted at the same

location on multiple occasions, such as for San Diego [Guazzotti, et al., 2001b;Toner, et

al., 2008], Riverside [Hughes, et al., 1999;Liu, et al., 2000;Pastor, et al., 2003], and

Atlanta [Liu, et al., 2003;Middlebrook, et al., 2003;Wenzel, et al., 2003]; for this

comparison the most recent study was used at each of these sites.

7.4.1 Size-Resolved Chemical Composition

The size-resolved number fractions of the main particle types for each field site

are shown in Figures 7.3-7.6 and discussed below in terms of location type. By reporting

number fractions versus particle diameter, transmission biases are negated, and these

plots are comparable across different instruments with different inlets and transmission

efficiencies [Pratt, 2009]. Bins with fewer than 100 particles are not shown due to high

uncertainties based on statistics.

7.4.1.1 Urban Areas

7.4.1.1.1 Aged Urban Areas

A striking feature of size-resolved depictions of the general particle classes for

aged urban areas (Figure 7.3) was the significant number fraction (average 44%, range

26-68%) of OC particles across the submicron size range (~0.1-1.0 μm) for all 8 sites

(Riverside, London, Houston, Atlanta, Fresno, Mexico City, Detroit, St. Louis). Above 1

μm the number fraction of OC particles decreases substantially, and by 2 μm OC particles

were less than 13% by number at all sites. The primary source of OC particles in urban

areas is generally light duty vehicle emissions [Pratt and Prather, 2009]. Significant

aging of these small combustion particles via gas-particle partitioning results in

219

Figure 7.3 Size-resolved particle type number fractions for aged urban areas (Riverside, London, Houston, Atlanta, Fresno, Mexico City, Detroit, and St. Louis).

220

supermicron (>1 μm) particles [Pratt and Prather, 2009]. In aged urban areas, the

contribution of secondary organics to the total submicron organic aerosol mass can range

from ~45-90% [Docherty, et al., 2008]. The other major submicron particle types were

ECOC and Biomass particles, contributing ~20% (range 0-45%) and ~19% (range 1-

41%) by number, respectively. Fresh EC was generally confined to the smallest measured

particles (0.1-0.3 μm), as expected for fresh fossil fuel emissions [Moffet and Prather,

2009;Putaud, et al., 2004]. The major source of Fresh EC particles in many urban areas

is generally heavy duty diesel vehicles [Pratt and Prather, 2009]. As noted in the above

description, ECOC particles are aged Fresh EC particles [Moffet and Prather, 2009;Pratt

and Prather, 2009;Spencer and Prather, 2006]; further aging of Fresh EC and ECOC

particles can result in reclassification as OC particles due to the significant accumulation

of OC mass [Spencer and Prather, 2006]. Biomass burning is a significant global source

[Bowman, et al., 2009;Langmann, et al., 2009;Reid, et al., 2009], residing mostly in the

submicron size range[Qin and Prather, 2006], and is often part of the regional

background in both urban and rural areas[Qin and Prather, 2006]. Overall, little

contribution from sea salt or dust was observed in the submicron size range, although

significant contributions were observed between 0.5-1.0 μm for sea salt in London.

In the supermicron size range (1.0-3.0 μm), sea salt was the most abundant

particle class (47%) at sites within ~200 km of an ocean (Riverside, Houston, Fresno, and

London). Atlanta is 400 km, but included as having high sea salt due to hurricane

transport from the coast. Dust was a fraction (average – 16%, range - 1-32%) of

supermicron particles at most sites. While most supermicron carbonaceous particles are

aged combustion particles [Pratt and Prather, 2009], ATOFMS source studies have

221

shown supermicron Fresh EC particles from heavy duty diesel vehicle emissions [Shields,

et al., 2007] and supermicron biomass burning particles were observed in Mexico City

[Moffet, et al., 2008a]. In Detroit, St. Louis, and Mexico City, large number fractions

(~40-60%) of supermicron particles are metal-rich particles from nearby industry [Moffet,

et al., 2008a;Snyder, et al., 2009].

7.4.1.1.2 Fresh Urban Areas

33% of the urban field sites (San Diego, Los Angeles, Bologna, and New York

City) showed low number fractions (average - 12%, range - 3-26%) of submicron OC

particles that result from significant atmospheric processing (Figure 7.4). The most

abundant submicron particle class at each of these four sites was ECOC (average – 48%,

range 32-56%). This suggests that these sites are not significantly influenced by the

condensation of secondary organics and/or heterogeneous oxidation processes, as large

amounts of OC on particles can mask the primary source signature [Pratt and Prather,

2009;Spencer and Prather, 2006]. All four of these sites are within a few miles of oceans

and the diurnal land/sea breeze cycle may serve to limit the stagnant conditions that lead

to increased secondary organic carbon formation [Hughes, et al., 1999;Qin, et al., in

prep]. The aged urban sites with sea salt, discussed in the above section, were all located

further inland with a greater likelihood of having greater secondary organic carbon

formation for a variety of reasons, including local topography (Riverside and Fresno) and

high local emissions (London, Atlanta, and Houston). There was also a great deal of

variability in the number fractions of Fresh EC and Biomass particles, likely dependent

on local emission amounts. However, Fresh EC was still among the smallest measured

222

Figure 7.4 Size-resolved particle type number fractions for fresh urban areas (San Diego, Los Angeles, Bologna, and New York).

223

particles [Putaud, et al., 2004], and Biomass particles were distributed across the

submicron size range, as previously observed for fossil fuel combustion and biomass

burning emissions [Clarke, et al., 2007]. In the supermicron size range, sea salt was the

dominant particle type across all four sites, as these sites were not located near strong

dust source regions.

7.4.1.2 Remote Continental Sites

The size-resolved chemistry of the rural/remote continental locations is shown in

Figure 7.5. Of the field studies in remote continental locations, three sites (Mt. Horeb,

Sugar Pine Dam, and Grand Canyon) were sampled during representative time periods,

while the fourth site (Yellowstone) was sampled during intense wildfire conditions,

common to many portions of the western United States (Figure 7.6). Wildfire conditions

are prevalent in western United states during the fall [Reid, et al., 2009], encompassing

large swaths of land [Urbanski, et al., 2009], and altering local air quality [Viswanathan,

et al., 2006]. The Mt. Horeb, Sugar Pine Dam, and the Grand Canyon studies were

characterized high number fractions of Biomass particles in the submicron and a high

fraction of Dust in the supermicron size range. All four sites also had lower OC particle

number fractions (average 22%, range 8-36%) compared to the Aged Urban sites.

Previous filter-based measurements have shown lower PM2.5 mass fractions at locations >

50 km away from urban sources [Putaud, et al., 2004]. The Sugar Pine Dam site was

influenced by sea salt particles, even though it is ~250 km inland, which was likely

attributable to intense storms with strong winds transporting sea salt to the site [Ault, et

al., 2010a]. The high number fraction of submicron Fresh EC at Mt. Horeb was likely

224

Figure 7.5 Size-resolved particle type number fractions for remote continental sites (Mt. Horeb, Sugar Pine, Grand Canyon, and Yellowstone).

225

due to local sources, such as tractors, in the farm setting of the study. The extremely high

fraction of dust in the supermicron size range at the Grand Canyon is likely due to the

fact that it is an arid dust-producing environment [Brewer and Moore, 2009;Park, et al.,

2009].

As noted above, the Yellowstone study was completed during intense wildfires

within the National Park [Hall, et al., 2006]. While, these conditions are not

representative of normal conditions in Yellowstone, the size-resolved chemistry during

this study is discussed herein given the frequency and broad impact of wildfires in the

western United States [Brewer and Moore, 2009;Urbanski, et al., 2009] and many other

regions globally these results are discussed [Bowman, et al., 2009;Langmann, et al.,

2009;Reid, et al., 2009]. The close proximity of the wildfires to the sampling site led to

biomass, OC, and ECOC particles constituting greater than 80% of the particles by

number across both the submicron and supermicron size ranges. The remote continental

studies demonstrate the importance of biomass burning and dust in non-urban settings, as

well as the sensitivity of particle chemistry to local sources that would likely not be

observed in urban settings due elevated particle concentrations from numerous

anthropogenic sources.

7.4.1.3 Marine Studies

For the six marine studies, the dominate particle type in the supermicron size

range was sea salt, as expected (Figure 7.6). Significant number fractions of sea salt

(average – 49%, range – 1-91%), were also observed from 0.5-1.0 μm. For the two sites

that were influenced by outflow from Asia (Maldives and Gosan), a significant fraction

226

Figure 7.6 Size-resolved particle type number fractions for marine studies (Gosan, Cape Verde, Maldives, Trinidad Head, Mace Head, and North Atlantic).

227

of supermicron dust was observed, due to long range transport from the major Saharan

and Asian dust source regions [Ault, et al., 2010c;Satheesh and Moorthy, 2005;Spencer,

et al., 2008]. During periods heavily influenced by dust source regions, marine locations

can even be characterized by higher number and mass concentrations of dust than sea salt

[Ault, et al., 2010c]. In the submicron size range, the six marine sites can be split into

four sites with high number fractions of Biomass particles (Maldives, Gosan, Cape

Verde, and Trinidad Head) and 2 sites not influenced by biomass burning (North Atlantic

and Mace Head). This difference is striking as the sites in Asia, North America, and

Africa all have similar fractions of carbonaceous particle types, but the sites in Europe do

not. This is not merely an instrumental issue as the same instrument was used for Cape

Verde, North Atlantic, and Mace Head. One possible explanation is that far greater

numbers of large fires occur in Africa, Asia, and North America in comparison to Europe

[Bowman, et al., 2009;Langmann, et al., 2009;Reid, et al., 2009], leading to much lower

Biomass particle number concentrations near Mace Head and the North Atlantic. Overall,

the marine studies divided neatly into 2 subclassifications based on the presence or

absence of Biomass particles, due related to the global distribution of biomass burning

emissions [Bowman, et al., 2009;Langmann, et al., 2009;Reid, et al., 2009].

The size-resolved chemistry for each marine study was separated into clean and

polluted time periods, based on whether aged Sea Salt or fresh Sea Salt was more

prevalent (Figure 7.7). The aged sea salt type has larger nitrate peaks than chloride, while

fresh sea salt is the converse, as has been shown previously [Gard, et al., 1998]. For

particle types other than Sea Salt, the size-resolved chemical class fractions were very

similar for the clean and polluted periods, with the submicron size range dominated by

228

Figure 7.7 Size-resolved particle type number fractions for clean and polluted time periods during three marine studies (Gosan, Maldives, and Trinidad Head).

229

ECOC and Biomass. This interesting finding demonstrates that, while the number

concentration of submicron particles decreases during clean time periods [Ault, et al.,

2010c], that the number fraction of particles are still mostly from combustion classes

(ECOC and Biomass).

7.4.2 Internal Mixing with Secondary Species

In the following sections, the internal mixing of each of the six general particle

classes is described for the following secondary species markers: oxidized organic carbon

(43C2H3O+, 43C2H3O-), ammonium (18NH4+), nitrate (46NO2

- and 62NO3-), and sulfate

(80SO3- and 97HSO4

-). The mixing state discussed below is for consideration of particles

across all size ranges (0.1-3.0 μm). Mixing with respect to submicron versus supermicron

diameter particles was investigated separately, to minimize any size bias, but the mixing

state by number fraction for each particle type was similar for both its submicron and

supermicron components at most sites and is presented together. This section looks only

at the fraction of particles mixed with each secondary marker. As particles aged in the

atmosphere and they become larger and the mass of the secondary species present

increases [Spencer and Prather, 2006]. Often particles take up water during atmospheric

aging [Seinfeld and Pandis, 2006]. In real-time laser desorption/ionization mass

spectrometry , the presence of water can cause the suppression of negative ion formation,

causing many aged particles sampled in ambient environments to lack negative ions

[Ault, et al., 2009;Neubauer, et al., 1998]. For this analysis only particles producing both

positive and negative ions were considered, while leading to a small under representation

of secondary species, the qualitative trends observed as a function of particle type are still

valuable for explaining the overall characteristics of particle mixing state. In general,

230

across all location types (Figures 7.8-7.12), there is large variability in the number

fractions of the general particle types mixed with each secondary species; the following

sections describe these mixing states for each location type.

7.4.2.1 Urban Areas

7.4.2.1.1 Aged Urban Areas

The complexity of particle mixing state is evident from the lack of straight

forward trends across the urban areas for similar particle classes (Figures 7.8 and 7.9);

this suggests that atmospheric processing is not consistent across all urban locations,

particularly for all particle types. The urban areas with significant OC particle

concentrations, discussed in this section, are expected to represent the sites most

impacted by secondary species, given the large number of particles that have taken up

secondary organic material. For the OC particle type, 78-95% of the particles by number

contain oxidized OC (43C2H3O+). In particular, Riverside, Fresno, and Mexico City were

characterized by particularly high fractions (91-95%) of OC particles internally mixed

with oxidized OC. This is consistent with high submicron mass fractions of oxidized

organics and high O:C ratios in atmospheric aerosols sampled in urban locations and

those downwind of urban areas [Zhang, et al., 2007]. The increased oxidation of the

organic aerosol is due to gas-particle partitioning of semi-volatile species and

heterogeneous oxidation processes during particle aging [Heald, et al., 2010]. These

results are further supported by the observation that higher number fractions of Fresh EC

and ECOC particles were internally mixed with oxidized organic carbon (46% and 58%,

231

Figure 7.8 Particle mixing state matrices for aged urban areas, depicting the number fractions of each particle type (Fresh EC, ECOC, OC, Biomass, Sea Salt, and Dust) during each study that were internally mixed with each secondary species (oxidized organic carbon (43C2H3O+), ammonium (NH4

+), nitrate (46NO2- and 62NO3

-), and sulfate (80SO3

- and 97HSO4-)).

232

Figure 7.9 Particles mixing state matrices for fresh urban areas, depicting the number fractions of each particle type (Fresh EC, ECOC, OC, Biomass, Sea Salt, and Dust) during each study that were internally mixed with each secondary species (oxidized organic carbon (43C2H3O+), ammonium (NH4



- and 97HSO4-)).

233

respectively) than for the same particle types at fresh urban (16% and 49%), remote

continental (12% and 54%), and marine (49% and 17%) locations.

However, despite the strong presence of secondary organic carbon, very few Sea

Salt or Dust particles are internally mixed with oxidized organic carbon, with less than

25% of those particle mass spectra containing 43C2H3O+ for all 6 studies with sea salt and

dust present. However, sea salt and dust particles were mixed strongly with nitrate (81%

and 89% by number, respectively); combined with a lack of ammonium present, it is

likely that species such as NaNO3 are present due to heterogeneous reactions between

nitric acid and the marine (Sea Salt) and crustal (Dust) species [Gard, et al., 1998]. The

form and degree of aging may also be related to the chemistry of the dust, as shown

previously [Sullivan, et al., 2007].The mixing of the submicron carbonaceous types

(Fresh EC, ECOC, OC, and Biomass) with nitrate is far less consistent, suggesting that

gas phase concentrations and species that these particles are exposed to vary considerably

site-to-site. Sulfate tends to be concentrated on the carbonaceous particle types and is

observed on few sea salt or dust particles, while ammonium does not hold a consistent

trend across these sites.

7.4.2.1.2 Fresh Urban Areas

The fresh urban areas tend to have fewer Fresh EC (17%) and ECOC (50%)

particle by number internally mixed with oxidized organic carbon (Fig. 7.9) than the aged

urban areas (46% and 58%, respectively). The decreased number fractions of Fresh EC

and ECOC particles mixed with oxidized organic carbon supports the discussion in

section 7.3.1.1.2 showing that these sites are less aged both in terms of the particle

234

classes present and the mixing of secondary species with those classes, likely due to local

meteorology preventing stagnation at the sites. Also, 11% of the dust particles by number

at all four of these sites were mixed with oxidized OC, which is slightly less than that

observed for dust particles in aged urban environments (16%). Similar to the aged urban

locations, the most common secondary species internally mixed with sea salt and dust

particles was nitrate, while sulfate and ammonium are observed on very few of these

particles. When present, ammonium is mixed with ECOC, OC, and Biomass particles, as

is sulfate, suggesting the presence of ammonium sulfate. The fraction of carbonaceous

particles mixed with nitrate is also inconsistent across the four sites. Higher number

fractions of carbonaceous particle classes are mixed with nitrate than sulfate at Los

Angeles and Bologna, while the reverse is true in San Diego, and the fractions are very

similar in New York. Thus, while the mixing state of the sea salt and dust, primarily in

the supermicron size range, is predictable across all four sites, the mixing state of the

carbonaceous particle types, primarily observed in the submicron size range, is not

consistent across all four sites.

7.4.2.2 Remote Continental Sites

For the remote continental sites, Dust (and Sea Salt when present) were most

likely to be mixed with nitrate and were not mixed with sulfate (Figure 7.10). The mixing

of sulfate with carbonaceous particle classes varies considerably between sites, with

ECOC, OC, and Biomass particles highly mixed at Sugar Pine Dam (99%, 100%, and

78%, by number, respectively) and much less mixed at the Grand Canyon (14%, 6%,

21%). A similar trend is observed for nitrate with ECOC, OC, and Biomass particles,

235

Figure 7.10 Particle mixing state matrices for remote continental sites, depicting the number fractions of each particle type (Fresh EC, ECOC, OC, Biomass, Sea Salt, and Dust) during each study that were internally mixed with each secondary species (oxidized organic carbon (43C2H3O+), ammonium (NH4



- and 97HSO4-)).

236

wherein they were highly mixed at Sugar Pine Dam (58%, 79%, and 70% by number,

respectively), but barely mixed at the Grand Canyon (4%, 2%, 17%). Sugar Pine Dam is

the only remote continental site with a high fraction of ammonium, which combined with

the high sulfate and nitrate fractions, suggests the presence of ammonium sulfate and

ammonium nitrate. It was the most polluted of the remote sites, likely due to transport

from California’s polluted Central Valley [Bao, et al., 2008]. The mixing of oxidized

organic carbon with OC particles (average - 86%) is consistent with the observed high

oxidized organic fractions of the total submicron organic aerosol mass in rural/remote

locations [Zhang, et al., 2007]. Both Mt. Horeb (71% and 58% for ECOC and Biomass)

and Yellowstone (81% and 76%) have ECOC and Biomass types highly mixed with

sulfate and to a lesser extent nitrate, but the OC particle type is not mixed with sulfate

and nitrate at either Mt. Horeb (5% and 4%) or Yellowstone (15% and 10%), as it was in

the urban areas. The remote continental sites are quite different in terms of how the

particle classes are mixed with secondary species. This reinforces the conclusion above

that in an environment with fewer sources, particles class fractions, as well as how they

are mixed, vary considerably.

7.4.2.3 Marine Studies

At the marine sites, many of the same mixing state trends observed in urban areas

continued, while a couple of notable differences emerged (Figure 7.11). Once again, sea

salt and, especially, dust were mixed with nitrate across all of the sites and were not

significantly internally mixed with oxidized organic carbon or ammonium. Oxidized

organic carbon, when present, was mixed with carbonaceous species. A notable

237

Figure 7.11 Particle mixing state matrices for marine studies, depicting the number fractions of each particle type (Fresh EC, ECOC, OC, Biomass, Sea Salt, and Dust) during each study that were internally mixed with each secondary species (oxidized organic carbon (43C2H3O+), ammonium (NH4



- and 97HSO4-)).

238

difference from the urban areas is that both dust and sea salt were mixed with significant

fractions of sulfate at many sites, which may result from the oxidation of ocean-derived

SO2 [Gaston, et al., 2010], although the fractions containing sulfate (up to 66 %) were

consistently lower than the fractions containing nitrate. Sulfate was also consistently

internally mixed with the carbonaceous particle types (Fresh EC, ECOC, OC, and

Biomass). The mixing of Fresh EC, ECOC, and Biomass classes with nitrate was quite

variable, ranging from moderate number fractions at Trinidad Head (64%, 49%, and

57%, respectively) and Gosan (37%, 42%, and 44%), compared to very low fractions in

the Maldives (<1%, <1%, and 4%) and North Atlantic (4% and < 1%, Fresh EC and

ECOC), likely due to influences from different upwind sources. The particle class most

likely to be mixed with nitrate and sulfate was consistently the OC particle type (55% and

69%, respectively). Overall, the trends were in particle mixing state were far more

consistent at the marine sites compared to the urban sites, which is logical as most

particles observed at these sites have had long atmospheric lifetimes and their exposure to

different gases becomes homogenized.

The differences in marine particle mixing state between clean and polluted time

periods are shown in Figure 7.12 for the Gosan, Maldives, and Trinidad Head studies.

During polluted marine time periods at Gosan, Sea Salt and Dust particles were highly

internally mixed with nitrate (92% and 89%, by number) with little mixing with sulfate

(~10%, Sea Salt has interference at 93NaCl2-). During the clean period, the carbonaceous

particle types (Fresh EC, ECOC, OC, and Biomass) were highly mixed with sulfate

(80%, 99%, 92%, and 97% by number, respectively) and barely mixed with nitrate (15%,

21%, 8%, and 16%). The high number fractions of Sea Salt and Dust particles mixed with

239

Figure 7.12 Particle mixing state matrices for clean and polluted time periods during 3 marine studies (Gosan, Maldives, Trinidad Head), depicting the number fractions of each particle type (Fresh EC, ECOC, OC, Biomass, Sea Salt, and Dust) during each study that were internally mixed with each secondary species (oxidized organic carbon (43C2H3O+), ammonium (NH4


-), and sulfate (80SO3- and 97HSO4

-)).

240

nitrate and carbonaceous particles mixed with sulfate continued during the polluted time

periods, but the number fraction of carbonaceous particles (Fresh EC, ECOC, OC,

Biomass) mixed with nitrate increases dramatically from the clean to polluted time

period, from 8-21% to 33-68%, by number, for these four particle types. This indicates

that nitrate is taken up during the aging process on to the carbonaceous particle types.

The Maldives field campaign had similarities in that the number fractions of the Fresh

EC, ECOC, and Biomass particle types internally mixed with sulfate were high for both

clean (86%, 100%,and 95% by number, respectively) and polluted time periods (83%,

98%, 93% by number, respectively), and the Sea Salt and Dust particles are highly mixed

with nitrate during both clean (57% and 69% by number, respectively) and polluted (51%

and 89% by number, respectively) time periods, as well. The key difference between the

Gosan and Maldives studies is that there is no increase in nitrate number fraction (< 2%

mixed) for either clean or polluted conditions during the Maldives study for carbonaceous

particles. This indicates that carbonaceous particles sampled in the Maldives were likely

not exposed to as much nitric acid or other nitrate precursors during transport. For

Trinidad Head there is greater variation in the internally mixed number fractions for each

particle type during the clean and polluted periods, but this location has similar overall

trends to the Maldives study, in that nitrate is not enhanced during polluted periods after

being absent for the clean periods particles. Overall this comparison shows that the

degree of aging is likely not related to the time in transit to the site, but what gas phase

concentrations the particles are exposed to en route.

241

7.5 Conclusions

The current scientific understanding of aerosols with respect to radiative forcing

is low [Solomon, et al., 2007]. In addition to research needs in understanding the

fundamental physical and chemical properties of aerosols, there is currently a disconnect

between empirical findings and the representation of aerosols in most models. Given the

large uncertainties associated with radiative forcing due to aerosol direct and indirect

effects [Solomon, et al., 2007], it is essential that improvements be made to the

representation of aerosols in models, particularly global climate models (GCMs), In the

2007 IPCC report, the general aerosol categories were reported as: black carbon, organic

carbon, biomass burning, sea salt, mineral dust (for some models), sulfate, and nitrate (for

some models) [Solomon, et al., 2007]. With the exception of biomass burning, particles in

these GCMs were considered to be externally mixed [Solomon, et al., 2007], which is not

indicative of aerosols in the atmosphere. Further, internal versus external mixing has been

shown to drastically alter the radiative impact of black carbon and sulfate [Jacobson,

2001]. Fifth and sixth generation GCMs are currently improving their representation of

aerosols and are moving away from exclusively using external mixtures to using both

internal and external mixtures [Ghan and Schwartz, 2007]. However, a great deal about

the mixing of chemical species within different atmospheric aerosols remains to be

incorporated into climate modeling. A lack of mixing state measurements has been also

identified as one of the limiting factors in fully characterizing the ability of cloud

condensation nuclei to activate and form cloud droplets [McFiggans, et al., 2006]. Single

particle mass spectrometry measurements can provide information on the chemical

mixing state, including both refractory and non-refractory species, of single atmospheric

242

particles. Herein, we begin to address this gap in knowledge by reviewing size-resolved,

single-particle mixing state data from 22 ATOFMS studies, primarily in the Northern

Hemisphere.

Six main particle classes (Fresh EC, ECOC, OC, Biomass, Sea Salt, and Dust

particles) were observed to represent on average 94% of particles by number (95% and

92% in the submicron (0.1-1.0 μm) and supermicron (1.0-3.0 μm) size ranges,

respectively). A significant fraction of the particles within the “other” category were

metal-containing particles in the supermicron size range that were produced from

industrial processes (i.e. Pb-Zn-Cl particles in Mexico City [Moffet, et al., 2008b]) and

metals mixed with carbonaceous species from combustion (i.e. vanadium mixed with

organic carbon [Ault, et al., 2010b]) in the submicron size range. While not characterized

in this overview, metal-containing particles have important implications for human health

[Chen, et al., 2007;Liu, et al., 2005] and as potential ice nuclei [Cziczo, et al., 2009b]. In

contrast to the 2007 IPCC report, which treated sulfate and nitrate as externally mixed,

these 22 ATOFMS studies observed sulfate and nitrate nearly exclusively as internal

mixtures with the six particle types listed above and not as external mixtures. The

assumptions of mixing state have particularly important ramifications for black carbon

and sulfate [Jacobson, 2001], necessitating improved representation of sulfate (and

nitrate) mixing state in future models.

The 22 ATOFMS studies were grouped into urban, remote continental, and

marine locations. Analysis of size-resolved, single-particle chemistry data from the 12

urban studies revealed two subgroups: 1) aged urban environments with high number

fractions of OC particles in the submicron size range and 2) fresh urban environments

243

with high number fractions of ECOC and biomass particles. The photochemical

oxidation/aging processes in the aged urban environments led to high number fractions of

OC particles, many of which were likely characterized by soot particle cores coated with

secondary species, particularly oxidized OC. These sites also had the highest number

fractions of secondary species (oxidized organic carbon, nitrate, sulfate, and ammonium)

mixed with all of the particle classes. Overall, particles in the fresh urban environments

were less internally mixed with secondary species. With the exception of a biomass

burning influenced site, the remote continental sites showed the greatest variability in

aerosol size and composition with Fresh EC (Mt. Horeb), Biomass/Sea Salt (Sugar Pine

Dam), and Dust (Grand Canyon) as the largest fraction of particles by number over the

accumulation and start of the coarse mode, due to intense sources locally. The marine

studies had two particular findings of note: 1) very low number fractions of OC particles

were observed at the 6 marine sites, and 2) carbonaceous particles (Fresh EC, EOC, OC,

Biomass) comprised the majority of particles below 0.5 μm (typically 80-95% by

number) during all marine studies. The number fraction of carbonaceous particles below

0.5 μm was not significantly lower during isolated “clean” periods with sulfate internally

mixed with carbonaceous particles and nitrate mixed with sea salt and dust. The

prevalence of submicron carbonaceous particles has important implications for the

treatment of CCN in “pristine” maritime atmospheres.

Across all studies, sea salt and dust particles were preferentially internally mixed

with nitrate, while low number fractions of these particles were internally mixed with

sulfate, ammonium, or oxidized organic carbon. The carbonaceous particle types (Fresh

EC, ECOC, and Biomass), observed primarily in the submicron size range were most

244

likely to be mixed with sulfate, although nitrate was also prevalent in aged/processed

environments. Significantly variability of in the mixing of secondary species with each

particle type was observed between studies, likely due to influences from different

atmospheric aging mechanisms. This demonstrates the need to incorporate sophisticated

methods for representing aerosol aging processes, rather than simply assuming mixing

based on aging timescales.

Many steps are currently being taken to more accurately represent aerosol mixing

state in climate models to reduce uncertainties associated with aerosol radiative forcings

[Ghan and Schwartz, 2007]. The findings of this overview are intended to provide more

comprehensive information on particle mixing state to provide inputs/assumptions for

climate models. It is clear that the particle categories discussed in the 2007 IPCC report

do not accurately depict aerosol mixing state. Future work is needed to measure the size-

resolved, single-particle mixing state of particles globally, particularly in remote

locations, where the greatest variability in chemistry was observed. Reduced

uncertainties in the estimated radiative forcings by the aerosol direct and indirect effects

are expected as modules are developed to more accurately depict aerosol processes as

they are observed in the atmosphere and as models are compared with ambient

observations.


The following present and former Prather group members are thanked for

assisting with data collection: Keith Coffee, Jessie Creamean, Hiroshi Furutani,

Cassandra Gaston, John Holecek, George Khairallah, Don-Yuan Liu, Joseph Mayer,

Ryan Moffet, Silva Pastor, Xueying Qin, Michelle Sipin, Matthew Spencer, and David

245

Suess. Data collection by the Gross group was assisted by James Schauer, Alexandra

Schmidt, and Melanie Yuen. Funding for the ATOFMS field studies was provided by:

Atmospheric Brown Cloud project under the United Nations Environmental Programme,

California Air Resources Board (CARB), UK Natural Environment Research Council,

California Energy Commission (CEC), Electric Power Research Institute (EPRI),

Environmental Protection Agency (including the EPA-STAR program), LADCO,

Michigan Department of Natural Resources (MDNR), National Oceanic & Atmospheric

Administration (NOAA), and National Science Foundation (NSF). Andrew Ault is

grateful for a Department of Energy Global Change Education Program graduate research

environmental fellowship. Kerri Pratt is grateful for a NSF graduate research fellowship

and a NOAA climate & global change postdoctoral fellowship. Jerome Fast and

Xiaohong Liu (Pacific Northwest National Laboratory) are thanked for discussions of the

motivation of this manuscript.

Chapter 7 is in preparation for submission to Atmospheric Environment: A. P.

Ault, A. Zhu, K. A. Pratt, D. S. Gross, M. Dall’Osto, M. Gaelli, C. O’Dowd, and K.A.

Prather, Characterization of Asian Outflow at Gosan, Korea by Single Particle Mass

Spectrometry.

246

7.7 References

S. Angelino, D. T. Suess and K. A. Prather, Formation of aerosol particles from reactions of secondary and tertiary alkylamines: Characterization by aerosol time-of-flight mass spectrometry, Environmental Science & Technology, 35 (15), 3130-3138, 2001.

A. P. Ault, J. M. Creamean, C. J. Gaston, C. R. Williams, F. M. Ralph and K. A. Prather,

Observations of Transported Asian Dust in California Precipitation during CalWater, Nat. Geosci., submitted, 2010a.

A. P. Ault, C. J. Gaston, Y. Wang, G. Dominguez, M. H. Thiemens and K. A. Prather,

Characterization of the Single Particle Mixing State of Individual Ship Plume Events Measured at the Port of Los Angeles, Environmental Science & Technology, 44 (6), 1954-1961, 2010b.



A. P. Ault, A. Zhu, V. Ramanathan and K. A. Prather, Characterization of Asian Outflow

at Gosan, Korea by Single Particle Mass Spectrometry, Journal Of Geophysical Research-Atmospheres, in prep, 2010c.

J. W. Bao, S. A. Michelson, P. O. G. Persson, I. V. Djalalova and J. M. Wilczak,

Observed and WRF-simulated low-level winds in a high-ozone episode during the Central California Ozone Study, Journal of Applied Meteorology and Climatology, 47 (9), 2372-2394, 2008.

S. E. Bauer, D. L. Wright, D. Koch, E. R. Lewis, R. McGraw, L. S. Chang, S. E.

Schwartz and R. Ruedy, MATRIX (Multiconfiguration Aerosol TRacker of mIXing state): an aerosol microphysical module for global atmospheric models, 8 (20), 6003-6035, 2008.

D. C. S. Beddows and H. H. Telle, Prospects of real-time single-particle biological

aerosol analysis: A comparison between laser-induced breakdown spectroscopy and aerosol time-of-flight mass spectrometry, Spectrochimica Acta Part B-Atomic Spectroscopy, 60 (7-8), 1040-1059, 2005.

P. V. Bhave, D. P. Fergenson, K. A. Prather and G. R. Cass, Source apportionment of

fine particulate matter by clustering single-particle data: Tests of receptor model accuracy, Environmental Science & Technology, 35 (10), 2060-2072, 2001.

247

D. Bowman, J. K. Balch, P. Artaxo, W. J. Bond, J. M. Carlson, M. A. Cochrane, C. M. D'Antonio, R. S. DeFries, J. C. Doyle, S. P. Harrison, F. H. Johnston, J. E. Keeley, M. A. Krawchuk, C. A. Kull, J. B. Marston, M. A. Moritz, I. C. Prentice, C. I. Roos, A. C. Scott, T. W. Swetnam, G. R. van der Werf and S. J. Pyne, Fire in the Earth System, Science, 324 (5926), 481-484, 2009.

P. Brewer and T. Moore, Source Contributions to Visibility Impairment in the

Southeastern and Western United States, Journal Of The Air & Waste Management Association, 59 (9), 1070-1081, 2009.

W. Cantrell and A. Heymsfield, Production of ice in tropospheric clouds - A review,

Bulletin Of The American Meteorological Society, 86 (6), 795-+, 2005. J. C. Chen, J. M. Cavallari, P. H. Stone and D. C. Christiani, Obesity is a modifier of

autonomic cardiac responses to fine metal particulates, Environmental Health Perspectives, 115 (7), 1002-1006, 2007.

S. H. Chung and J. H. Seinfeld, Global distribution and climate forcing of carbonaceous

aerosols, Journal Of Geophysical Research-Atmospheres, 107 (D19), 2002. A. Clarke, C. McNaughton, V. Kapustin, Y. Shinozuka, S. Howell, J. Dibb, J. Zhou, B.

Anderson, V. Brekhovskikh, H. Turner and M. Pinkerton, Biomass burning and pollution aerosol over North America: Organic components and their influence on spectral optical properties and humidification response, Journal Of Geophysical Research-Atmospheres, 112 (D12), 2007.

B. Croft, U. Lohmann and K. von Salzen, Black carbon ageing in the Canadian Centre for

Climate modelling and analysis atmospheric general circulation model, Atmospheric Chemistry And Physics, 5, 1931-1949, 2005.

M. J. Cubison, B. Ervens, G. Feingold, K. S. Docherty, I. M. Ulbrich, L. Shields, K.

Prather, S. Hering and J. L. Jimenez, The influence of chemical composition and mixing state of Los Angeles urban aerosol on CCN number and cloud properties, Atmospheric Chemistry And Physics, 8 (18), 5649-5667, 2008.

D. J. Cziczo, K. D. Froyd, S. J. Gallavardin, O. Moehler, S. Benz, H. Saathoff and D. M.

Murphy, Deactivation of ice nuclei due to atmospherically relevant surface coatings, Environmental Research Letters, 4 (4), 2009a.

D. J. Cziczo, O. Stetzer, A. Worringen, M. Ebert, S. Weinbruch, M. Kamphus, S. J.

Gallavardin, J. Curtius, S. Borrmann, K. D. Froyd, S. Mertes, O. Mohler and U. Lohmann, Inadvertent climate modification due to anthropogenic lead, Nature Geoscience, 2 (5), 333-336, 2009b.

248

M. Dall'Osto, M. J. Booth, W. Smith, R. Fisher and R. M. Harrison, A study of the size distributions and the chemical characterization of airborne particles in the vicinity of a large integrated steelworks, Aerosol Science And Technology, 42 (12), 981-991, 2008.

K. A. Denkenberger, R. C. Moffet, J. C. Holecek, T. P. Rebotier and K. A. Prather, Real-

time, single-particle measurements of oligomers in aged ambient aerosol particles, Environmental Science & Technology, 41 (15), 5439-5446, 2007.

K. S. Docherty, E. A. Stone, I. M. Ulbrich, P. F. DeCarlo, D. C. Snyder, J. J. Schauer, R.

E. Peltier, R. J. Weber, S. M. Murphy, J. H. Seinfeld, B. D. Grover, D. J. Eatough and J. L. Jimenez, Apportionment of Primary and Secondary Organic Aerosols in Southern California during the 2005 Study of Organic Aerosols in Riverside (SOAR-1), Environmental Science & Technology, 42 (20), 7655-7662, 2008.

F. Drewnick, M. Dall'Osto and R. Harrison, Characterization of aerosol particles from

grass mowing by joint deployment of ToF-AMS and ATOFMS instruments, Atmospheric Environment, 42 (13), 3006-3017, 2008.

T. Ferge, E. Karg, A. Schroppel, K. R. Coffee, H. J. Tobias, M. Frank, E. E. Gard and R.

Zimmermann, Fast determination of the relative elemental and organic carbon content of aerosol samples by on-line single-particle aerosol time-of-flight mass spectrometry, Environmental Science & Technology, 40 (10), 3327-3335, 2006.

P. Forster, V. Ramaswamy, P. Artaxo, T. Berntsen, R. Betts, D. W. Fahey, J. Haywood,

J. Lean, D. C. Lowe, G. Myhre, J. Nganga, R. Prinn, G. Raga, M. Schulz and R. Van Dorland, Changes in Atmospheric Constituents and in Radiative Forcing 2007.

H. Furutani, M. Dall'osto, G. C. Roberts and K. A. Prather, Assessment of the relative

importance of atmospheric aging on CCN activity derived from field observations, Atmos. Environ., 42 (13), 3130-3142, 2008.

S. Fuzzi, M. O. Andreae, B. J. Huebert, M. Kulmala, T. C. Bond, M. Boy, S. J. Doherty,

A. Guenther, M. Kanakidou, K. Kawamura, V. M. Kerminen, U. Lohmann, L. M. Russell and U. Poschl, Critical assessment of the current state of scientific knowledge, terminology, and research needs concerning the role of organic aerosols in the atmosphere, climate, and global change, Atmospheric Chemistry And Physics, 6, 2017-2038, 2006.



249

E. E. Gard, M. J. Kleeman, D. S. Gross, L. S. Hughes, J. O. Allen, B. D. Morrical, D. P. Fergenson, T. Dienes, M. E. Galli, R. J. Johnson, G. R. Cass and K. A. Prather, Direct observation of heterogeneous chemistry in the atmosphere, Science, 279 (5354), 1184-1187, 1998.

C. J. Gaston, K. A. Pratt, X. Y. Qin and K. A. Prather, Real-Time Detection and Mixing

State of Methanesulfonate in Single Particles at an Inland Urban Location during a Phytoplankton Bloom, Environmental Science & Technology, 44 (5), 1566-1572, 2010.

S. J. Ghan and S. E. Schwartz, Aerosol properties and processes - A path from field and

laboratory measurements to global climate models, Bulletin Of The American Meteorological Society, 88 (7), 1059-+, 2007.

D. S. Gross, R. Atlas, J. Rzeszotarski, E. Turetsky, J. Christensen, S. Benzaid, J. Olson,

T. Smith, L. Steinberg, J. Sulman, A. Ritz, B. Anderson, C. Nelson, D. R. Musicant, L. Chen, D. C. Snyder and J. J. Schauer, Environmental chemistry through intelligent atmospheric data analysis, Environmental Modelling & Software, 25 (6), 760-769, 2010.

D. S. Gross, M. E. Galli, M. Kalberer, A. S. H. Prevot, J. Dommen, M. R. Alfarra, J.

Duplissy, K. Gaeggeler, A. Gascho, A. Metzger and U. Baltensperger, Real-time measurement of oligomeric species in secondary organic aerosol with the aerosol time-of-flight mass spectrometer, Analytical Chemistry, 78 (7), 2130-2137, 2006.

D. S. Gross, M. E. Galli, P. J. Silva, S. H. Wood, D. Y. Liu and K. A. Prather, Single

particle characterization of automobile and diesel truck emissions in the Caldecott Tunnel, Aerosol Science And Technology, 32 (2), 152-163, 2000.

S. A. Guazzotti, K. R. Coffee and K. A. Prather, Continuous measurements of size-

resolved particle chemistry during INDOEX-Intensive Field Phase 99, Journal Of Geophysical Research-Atmospheres, 106 (D22), 28607-28627, 2001a.


measurements of the chemical composition of size-resolved particles during a Santa Ana wind episode, California USA, Atmospheric Environment, 35 (19), 3229-3240, 2001b.

B. D. Hall, M. L. Olson, A. P. Rutter, R. R. Frontiera, D. P. Krabbenhoft, D. S. Gross, M.

Yuen, T. M. Rudolph and J. J. Schauer, Atmospheric mercury speciation in Yellowstone National Park, Science Of The Total Environment, 367 (1), 354-366, 2006.

C. L. Heald, J. H. Kroll, J. L. Jimenez, K. S. Docherty, P. F. DeCarlo, A. C. Aiken, Q.

Chen, S. T. Martin, D. K. Farmer and P. Artaxo, A simplified description of the

250

evolution of organic aerosol composition in the atmosphere, Geophysical Research Letters, 37, 2010.

R. M. Healy, I. P. O'Connor, S. Hellebust, A. Arnaud, J. R. Sodeau and J. C. Wenger,

Characterisation of single particles from in-port ship emissions, Atmospheric Environment, 43 (40), 6408-6414, 2009.

E. J. Highwood and R. P. Kinnersley, When smoke gets in our eyes: The multiple

impacts of atmospheric black carbon on climate, air quality and health, Environment International, 32 (4), 560-566, 2006.

L. S. Hughes, J. O. Allen, M. J. Kleeman, R. J. Johnson, G. R. Cass, D. S. Gross, E. E.

Gard, M. E. Galli, B. D. Morrical, D. P. Fergenson, T. Dienes, C. A. Noble, P. J. Silva and K. A. Prather, Size and composition distribution of atmospheric particles in southern California, Environmental Science & Technology, 33 (20), 3506-3515, 1999.

M. Z. Jacobson, Strong radiative heating due to the mixing state of black carbon in

atmospheric aerosols, 409 (6821), 695-697, 2001. D. Koch, Transport and direct radiative forcing of carbonaceous and sulfate aerosols in

the GISS GCM, Journal Of Geophysical Research-Atmospheres, 106 (D17), 20311-20332, 2001.

D. Koch, T. C. Bond, D. Streets, N. Unger and G. R. van der Werf, Global impacts of

aerosols from particular source regions and sectors, Journal Of Geophysical Research-Atmospheres, 112 (D2), 2007.

B. Langmann, B. Duncan, C. Textor, J. Trentmann and G. R. van der Werf, Vegetation

fire emissions and their impact on air pollution and climate, Atmospheric Environment, 43 (1), 107-116, 2009.



D. Y. Liu, D. Rutherford, M. Kinsey and K. A. Prather, Real-time monitoring of

pyrotechnically derived aerosol particles in the troposphere, Analytical Chemistry, 69 (10), 1808-1814, 1997.

D. Y. Liu, R. J. Wenzel and K. A. Prather, Aerosol time-of-flight mass spectrometry

during the Atlanta Supersite Experiment: 1. Measurements, Journal Of Geophysical Research-Atmospheres, 108 (D7), 2003.

251

Y. C. Liu, M. A. Woodin, T. J. Smith, R. F. Herrick, P. L. Williams, R. Hauser and D. C. Christiani, Exposure to fuel-oil ash and welding emissions during the overhaul of an oil-fired boiler, Journal Of Occupational And Environmental Hygiene, 2 (9), 435-443, 2005.

U. Lohmann and C. Hoose, Sensitivity studies of different aerosol indirect effects in

mixed-phase clouds, Atmospheric Chemistry And Physics, 9 (22), 8917-8934, 2009.

G. McFiggans, P. Artaxo, U. Baltensperger, H. Coe, M. C. Facchini, G. Feingold, S.

Fuzzi, M. Gysel, A. Laaksonen, U. Lohmann, T. F. Mentel, D. M. Murphy, C. D. O'Dowd, J. R. Snider and E. Weingartner, The effect of physical and chemical aerosol properties on warm cloud droplet activation, Atmospheric Chemistry And Physics, 6, 2593-2649, 2006.

A. M. Middlebrook, D. M. Murphy, S. H. Lee, D. S. Thomson, K. A. Prather, R. J.

Wenzel, D. Y. Liu, D. J. Phares, K. P. Rhoads, A. S. Wexler, M. V. Johnston, J. L. Jimenez, J. T. Jayne, D. R. Worsnop, I. Yourshaw, J. H. Seinfeld and R. C. Flagan, A comparison of particle mass spectrometers during the 1999 Atlanta Supersite Project, Journal Of Geophysical Research-Atmospheres, 108 (D7), 2003.

R. C. Moffet, B. de Foy, L. T. Molina, M. J. Molina and K. A. Prather, Measurement of

ambient aerosols in northern Mexico City by single particle mass spectrometry, Atmospheric Chemistry And Physics, 8 (16), 4499-4516, 2008a.

R. C. Moffet, Y. Desyaterik, R. J. Hopkins, A. V. Tivanski, M. K. Gilles, Y. Wang, V.

Shutthanandan, L. T. Molina, R. G. Abraham, K. S. Johnson, V. Mugica, M. J. Molina, A. Laskin and K. A. Prather, Characterization of aerosols containing Zn, Pb, and Cl from an industrial region of Mexico City, Environmental Science & Technology, 42 (19), 7091-7097, 2008b.


properties of soot with implications for radiative forcing estimates, Proceedings Of The National Academy Of Sciences Of The United States Of America, 106 (29), 11872-11877, 2009.

D. M. Murphy, The design of single particle laser mass spectrometers, Mass

Spectrometry Reviews, 26 (2), 150-165, 2007. G. Myhre, Consistency Between Satellite-Derived and Modeled Estimates of the Direct

Aerosol Effect, Science, 325 (5937), 187-190, 2009.

252

K. R. Neubauer, M. V. Johnston and A. S. Wexler, Humidity effects on the mass spectra of single aerosol particles, Atmospheric Environment, 32 (14-15), 2521-2529, 1998.

C. A. Noble and K. A. Prather, Real-time single particle mass spectrometry: A historical

review of a quarter century of the chemical analysis of aerosols, Mass Spectrometry Reviews, 19 (4), 248-274, 2000.

S. H. Park, S. L. Gong, W. Gong, P. A. Makar, M. D. Moran, C. A. Stroud and J. Zhang,

Sensitivity of surface characteristics on the simulation of wind-blown-dust source in North America, Atmospheric Environment, 43 (19), 3122-3129, 2009.

S. H. Pastor, J. O. Allen, L. S. Hughes, P. Bhave, G. R. Cass and K. A. Prather, Ambient

single particle analysis in Riverside, California by aerosol time-of-flight mass spectrometry during the SCOS97-NARSTO, Atmospheric Environment, 37, S239-S258, 2003.


effects, Angewandte Chemie-International Edition, 44 (46), 7520-7540, 2005. K. A. Prather, C. D. Hatch and V. H. Grassian, Analysis of Atmospheric Aerosols,

Annual Review of Analytical Chemistry, 1, 485-514, 2008. K. A. Prather, T. Nordmeyer and K. Salt, Real-Time Characterization Of Individual

Aerosol-Particles Using Time-Of-Flight Mass-Spectrometry, Analytical Chemistry, 66 (9), 1403-1407, 1994.

K. A. Pratt, New Insights into Single-Particle Mixing State using Aircraft Aerosol Time-

of-Flight Mass Spectrometry, University of California, San Diego, La Jolla, CA, 2009.


Twohy, A. J. Prenni and K. A. Prather, In situ detection of biological particles in cloud ice-crystals, Nature Geoscience, 2 (6), 397-400, 2009a.

K. A. Pratt, L. E. Hatch and K. A. Prather, Seasonal volatility dependence of ambient

particle phase amines, Environmental Science & Technology, 43 (14), 5276-5281, 2009b.


Furutani, M. Gonin, K. Fuhrer, Y. X. Su, S. Guazzotti and K. A. Prather, Development and Characterization of an Aircraft Aerosol Time-of-Flight Mass Spectrometer, Analytical Chemistry, 81 (5), 1792-1800, 2009c.

253

K. A. Pratt and K. A. Prather, Real-Time, Single-Particle Volatility, Size, and Chemical Composition Measurements of Aged Urban Aerosols, Environmental Science & Technology, 43 (21), 8276-8282, 2009.

K. A. Pratt and K. A. Prather, Aircraft measurements of vertical profiles of aerosol

mixing states, Journal Of Geophysical Research-Atmospheres, 115 (D11305), 2010a.

K. A. Pratt and K. A. Prather, Mass spectrometry of atmospheric aerosols – Recent

developments & applications. Part II: On-line mass spectrometry techniques., Mass Spectrometry Reviews, Submitted, 2010b.

J. P. Putaud, F. Raes, R. Van Dingenen, E. Bruggemann, M. C. Facchini, S. Decesari, S.

Fuzzi, R. Gehrig, C. Huglin, P. Laj, G. Lorbeer, W. Maenhaut, N. Mihalopoulos, K. Mulller, X. Querol, S. Rodriguez, J. Schneider, G. Spindler, H. ten Brink, K. Torseth and A. Wiedensohler, European aerosol phenomenology-2: chemical characteristics of particulate matter at kerbside, urban, rural and background sites in Europe, Atmospheric Environment, 38 (16), 2579-2595, 2004.



X. Y. Qin and K. A. Prather, Impact of biomass emissions on particle chemistry during

the California Regional Particulate Air Quality Study, International Journal Of Mass Spectrometry, 258 (1-3), 142-150, 2006.

F. Raes, R. Van Dingenen, E. Vignati, J. Wilson, J. P. Putaud, J. H. Seinfeld and P.

Adams, Formation and cycling of aerosols in the global troposphere, Atmospheric Environment, 34 (25), 4215-4240, 2000.

V. Ramanathan and G. Carmichael, Global and regional climate changes due to black

carbon, Nature Geoscience, 1 (4), 221-227, 2008. T. P. Rebotier and K. A. Prather, Aerosol time-of-flight mass spectrometry data analysis:

A benchmark of clustering algorithms, Analytica Chimica Acta, 585 (1), 38-54, 2007.

J. S. Reid, E. J. Hyer, E. M. Prins, D. L. Westphal, J. L. Zhang, J. Wang, S. A.

Christopher, C. A. Curtis, C. C. Schmidt, D. P. Eleuterio, K. A. Richardson and J. P. Hoffman, Global Monitoring and Forecasting of Biomass-Burning Smoke: Description of and Lessons From the Fire Locating and Modeling of Burning Emissions (FLAMBE) Program, Ieee Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2 (3), 144-162, 2009.

254

N. Riemer, H. Vogel, B. Vogel and F. Fiedler, Modeling aerosols on the mesoscale-gamma: Treatment of soot aerosol and its radiative effects, Journal Of Geophysical Research-Atmospheres, 108 (D19), 2003.

N. Riemer, M. West, R. Zaveri and R. Easter, Estimating black carbon aging time-scales

with a particle-resolved aerosol model, Journal Of Aerosol Science, 41 (1), 143-158, 2010.

N. Riemer, M. West, R. A. Zaveri and R. C. Easter, Simulating the evolution of soot

mixing state with a particle resolved aerosol model, Journal Of Geophysical Research-Atmospheres, in press, 2009.

S. K. Satheesh and K. K. Moorthy, Radiative effects of natural aerosols: A review,

Atmospheric Environment, 39 (11), 2089-2110, 2005. R. B. Schlesinger, N. Kunzli, G. M. Hidy, T. Gotschi and M. Jerrett, The health relevance

of ambient particulate matter characteristics: Coherence of toxicological and epidemiological inferences, Inhalation Toxicology, 18 (2), 95-125, 2006.

J. H. Seinfeld and S. N. Pandis, Atmospheric Chemistry and Physics, Wiley Interscience,

Hoboken, NJ, 2006. L. G. Shields, D. T. Suess and K. A. Prather, Determination of single particle mass


P. J. Silva, R. A. Carlin and K. A. Prather, Single particle analysis of suspended soil dust

from Southern California, Atmospheric Environment, 34 (11), 1811-1820, 2000. P. J. Silva, D. Y. Liu, C. A. Noble and K. A. Prather, Size and chemical characterization


P. J. Silva and K. A. Prather, On-line characterization of individual particles from

automobile emissions, Environmental Science & Technology, 31 (11), 3074-3080, 1997.

P. J. Silva and K. A. Prather, Interpretation of mass spectra from organic compounds in

aerosol time-of-flight mass spectrometry, Analytical Chemistry, 72 (15), 3553-3562, 2000.

D. C. Snyder, J. J. Schauer, D. S. Gross and J. R. Turner, Estimating the contribution of

point sources to atmospheric metals using single-particle mass spectrometry, Atmospheric Environment, 43 (26), 4033-4042, 2009.

255



S. Solomon, D. Qin, M. Manning, Z. Chen, M. Marquis, K. B. Averyt, M. Tignor and H.





resolved chemical composition of aerosol particles during a monsoonal transition period over the Indian Ocean, 113 (D16), 2008.

M. T. Spencer and K. A. Prather, Using ATOFMS to determine OC/EC mass fractions in

particles, Aerosol Science And Technology, 40 (8), 585-594, 2006. M. T. Spencer, L. G. Shields, D. A. Sodeman, S. M. Toner and K. A. Prather,

Comparison of oil and fuel particle chemical signatures with particle emissions from heavy and light duty vehicles, Atmospheric Environment, 40 (27), 5224-5235, 2006.



Y. X. Su, M. F. Sipin, K. A. Prather, R. M. Gelein, A. Lunts and G. Oberdorster,

ATOFMS characterization of individual model aerosol particles used for exposure studies, Aerosol Science And Technology, 39 (5), 400-407, 2005.



R. C. Sullivan, M. J. K. Moore, M. D. Petters, S. M. Kreidenweis, G. C. Roberts and K.

A. Prather, Effect of chemical mixing state on the hygroscopicity and cloud nucleation properties of calcium mineral dust particles, Atmospheric Chemistry And Physics, 9 (10), 3303-3316, 2009a.

256

R. C. Sullivan, M. J. K. Moore, M. D. Petters, S. M. Kreidenweis, G. C. Roberts and K. A. Prather, Timescale for hygroscopic conversion of calcite mineral particles through heterogeneous reaction with nitric acid, Physical Chemistry Chemical Physics, 11 (36), 7826-7837, 2009b.

R. C. Sullivan and K. A. Prather, Recent advances in our understanding of atmospheric

chemistry and climate made possible by on-line aerosol analysis instrumentation, Analytical Chemistry, 77 (12), 3861-3885, 2005.

R. C. Sullivan and K. A. Prather, Investigations of the diurnal cycle and mixing state of

oxalic acid in individual particles in Asian aerosol outflow, Environmental Science & Technology, 41 (23), 8062-8069, 2007.

S. M. Toner, Anthropogenic Particulate Source Characterization and Source

Apportionment Using Aerosol Time-of-Flight Mass Spectrometry, Doctoral Thesis thesis, University of California San Diego, La Jolla, 2007.

S. M. Toner, L. G. Shields, D. A. Sodeman and K. A. Prather, Using mass spectral source

signatures to apportion exhaust particles from gasoline and diesel powered vehicles in a freeway study using UF-ATOFMS, Atmospheric Environment, 42 (3), 568-581, 2008.

S. M. Toner, D. A. Sodeman and K. A. Prather, Single particle characterization of

ultrafine and accumulation mode particles from heavy duty diesel vehicles using aerosol time-of-flight mass spectrometry, Environmental Science & Technology, 40 (12), 3912-3921, 2006.

S. P. Urbanski, J. M. Salmon, B. L. Nordgren and W. M. Hao, A MODIS direct broadcast

algorithm for mapping wildfire burned area in the western United States, Remote Sensing of Environment, 113 (11), 2511-2526, 2009.

S. Viswanathan, L. Eria, N. Diunugala, J. Johnson and C. McClean, An analysis of

effects of San Diego wildfire on ambient air quality, Journal Of The Air & Waste Management Association, 56 (1), 56-67, 2006.

R. J. Wenzel, D. Y. Liu, E. S. Edgerton and K. A. Prather, Aerosol time-of-flight mass

spectrometry during the Atlanta Supersite Experiment: 2. Scaling procedures, Journal Of Geophysical Research-Atmospheres, 108 (D7), 2003.

J. R. Whiteaker and K. A. Prather, Detection of pesticide residues on individual particles,

Analytical Chemistry, 75 (1), 49-56, 2003a. J. R. Whiteaker and K. A. Prather, Hydroxymethanesulfonate as a tracer for fog

processing of individual aerosol particles, Atmospheric Environment, 37 (8), 1033-1043, 2003b.

257

J. R. Whiteaker, D. T. Suess and K. A. Prather, Effects of meteorological conditions on

aerosol composition and mixing state in Bakersfield, CA, Environmental Science & Technology, 36 (11), 2345-2353, 2002.

J. Williams, M. de Reus, R. Krejci, H. Fischer and J. Strom, Application of the

variability-size relationship to atmospheric aerosol studies: estimating aerosol lifetimes and ages, Atmospheric Chemistry And Physics, 2, 133-145, 2002.

Q. Zhang, J. L. Jimenez, M. R. Canagaratna, J. D. Allan, H. Coe, I. Ulbrich, M. R.

Alfarra, A. Takami, A. M. Middlebrook, Y. L. Sun, K. Dzepina, E. Dunlea, K. Docherty, P. F. DeCarlo, D. Salcedo, T. Onasch, J. T. Jayne, T. Miyoshi, A. Shimono, S. Hatakeyama, N. Takegawa, Y. Kondo, J. Schneider, F. Drewnick, S. Borrmann, S. Weimer, K. Demerjian, P. Williams, K. Bower, R. Bahreini, L. Cottrell, R. J. Griffin, J. Rautiainen, J. Y. Sun, Y. M. Zhang and D. R. Worsnop, Ubiquity and dominance of oxygenated species in organic aerosols in anthropogenically-influenced Northern Hemisphere midlatitudes, Geophysical Research Letters, 34 (13), 2007.

R. Zimmermann, T. Ferge, M. Galli and R. Karlsson, Application of single-particle laser

desorption/ionization time-of-flight mass spectrometry for detection of polycyclic aromatic hydrocarbons from soot particles originating from an industrial combustion process, Rapid Communications In Mass Spectrometry, 17 (8), 851-859, 2003.

B. Zuberi, K. S. Johnson, G. K. Aleks, L. T. Molina and A. Laskin, Hydrophilic

properties of aged soot, Geophysical Research Letters, 32 (1), 2005.

258

8 Conclusions and Future Directions

8.1 Conclusions

Understanding the influence of transported particles on receptor sites is critical for

fully describing the atmospheric processes that impact climate and human health. This

dissertation has focused on observing the chemical and physical changes particles

undergo during transport on local (ship plumes), regional (port emissions and Asian

outflow), and intercontinental scales (Asian dust in North American precipitation). These

studies have shed new light on changes to transported particles over the timescales of

minutes to weeks. ATOFMS has been shown to be an effective tool for investigating

these research objectives. These findings have strengthened our understanding of the

processes that particles undergo during transport and provided motivation for more

detailed studies in the future.

Chapter 2 discussed results from analysis of particles observed at the Scripps

Institution of Oceanography (SIO) Pier in La Jolla, CA. Elemental carbon mixed with

organic carbon and vanadium-nickel-iron particle types were observed and linked with

regional transport from the region around the Ports of Los Angeles and Long Beach.

These two particle types were heavily aged and lacked negative spectra, likely due to

uptake of water, sulfate, and nitrate. The mass concentration of particles from the port

region was observed to dwarf that of local emissions at times of peak transport; size

resolved chemical analysis further pointed to a large mass contribution from the

submicron mode during these time periods. Mass concentration comparisons with a site

259

in downtown San Diego show that this transport affects the entirety of San Diego not just

the area surrounding the SIO Pier.

To verify that vanadium was coming from ships as hypothesized in Chapter 2,

sampling was conducted at a site adjacent to the main channel of the Port of Los Angeles.

Here plumes from specific ships were identified and analyzed with a suite of gas and

particle phase instrumentation. Plumes were grouped into two categories by their levels

of SORRRRRR2RRRRRR and NORRRRRRxRRRRRR, and ORRRRRR3RRRRRR: those with a high concentration of OC-V-sulfate particles,

which we believe originate from residual fuel burning ships, and those with high

concentrations of fresh soot, likely from distillate fuel burning ships. Both types were

observed to be highly correlated in time with specific plumes identified through ship logs,

radio signals, and visual documentation. A key finding was that sulfate and sulfuric acid

were enhanced preferentially on organic carbon particles mixed with vanadium. We

hypothesized that this is likely due to catalytic oxidation of the SORRRRRR2RRRRRR in the plume by

vanadium in the particle phase, either through aqueous or heterogeneous reactions.

The development of the mobile laboratory described in Chapter 4 is a significant

step forward in our ability to conduct field studies at diverse locations and transport

equipment between sites. Previous field deployment required that instruments be

transported in crates or box trucks and required housing and power upon arrival at a

sampling location. Using the mobile laboratory in combination with generators allows for

instruments to be towed to a site and be up and running in less than 15 minutes. This

allowed for measurements to be obtained at 5 sites in one day at different points around

San Diego Bay. The diffusion of particles from the source region in downtown versus the

cleaning effect of the sea breeze during the afternoon of March 13, 2009 provided an

260

interesting contrast between sources. This stood in stark contrast to the pristine conditions

observed during the earlier sampling on February 12, 2009 after a week of rain. These

results show the variation in particle chemistry and concentration as particles are

transported are transported away from and into a region that is often considered to have a

homogenous aerosol population.

The potential for Asian dust to influence precipitation in the United States,

specifically California, has not been fully elucidated, only gaining attention within the

past decade. Considering Asia accounts for roughly one third of global dust emissions,

these processes need to be explored in detail. The results in Chapter 5 detail initial

investigations into the links between Asian dust, clouds, and precipitation over

California. Monitoring the particle residues in precipitation revealed a stark shift from

organic carbon residues, likely from biomass burning, to dust residues over the course of

an extratropical cyclone with an atmospheric river. This contrasted with a separate

extratropical cyclone with an atmospheric river that did not observe this shift. Back

trajectory analysis revealed that during the influenced storm Asian dust was transported

at the altitudes of precipitating clouds. This was followed by a shift in the cloud drop size

distribution and an increase in precipitation intensity. While this is only one data point, it

is an intriguing observation that needs additional verification. These results demonstrate

the necessity of understanding the role of aerosol transport in altering aerosol-cloud-

precipitation interactions.

Since the outflow of particles from the Asian continent has the potential to

influence climate a continent away it is critical to understand the atmospheric processing

that these particles undergo during transport. Chapter 6 discusses results from the ground

261

measurements associated with the Pacific Dust Experiment (PACDEX) at the Gosan

Supersite on Jeju Island off the south coast of South Korea. The highest concentrations

observed at the site corresponded to air masses transported from China. These time

periods were compared with back trajectories coming from Korea, the Sea of Japan, and

over the ocean. Size resolved mass distributions indicate that even after one day of

transport, submicron particles comprise the majority of the mass in Asian outflow.

Additionally, these particle mixing state was heavily influenced by particle source and

size. The mixing state information in this paper should provide a useful data set for the

validation of single particle models investigating particle mixing state and aging.

Chapter 7 describes observations of single particle mixing state from studies across

the northern hemisphere. Based on comparisons of the size resolved chemical mixing

state four different site classifications emerged: aged urban, fresh urban, remote

continental, and marine. The size-resolved particle chemistry varied considerably

between the four types of sampling sites. A size dependence was observed for particle

chemistry, with most sites displaying a shift from a mostly carbonaceous submicron

aerosol to predominantly crustal and sea salt aerosol in the supermicron size range.

Comparisons of the secondary species associated with each of the primary particle types

demonstrated considerable differences across the diverse sampling locations.

Consistently observed trends included nitrate mixed with particles in the supermicron

size range and sulfate concentrated in the submicron mode. The similarities and

differences observed across these studies provide inputs for modelers attempting to

accurately represent single particle mixing state in climate models.

262

8.2 Los Angeles Basin Mobile Study (LABMS) Results

Prior to the mobile study around the San Diego Bay in 2009, a mobile laboratory

study was conducted in the eastern half of the Los Angeles (LA) Basin in 2007. These

results were not included in the manuscript described in Chapter 4 and are discussed

below.

The LABMS measurements were made over a 5 day period; the sampling periods

are shown as green bars in Figure 8.1. These measurements involved the mobile

laboratory equipped with an ATOFMS, a UF-ATOFMS, gas phase instrumentation (ORRRRRR3RRRRRR,

NORRRRRRxRRRRRR, and SORRRRRR2RRRRRR), and particle phase instrumentation (SMPS and APS). As described in

Chapter 4 power at each of the sites was obtained from two Honda generators inside the

bed of a pickup truck. Sites were chosen to obtain measurements on neighborhood scale

temporally and were distributed throughout the eastern half of the LA Basin. Sampling

sites included the University of California, Riverside on November 11PPPPPP

thPPPPPP and 12 PPPPPP

thPPPPPP (UCR1

and UCR2), the Chino dairy farms (Chino Farms), a park near downtown Riverside

(Riverside Park), a church parking lot in Chino Hills (Chino Hills Church), a park in

Rubidoux (Rutland Park), and a baseball field in Norco (Norco). The sampling conditions

during this 5 day period were characterized by low mass concentrations of particles less

than 2.5 microns (PMRRRRRR2.5RRRRRR), reaching their lowest levels for November 2007. The PMRRRRRR2.5RRRRRR

mass concentrations for two sites in the eastern Los Angeles Basin are also shown in

Figure 8.1.

During the fall, Riverside meteorology frequently leads to significant build up

periods followed by periods where winds come from the east bringing air masses with

few anthropogenic particles. These easterly winds, known as Santa Ana’s, are caused by

263

Figure 8.1 PMRRRRRR2.5RRRRRR mass concentrations during LABMS for Rubidoux (black line) and

Mira Loma (blue line). Sampling time periods are marked by green bars.

264

Figure 8.2 500 meter back trajectories from each of the sites sampled by the mobile laboratory during the LABMS: UCR1 (red), Chino Farms (orange), UCR2 (yellow), Riverside Park (green), Chino Hills (aqua), Rutland Park (blue), and Norco (dark blue). Diamond markers on the back trajectories represent each hour.

265

Figure 8.3 Size resolved chemical mixing state from both ATOFMS and UF-ATOFMS measurements at each of the mobile laboratory sites. a) UCR1, b) Chino Farms, c) UCR2), d) Riverside Park, e) Chino Hills, f) Rutland Park, g) Norco.

266

the clockwise flow around a high pressure system over the American northwest leading

to flow from the Great Basin of Nevada, Utah, and Eastern California towards Southern

California. The low concentrations discussed above were due to Santa Ana conditions

observed in the LA Basin which disrupted the typical diurnal shift between sea and land

breezes that often leads to pollution build up in the eastern half of the LA Basin. The

presence of Santa Ana winds was verified by back trajectory analysis at 500 meters for

each site during the study, which is shown in Figure 8.2; each of the sample sites can be

seen at the origin the back trajectory. These back trajectories indicate that air masses

approached the basin from the east, where as air masses typically approach from the west

during the mid-afternoon of days exhibiting typical diurnal patterns [Qin, et al., in prep].

The size resolved chemical mixing state (i.e. the fraction of each particle type in

each 50 nm bin) at each of the seven sites above is shown in Figure 8.3. The size range of

100-500 nm represents data from the UF-ATOFMS while the 500-2500 nm size range

was obtained from the ATOFMS. The main particle types observed are typical of

ATOFMS measurements in Riverside as reported by previous studies [Qin, et al., in

prep;Shields, et al., 2008]. These types include fresh elemental carbon, elemental carbon

mixed with organic carbon (ECOC), organic carbon (OC), biomass burning (Biomass),

vanadium-containing particles (vanadium), sea salt, and dust. For the purposes of these

comparisons, minor particle types were combined into the other category. The

measurements at the first sampling site UCR1 occurred as the Santa Ana winds were

cleaning out the built up pollution in the eastern LA basin. This can be identified

chemically by the highest fraction of organic carbon occurring during this time period

267

between 150 – 1000 nm. This time period also had the highest fractional contribution of

sea salt particles in the supermicron size range, which was likely due to diurnal transport

from the coast prior to the Santa Ana conditions. The size resolved chemical composition

for Chino Farms, UCR2, Riverside Park, Chino Hills are all reasonably similar and

characterized by ECOC, OC, and biomass in the submicron size range and dust in the

supermicron size range. Sampling at the Norco site occurred as the Santa Ana conditions

were lessening and the back trajectory after ~ 1 day loops from its initial easterly

direction towards the western LA basin and ocean. The influence of coastal emissions is

confirmed by a large fraction of vanadium containing particles in the submicron size

range that are likely from residual fuel burning ships in the Long Beach region [Ault, et

al., 2010;Ault, et al., 2009].

The results from the LABMS showed the impact of Santa Ana conditions across

the basin and the ubiquity of these conditions across the eastern LA basin. This study has

provided motivation for more detailed measurements with the mobile laboratory during

polluted conditions to determine the degree of spatial variability with respect to particle

chemistry over this region. Additional experiments with one stationary instrument and

one instrument (with the same inlet) in the mobile laboratory could provide insight into

the variation of aerosols over neighborhood to city-wide scales.

8.3 Characterization of Aerosol Optical Properties at Gosan, Korea

Additional studies beyond those described in Chapter 6 were conducted at the

Gosan supersite aiming to measure the optical properties of the sampled particles.

Particle optical properties were determined using the method previously developed in the

Prather laboratory [Moffet and Prather, 2005;Moffet, et al., 2008]. Briefly, the light

268

scattered by individual particles as they pass through the sizing lasers is focused by an

ellipsoidal mirror onto a photomultiplier tube (PMT). When particle size and chemical

composition are saved, the peak height and area from both lasers are also saved for the

particle. Once particles are grouped by chemical composition the raw scattering response

of each particle is plotted versus its size. Particles that travel through the center of the

Gaussian laser beam profile form an upper limit of scattered light intensity that can be

traced and fit to Mie Theory to determine particle refractive index (n) and density for the

class of particles. As the particle beam is broader than the laser beam a significant

fraction of particles will only clip the beam and scatter less light. To determine the upper

limit a monotonic regression of scattered intensities that accounts for the distribution of

full and clipped scattering responses was utilized. Preliminary results described here

indicate distinct optical differences between different particles types, as well as key

differences based on relative humidity. Examples of the scattering from several of the

most prevalent particle types (ECOC, biomass, dust, fresh sea salt, and aged sea salt) are

included in Figure 8.4. For spherical particle types, such as sea salt in a humid marine

environment, ripples are observed in the scattering data above roughly 1.5 microns.

These ripples are predicted from Mie theory as shown in the introduction (Chapter 1) and

confirm that the particles are spherical. The lack of a clear upper scattering limit for the

dust particle type is indicative of a non-spherical particle type, as is expected for dust.

Changes in relative humidity can have a significant impact on the scattering

properties of aerosols. As particles take up greater amounts of water their refractive index

approaches that of water (n = 1.33), which leads to a greater scattering response. The size

269

Figure 8.4 Single particle scattering data (response vs. size) from the most abundant particle types observed at Gosan, Korea: ECOC (red), biomass burning (yellow), fresh sea salt (aqua), aged sea salt (teal), dust (brown).

270

Figure 8.5 The scattering response as a function of size for biomass burning (top) and aged sea salt (bottom) with particles sampled during different relative humidities marked by color.

271

Figure 8.6 The top fit of the scattering response as a function of size for aged sea salt particles at different relative humidities.

272

resolved scattering data are shown in Figure 8.5 for biomass burning (top) and aged sea

salt (bottom) particles. For both particle types the upper limit of the scattering response

can be seen to increase as the relative humidity increases, indicating that the particles are

in fact taking up water. The differences in particle scattering response as a function of

relative humidity are more explicitly displayed for the aged sea salt particle type in

Figure 8.6. Here the upper limit fits can be seen to increase as relative humidity increases

(except for the 90-100% bin, which does not have sufficient counts for an accurate fit).

These results will provide a set of refractive indices and densities for aged air

masses as they begin the process of intercontinental transport and provide a more

fundamental understanding of the optical properties of aged aerosols. Taken together

these chemically resolved optical properties have the potential to greatly improve the

representation of aerosol properties in climate models by providing empirically derived

input values and benchmarks for comparison of current model results.

8.4 Future Directions

The research experiences described in this dissertation provide significant

motivation for follow up research by future group members. With respect to the analysis

of ship emissions, two directions of research could bring further advances beyond the

initial findings presented here. It is of particular interest to analyze fuel samples used in

ocean going vessels, specifically comparing the spectra associated with the combustion of

residual (a.k.a. bunker) fuels and distillated (a.k.a. diesel) fuels. Controlled sampling

from one or more ships by combining ATOFMS measurements with bulk sampling

methods [Agrawal, et al., 2008a;Agrawal, et al., 2008b], as has been previously achieved

for monitoring car and truck emissions [Shields, et al., 2007;Sodeman, et al., 2005;Toner,

273

et al., 2006] could provide verification of our hypothesis outlined in Chapter 3 that the

amount of fresh soot may be related to engine conditions. A second set of experiments

exploring the oxidation process of SORRRRRR2RRRRRR to sulfate could provide additional

understanding of the observations made in this work. These studies could be done in the

laboratory using a flow tube reactor, wherein fresh combustion particles would be

exposed to varying levels of SORRRRRR2RRRRRR to explore reaction kinetics and products. This study

could utilize techniques such as ion chromatography and microscopy, in addition to

ATOFMS, to gain further insight as to how the reaction progresses. Further, using the

ATOFMS to track an evolving ship plume in the ambient atmosphere (e.g. [Murphy, et

al., 2009]) would be interesting on two levels: (1) the sulfate and sulfuric acid peak areas

could be monitored during aging to obtain time scales of aging on the particles and (2)

the OC-V-sulfate particles in Chapter 2 near ship stacks likely aged and become the V-

Ni-Fe(sulfate-nitrate) particles observed both at the Port of Los Angeles and San Diego.

Monitoring the evolution of the mass spectra as these particles age could give insight as

to how the aging proceeds chemically. These laboratory and the field studies should also

characterize the cloud condensation nuclei (CCN) activity as a function of particle age.

Combining the chemistry and cloud nucleating potential of these particles could be key to

unraveling many of the lingering issues related to ship track clouds over the world’s

oceans.

The mobile laboratory represents a fantastic opportunity to explore both climate

and health related aerosol issues by taking advantage of its mobility. The addition of

universal power supplies (UPS) represents a minor modification to the mobile laboratory

that would increase its flexibility by decreasing down time between sites and potentially

274

allow for truly mobile sampling. At stationary sites additional UPS’s will serve to filter

incoming power, protecting against surges, dirty power, and short term power outages. In

time, developing a method for remote data collection and instrument monitoring, possibly

via a small satellite dish, would allow for greater real-time access to the instruments,

decrease instrument downtime by quickly alerting scientists to instrument malfunctions,

and provide the ability to backup data on the fly without relying on local internet (similar

to how NOAA’s sampling sites are operated). Additional studies that would be of

considerable interest involving the mobile laboratory include investigation of aging

processes placing a stationary ATOFMS at the source and positioning the mobile

laboratory at various distances downwind. This would build on initial attempts to

measure with one ATOFMS at a stationary location (UC-Riverside) and mobile

laboratory measurements made in the eastern Los Angeles Basin in fall 2007 that did not

yield usable results. An additional experiment that would be difficult, but potentially very

interesting would involve mobile sampling linked with aircraft sampling. For example,

measurements could be made each day on the ground at a different point along the flight

path of a research aircraft. The research aircraft would then transect the location at

various elevations to explore the link between the aerosol populations on the ground and

in vertical profile. This could provide a great deal of information regarding the exchange

of aerosols between the surface and aloft.

The initial measurements from CalWater Early Start described in Chapter 5 provide

the foundation for numerous research projects that could link precipitation sampling,

trajectory analysis, and meteorological measurements to explore aerosol-cloud-

precipitation interactions. In designing a future study utilizing this suite of measurements,

275

it would be extremely interesting to have an ATOFMS at one of the atmospheric river

observatories (AROs) on the west coast of California and a second ATOFMS at the Sugar

Pine site. Since the ARO sites are fairly remote it would be very interesting to compare

precipitation samples between these two sites to look at what typical residue chemistries

are and to observe the extent of dust incorporation at both sites under transport

conditions. A second, more focused approach, could involve sampling at a valley site

(perhaps Sloughhouse), at Sugar Pine, and at Norden simultaneously to explore changes

in the precipitation chemistry as the clouds are orographically lifted up the windward side

of the Sierra Nevada mountain range. The ultimate experiment might involve sampling at

all three of those sites while having a plane transect across the three at different altitudes

over the course of the day. This could provide unprecedented understanding of the

evolution of precipitation and cloud properties from a chemical and meteorological

perspective.

While the Gosan data set described in Chapters 6 provided a number of novel

results, there are many different directions that research of this type could be directed in

the future. A key starting place would be standardizing and better characterizing the

optical scattering measurements. At present there is not a good method for assessing data

quality at the time of collection. One way to improve the chances of success would

involve running a size calibration using more than the standard 8-10 sizes currently used.

A calibration of 25 sizes including 5 below a micron and PSLs every 0.1 microns from

1.0 microns to 3.0 microns would allow for a detailed understanding of the state of the

instrument as it compares to Mie theory. Currently the size calibration does not include

enough points above 1 micron to allow for a correction between Mie theory predictions

276

and actual data with a thorough calibration curve. Thus any differences in observed

scattering vs. theoretically predicted scattering could be corrected for ahead of time or

accounted for after the fact. A simple and interesting experiment with the optical

measurements could involve changing one of the scattering lasers to a blue or red

wavelength. Since Mie theory is wavelength dependent, variation in refractive index as a

function of wavelength could be observed for different particle types. The scattering

properties could then be integrated into an optical property model to constrain forcings

for specific particle types. Regarding sampling using tandem DMA-ATOFMS, a

complete characterization of the transmission efficiency for each aerosol inlet (nozzle or

aerodynamic lens) in high size resolution would be extremely useful for future

measurements, to allow for scaling to atmospheric concentrations after sampling. A

simple study with an ATOFMS run next to a DMA-ATOFMS setup would be extremely

interesting as it would allow for effective density and shape factor information to be

applied directly to measurements across the entire size range of the instrument.

The discussion of multiple ATOFMS datasets in Chapter 7 leaves open many

different possibilities for comparing datasets across instruments and research groups. An

obvious starting place would be to go into greater detail comparing the studies discussed

already, as well as including additional studies that were not included due to time and

space limitations. As new studies are collected I hope that efforts will be made to

compare with older data sets in order to provide interesting insights and greater

understanding of the temporal and spatial variability of atmospheric aerosols. An

additional step would involve comparing ATOFMS results with results from other single

particle mass spectrometers such as the PALMS [Murphy and Thomson, 1995], the

277

SPLAT instruments [Zelenyuk and Imre, 2005;Zelenyuk, et al., 2009], and the RSMS

[Carson, et al., 1995]. This will involve challenges such as different laser wavelengths,

single vs. dual polarity spectra, and differing inlets. The fact that these instruments have

operational differences that can give a broader set of results means that taken together

they have the potential to provide a more clear understanding of what we have learned in

nearly two decades of on-line single particle mass spectrometry analysis.

0B0B0B0B0B0B8.5 Final Thought

I hope the passion that I have had for this work over the course of my graduate

career is reflected in this dissertation. I have enjoyed the breadth of experiences from the

intensity and excitement of field studies to the integration of multiple techniques and data

sets after the fact to answer a pressing question. I hope that the scientific questions

explored in this thesis motivate others to ask the tough questions and not be afraid to

tackle new projects as a means of advancing our understanding of this gigantic reaction

vessel we call the atmosphere.

278

8.6 References

H. Agrawal, Q. G. J. Malloy, W. A. Welch, J. Wayne Miller and D. R. Cocker III, In-use

gaseous and particulate matter emissions from a modern ocean going container vessel, Atmospheric Environment, 42 (21), 5504, 2008a.


a Crude Oil Tanker at Sea, Environ. Sci. Technol., 2008b. A. P. Ault, C. J. Gaston, Y. Wang, G. Dominguez, M. H. Thiemens and K. A. Prather,

Characterization of the Single Particle Mixing State of Individual Ship Plume Events Measured at the Port of Los Angeles, Environmental Science & Technology, 44 (6), 1954-1961, 2010.



P. G. Carson, K. R. Neubauer, M. V. Johnston and A. S. Wexler, Online Chemical-

Analysis Of Aerosols By Rapid Single-Particle Mass-Spectrometry, Journal Of Aerosol Science, 26 (4), 535-545, 1995.

R. C. Moffet and K. A. Prather, Extending ATOFMS measurements to include refractive

index and density, Analytical Chemistry, 77 (20), 6535-6541, 2005. R. C. Moffet, X. Y. Qin, T. Rebotier, H. Furutani and K. A. Prather, Chemically


D. M. Murphy and D. S. Thomson, Laser Ionization Mass-Spectroscopy Of Single

Aerosol-Particles, Aerosol Science And Technology, 22 (3), 237-249, 1995. S. M. Murphy, H. Agrawal, A. Sorooshian, L. T. Padro, H. Gates, S. Hersey, W. A.

Welch, H. Jung, J. W. Miller, D. R. Cocker, A. Nenes, H. H. Jonsson, R. C. Flagan and J. H. Seinfeld, Comprehensive Simultaneous Shipboard and Airborne Characterization of Exhaust from a Modern Container Ship at Sea, Environmental Science & Technology, 43 (13), 4626-4640, 2009.



279

L. G. Shields, X. Y. Qin, S. M. Toner and K. A. Prather, Detection of ambient ultrafine

aerosols by single particle techniques during the SOAR 2005 campaign, Aerosol Science And Technology, 42 (8), 674-684, 2008.

L. G. Shields, D. T. Suess and K. A. Prather, Determination of single particle mass




S. M. Toner, D. A. Sodeman and K. A. Prather, Single particle characterization of

ultrafine and accumulation mode particles from heavy duty diesel vehicles using aerosol time-of-flight mass spectrometry, Environmental Science & Technology, 40 (12), 3912-3921, 2006.

A. Zelenyuk and D. Imre, Single particle laser ablation time-of-flight mass spectrometer:

An introduction to SPLAT, Aerosol Science And Technology, 39 (6), 554-568, 2005.

A. Zelenyuk, J. Yang, E. Choi and D. Imre, SPLAT II: An Aircraft Compatible, Ultra-

Sensitive, High Precision Instrument for In-Situ Characterization of the Size and Composition of Fine and Ultrafine Particles, Aerosol Science And Technology, 43 (5), 411-424, 2009.

UNIVERSITE OF CALIFORNIA, SAN DIEGO - eScholarship.org

Documents

Transcript of UNIVERSITE OF CALIFORNIA, SAN DIEGO - eScholarship.org