ABSTRACT 
 
 
 
 
Title of Document: EMISSIONS, TRANSPORT, AND 
EVOLUTION OF ATMOSPHERIC 
POLLUTANTS FROM CHINA: AN 
OBSERVATIONAL STUDY  
  
 Can Li, Doctor of Philosophy, 2008 
  
Directed By: Professor Russell R. Dickerson 
Professor Zhanqing Li 
Department of Atmospheric and Oceanic Science 
     
China?s air pollution issue, a byproduct of recent phenomenal economic 
growth, has received increasing attention in light of its local and large-scale impacts. I 
investigated the emissions, transport, and evolution of pollutants from China using 
measurements near some source regions in northern China in 2005.  Surface pollution 
near Beijing in March was overall heavy but changed dramatically, as passing mid-
latitude cyclones led to fast transitions between polluted prefrontal and clean 
postfrontal conditions. Large differences found between measurements and 
inventories suggest substantial uncertainties in emission estimates. Small, coal-fired 
boilers are shown unlikely to be the major source of inventory error; experiments 
measuring traffic emissions are called for. Ground-level aerosols absorb light and are 
from both wind-blown dust and anthropogenic emissions. Their effects on climate are 
to be further studied. 
  
Pollutants at higher altitudes are more likely to have large-scale impact than 
pollutants that remain near the surface. The aircraft campaign in April was among the 
first efforts to measure the vertical distribution of pollutants over inland China. The 
largest pollutant levels observed in the free troposphere during the campaign were 
related to dry convective lofting over an industrial region. This differs from earlier 
experiments over the Pacific, which recognized the warm conveyor belt (WCB) as the 
main lofting mechanism. Dry convection over the continent may be followed by 
WCB lifting as the systems move out over the ocean. Their relative roles are yet to be 
determined.   
 Analyses of meteorological and satellite cloud data reveal the importance of 
in-cloud processing in oxidizing SO2 transported behind cold fronts. Through 
integration of satellite sensors, in-situ measurements, trajectory and chemical 
transport models, I tracked a pollution plume as it traveled away from source region. 
The decay of SO2 in the plume over three days was quantified, suggesting an SO2 
lifetime of 1-4 d. Formation of sulfate and loss of dust together changed the aerosol 
loading of the plume. This analysis showcases the potential for employing satellites to 
trace transport events and pollutant evolution, and highlights the main uncertainties in 
quantitative application of satellite data.  
 
 
 
 
 
  
 
 
 
 
 
 
 
 
 
 
EMISSIONS, TRANSPORT, AND EVOLUTION OF ATMOSPHERIC 
POLLUTANTS FROM CHINA: AN OBSERVATIONAL STUDY    
 
 
 
By 
 
 
Can Li 
 
 
 
 
 
Dissertation submitted to the Faculty of the Graduate School of the  
University of Maryland, College Park, in partial fulfillment 
of the requirements for the degree of 
Doctor of Philosophy 
2008 
 
 
 
 
 
 
 
 
 
Advisory Committee: 
Professor Zhanqing Li, Chair/Advisor 
Professor Russell R. Dickerson, Advisor 
Professor Sheryl H. Ehrman 
Dr. Nickolay A. Krotkov 
Professor Ross J. Salawitch 
Dr. Jeffrey W. Stehr 
  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
? Copyright by 
Can Li 
2008 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ii 
 
Dedication 
To My Fellow Countrymen 
 
 iii 
 
Acknowledgements 
 First and foremost, I wish to thank my advisors, Professors Russ Dickerson 
and Zhanqing Li, for their guidance, patience, and support over the years. Working 
with them has been a truly ?eye-opening? experience for me. I enjoyed close 
collaboration with Drs. Lackson Marufu, Jeff Stehr, Brett Taubman, Jennifer Hains, 
and many others of Russ?s air chemistry group. I was also privileged to have some of 
the best colleagues and officemates in Wen Mi, Hye Lim Yoo, Feng Niu, Maureen 
Cribb, Drs. Tianle Yuan, Kwon-Ho Lee, and others in Zhanqing?s team. 
 My deepest gratitude goes to my colleagues at IAP/CAS and Liaoning 
Meteorological Administration, without whom this research would have been 
impossible. Dr. Tianxue Wen and Yanan Liu of IAP analyzed the aerosol 
composition. I cannot thank Drs. Nick Krotkov and Kai Yang of 
NASA/GSFC/UMBC enough for their help on OMI data and lots of insightful 
discussion. Dr. Mian Chin of NASA/GSFC kindly helped me with GOCART output. 
Drs. David Streets and Q. Zhang of ANL provided emission inventory data. Dr. 
Shimizu of NIES, Japan provided lidar data. 
 My thank goes out to all the faculty and staff of AOSC and ESSIC for their 
help. I am indebted to my fellow students and friends, especially Drs. Haifeng Qian, 
Junjie Liu, and Jie Liu, for five delightful years of my life. 
 Last but not least, I would like to thank my parents for supporting me and 
giving me courage to pursue my dream. I owe a great deal to my always caring, 
understanding, and loving wife, Chen. She has been my greatest inspiration and will 
always be.      
 
 iv 
 
Table of Contents 
 
 
Dedication..................................................................................................................... ii 
Acknowledgements......................................................................................................iii 
Table of Contents......................................................................................................... iv 
List of Tables ............................................................................................................... vi 
List of Figures............................................................................................................. vii 
Chapter 1: Introduction................................................................................................. 1 
1.1 Background......................................................................................................... 1 
1.2 Some Important Atmospheric Pollutants............................................................ 5 
1.3 Air Pollution in China......................................................................................... 9 
1.3.1 Industrialization, urbanization, and air pollution......................................... 9 
1.3.2 Air pollution in China: State of the issue................................................... 13 
1.3.3 Air pollution in China: impacts.................................................................. 15 
1.4 Overview of Research....................................................................................... 17 
1.4.1 Emission inventory: method, uncertainties, and evaluation ...................... 17 
1.4.2 Long-range transport of atmospheric pollutants: mechanisms .................. 19 
1.4.3 Aerosol chemical composition in China.................................................... 25 
1.4.4 EAST-AIRE: an international regional experiment................................... 26 
1.4.5 Outline of research..................................................................................... 27 
Chapter 2: Experimental Methods .............................................................................. 29 
2.1 Sampling Site and Platform .............................................................................. 29 
2.1.1 Xianghe site ............................................................................................... 29 
2.1.2 Y-12 aircraft............................................................................................... 30 
2.2 Measurements of Trace Gases .......................................................................... 32 
2.2.1 Carbon monoxide....................................................................................... 32 
2.2.2 Sulfur dioxide............................................................................................. 34 
2.2.3 Ozone ......................................................................................................... 35 
2.2.4 Total reactive nitrogen oxides.................................................................... 37 
2.3 Aerosol Optical Properties................................................................................ 38 
2.3.1 Aerosol scattering coefficients................................................................... 38 
2.3.2 Aerosol absorption coefficient................................................................... 40 
2.4 Aerosol Chemical Composition........................................................................ 42 
2.4.1 Aerosol sampling ....................................................................................... 42 
2.4.2 Chemical analysis ...................................................................................... 43 
Chapter 3: Xianghe Ground Experiment .................................................................... 45 
3.1 Abundance of Pollutants: Statistics and Variations.......................................... 45 
3.1.1 Concentrations of trace gases..................................................................... 45 
3.1.2 Aerosol optical properties and composition .............................................. 46 
3.1.3 Diurnal variation of pollutants................................................................... 49 
3.1.4 Synoptic variation of pollutants................................................................. 52 
3.2 Pollutant Ratios and Implications for Regional Emissions............................... 61 
3.2.1 Pollutant ratios from correlation analysis .................................................. 61 
3.2.2 Implications for CO and SO2 emissions .................................................... 65 
3.2.3 Cluster analysis and implications for regional emission inventory ........... 67 
 
 v 
 
3.2.4 Some other issues of inventory evaluation ................................................ 72 
3.2.5 Black carbon emissions.............................................................................. 74 
3.3 Enrichment of Elements and Ions in Aerosols.................................................. 76 
3.3.1 Enrichment factors and enrichment of lead ............................................... 76 
3.3.2 Enrichment of chloride ion ........................................................................ 80 
3.4 Characterize Emissions from a Small Coal-Fired Boiler.................................. 81 
3.4.1 The boiler................................................................................................... 81 
3.4.2 Method ....................................................................................................... 83 
3.4.3 Emission characteristics of the boiler and implications............................. 84 
3.5 Summary........................................................................................................... 89 
Chapter 4: Shenyang Aircraft Campaign.................................................................... 95 
4.1 Overview of the Aircraft Campaign ................................................................. 95 
4.1.1 Weather...................................................................................................... 95 
4.1.2 Flight routes ............................................................................................... 97 
4.1.3 Synoptic variation of air pollution during the aircraft campaign............... 99 
4.1.4 Vertical distribution of pollutants: comparison to the Northeastern U.S. 103 
4.2 Upwind Dry Convection and Long-Range Transport: A Case Study............. 107 
4.2.1 Prefrontal flight on April 5, 2005 ............................................................ 107 
4.2.2 Postfrontal flight on April 7, 2005........................................................... 115 
4.2.3 Discussion................................................................................................ 119 
4.3 Summary......................................................................................................... 121 
Chapter 5:  Evolution of Pollutants during Transport: A Perspective from Space... 125 
5.1 In-Cloud Processing of Upwind Sulfur Dioxide............................................. 125 
5.1.1 The relative nitrate-to-sulfate ratio in Xianghe aerosol samples ............. 125 
5.1.2 Synoptic change in [NO3-]/[SO42-]........................................................... 127 
5.1.3 SO42-/SOx and NO3-/NOy mole ratios ...................................................... 129 
5.1.4 In-cloud processing of SO2 ...................................................................... 134 
5.1.5 Implications for regional atmospheric chemistry .................................... 138 
5.2 Transport and Evolution of a Pollution Plume from Northern China: A 
Satellite-Based Case Study ................................................................................... 139 
5.2.1 Introduction.............................................................................................. 139 
5.2.2 Data and models....................................................................................... 140 
5.2.3 AMF correction for OMI operational SO2 product.................................. 144 
5.2.4 Transport and evolution of the plume...................................................... 150 
5.2.5 Changes in AOD and total mass of SO2 .................................................. 158 
5.2.6 Discussion................................................................................................ 160 
5.2.7 Summary.................................................................................................. 167 
Chapter 6:  Conclusions............................................................................................ 169 
6.1 Summary of Results........................................................................................ 169 
6.2 Recommendations........................................................................................... 175 
Appendix I ................................................................................................................ 179 
Appendix II ............................................................................................................... 185 
Appendix III.............................................................................................................. 189 
Bibliography ............................................................................................................. 190 
 
 
 vi 
 
List of Tables 
Table 1.1. National Ambient Air Quality Standards for some atmospheric 
pollutants....................................................................................................... 9 
Table 1.2. Concentration limits for some pollutants in the National Ambient Air 
Quality Standard of China .......................................................................... 14 
 
Table 3.1. Statistics of trace gases and aerosol optical properties in Xianghe... 47 
Table 3.2. Statistics of water-soluble inorganic ions in aerosol samples from 
Xianghe (?g/m3).......................................................................................... 48 
Table 3.3. Statistics of elements in aerosol samples from Xianghe (?g/m3)...... 49 
Table 3.4. Concentrations of some elements and water-soluble ions in daytime 
and nighttime samples (average ? standard deviation, ?g/m3) ................... 52 
Table 3.5. Results of the pollutant concentration correlation analysis............... 61 
Table 3.6. Observed and national inventory pollutant ratios (ppbv/ppbv)......... 65 
Table 3.7. Regional inventory pollutant ratios within different areas around 
Xianghe (ppb/ppb) ...................................................................................... 69 
Table 3.8. Results of correlation analysis for different trajectory groups.......... 69 
 
Table 4.1. Summary of weather conditions during the Shenyang aircraft 
campaign..................................................................................................... 97 
Table 4.2. Summary of the flights during the Shenyang aircraft campaign....... 98 
Table 4.3. Statistics of ?ngstr?m exponents (440-675 nm) in 2005 at three 
AERONET sites in China and the U.S. .................................................... 106 
 
Table 5.1. Average nitrate and sulfate concentrations from this study and some 
previous studies......................................................................................... 126 
 
 
 
 vii 
 
List of Figures 
Figure 1.1. Some examples of the impacts of atmospheric pollutants on climate, 
ecosystem, and humans................................................................................. 5 
Figure 1.2. GDP and percentage contribution from industry for China during 
1978-2004. .................................................................................................. 11 
Figure 1.3. Correlation between annual growth rate of GDP and industry of 
China during 1989-2004. ............................................................................ 11 
Figure 1.4. Coal production and percentage contribution to total energy 
consumption of China during 1978-2004. .................................................. 12 
Figure 1.5. Total SO2 emissions, industrial SO2 emissions, and industrial output 
value of China during 1995-2005. .............................................................. 12 
Figure 1.6. Schematic of major pathways for pollutant outflows from East Asia 
and inflows into North America ................................................................. 22 
Figure 1. 7. Major air streams associated with a mid-latitude cyclone.............. 23 
Figure 1. 8. Schematic of CO redistribution by a deep convection process. ..... 24 
 
Figure 2.1. Location of the Xianghe site............................................................ 30 
Figure 2.2. Y-12 aircraft used for the aircraft campaign. .................................. 32 
Figure 2.3. Schematic diagram of the modified CO analyzer............................ 34 
Figure 2.4. Schematic diagram of the modified SO2 analyzer........................... 35 
Figure 2.5. Schematic diagram of the TEI O3 analyzer ..................................... 36 
Figure 2.6. Schematic diagram of the modified NO/NOy detector.................... 38 
Figure 2.7. Schematic diagram of the TSI 3563 integrating nephelometer ....... 40 
Figure 2.8. Schematic diagram of the PSAP...................................................... 42 
Figure 2.9. Left: setup of the bulk aerosol sampler during the Xianghe 
experiment; Right: schematic diagram of the bulk aerosol sampler. ......... 43 
 
Figure 3.1. Diurnal cycles of CO, NOy, aerosol absorption SO2, aerosol 
scattering, and O3 at the Xianghe site. ........................................................ 51 
Figure 3.2. Hourly averaged aerosol scattering, aerosol absorption, relative 
humidity, temperature, surface pressure, wind direction, and wind speed 
observed at the Xianghe site. ...................................................................... 54 
Figure 3.3. Hourly averaged concentrations of SO2, CO, NOy, and O3 observed 
at the Xianghe site....................................................................................... 55 
Figure 3.4. Time series of some water-soluble ions in aerosol samples from 
Xianghe....................................................................................................... 58 
Figure 3.5. Time series of some pollutant tracer metal elements in aerosol 
samples from Xianghe. ............................................................................... 59 
Figure 3.6. Time series of some crustal elements in aerosol samples from 
Xianghe....................................................................................................... 60 
Figure 3.7. Correlations of CO and NOy, CO and SO2, aerosol scattering  and 
SO2, aerosol scattering  and aerosol absorption at the Xianghe site........... 62 
 
 viii 
 
Figure 3.8. Inventory ratios of CO/SO2 and CO/NOx  in March, 2001. ............ 68 
Figure 3.9. Correlation between BC and CO at the Xianghe site. ..................... 75 
Figure 3.10. Box and whisker chart of the enrichment factors.......................... 79 
Figure 3.11. Enrichment factor of Pb and hourly averaged CO. ....................... 79 
Figure 3.12. Scatter plot of Cl- and Na+ concentrations..................................... 81 
Figure 3.13. The coal-fired boiler at the Xianghe site. ...................................... 82 
Figure 3.14. One plume from the boiler detected on March 14, 2005............... 84 
Figure 3.15. Correlations of ?CO, ?SO2, ?NOy, and ?CO2 for all plume data.85 
Figure 3.16. Same as Figure 3.14 but for cold start data. .................................. 86 
 
Figure 4.1. Meteorology records from the airport near Shenyang..................... 96 
Figure 4.2. Flight altitudes and SO2 concentrations along the flight tracks..... 100 
Figure 4.3. Flight altitudes and aerosol scattering along the flight tracks. ...... 102 
Figure 4.4.Surface measured aerosol optical thickness over Shenyang. ......... 103 
Figure 4.5. a) Mean vertical profiles of SO2 during the Shenyang experiment. b) 
Mean vertical profiles of aerosol scattering.............................................. 105 
Figure 4.6. Vertical profiles of ?ngstr?m Exponents over N China in spring and 
over the NE U.S. in summer. .................................................................... 106 
Figure 4.7. Surface analysis for 00 UTC April 5, 2005................................... 109 
Figure 4.8. Vertical profiles of ozone, RH, and temperature S of Shenyang on 
April 5. ...................................................................................................... 110 
Figure 4.9. Vertical profile of aerosol scattering coefficients for flight S of 
Shenyang on April 5. ................................................................................ 110 
Figure 4.10.Vertical profile of SO2, for flight S of Shenyang on April 5........ 111 
Figure 4.11. Vertical distribution of CO from the flight of April 5................. 111 
Figure 4.12. Backward trajectories for the location of southern spiral............ 113 
Figure 4.13. GOES-9 IR image for 00 UTC April 4, 2005.............................. 114 
Figure 4.14. MODIS true color image for 04 UTC April, 4, 2005.................. 115 
Figure 4.15. Surface weather analysis for 00 UTC April 7, 2005. .................. 117 
Figure 4.16. Backward trajectories for the southern spiral on April 7............. 118 
Figure 4.17. Profiles of temperature, trace gases and aerosol scattering N of 
Shenyang on April 7. ................................................................................ 118 
Figure 4.18. Profiles of temperature, trace gases and aerosol scattering S of 
Shenyang on April 7. ................................................................................ 119 
Figure 4.19. Climatology of precipitation near Baltimore and Beijing ........... 120 
 
Figure 5.1. Nitrate and sulfate in bulk aerosol samples from Xianghe............ 129 
Figure 5.2. SO42-/SOx and NO3-/NOy mole ratios as a function of CO............ 130 
Figure 5.3. Scatter plots of SO42-/SOx and NO3-/NOy mole ratio .................... 132 
Figure 5.4. Upper: 72-hr backward trajectories from Xianghe at 04 UTC on 
March 17, 2005. Middle: MODIS-retrieved cloud properties along the 
trajectories. Lower: histograms of MODIS cloud properties................... 136 
Figure 5.5. Percentage of ?cloud?, ?dust?, ?cloud and dust?, and ?unexplained? 
samples...................................................................................................... 138 
 
 ix 
 
Figure 5.6. Histograms of MODIS cloud top pressure near the plume ........... 147 
Figure 5.7. Normalized trajectory-projected and GOCART SO2 vertical profile 
over N Japan on April 6, 2005.................................................................. 148 
Figure 5.8. Normalized trajectory-projected and GOCART SO2 vertical profile 
over N Pacific on April 7, 2005................................................................ 148 
Figure 5.9. Upper: aerosol extinction coefficients  in Sapporo, Japan on April 6, 
2005; Lower: vertical profile of aerosol extinction by spherical, non-
spherical, and total aerosols ...................................................................... 149 
Figure 5.10. Normalized trajectory-projected and GOCART aerosol vertical 
profile over N Pacific on April 7, 2005. ................................................... 150 
Figure 5.11. OMI-retrieved and trajectory-calculated SO2 from April 5 to 7. 151 
Figure 5.12. MODIS-retrieved AOD, OMI aerosol index, and trajectory 
projected AOD from April 5 to 7, 2005.................................................... 152 
Figure 5.13. GOCART simulated SO2, and AOD at 06 UTC, April 5, 2005.. 153 
Figure 5.14. Same as Figure 5.13 but for 03 UTC, April 6, 2005 ................... 156 
Figure 5.15. Same as Figure 5.13 but for 03 UTC, April 7, 2005 ................... 157 
Figure 5.16. Total SO2 mass of the plume and average AOD near the main body 
of the plume.. ............................................................................................ 160 
Figure 5.17. Trajectory-projected AOD on April 6. ........................................ 162 
Figure 5.18. Total SO2 mass of the plume, and MODIS and trajectory AOD. 163 
Figure 5.19. Decay plot of SO2 in the plume from April 5 to 7, 2005............. 165 
Figure 5.20. Change of AOD of aerosol species in GOCART model............. 166 
 
Appendix I.1. Surface weather analysis at 00 UTC on April 1, 2005. ............ 179 
Appendix I.2. Surface weather analysis at 00 UTC on April 2, 2005. ............ 179 
Appendix I.3. Surface weather analysis at 00 UTC on April 3, 2005. ............ 180 
Appendix I.4. Surface weather analysis at 00 UTC on April 4, 2005. ............ 180 
Appendix I.5. Surface weather analysis at 00 UTC on April 5, 2005. ............ 181 
Appendix I.6. Surface weather analysis at 00 UTC on April 6, 2005. ............ 181 
Appendix I.7. Surface weather analysis at 00 UTC on April 7, 2005. ............ 182 
Appendix I.8. Surface weather analysis at 00 UTC on April 8, 2005. ............ 182 
Appendix I.9. Surface weather analysis at 00 UTC on April 9, 2005. ............ 183 
Appendix I.10. Surface weather analysis at 00 UTC on April 10, 2005. ........ 183 
Appendix I.11. Surface weather analysis at 00 UTC on April 11, 2005. ........ 184 
Appendix I.12. Surface weather analysis at 00 UTC on April 12, 2005. ........ 184 
 
Appendix II.1. Vertical profiles of RH, aerosol scattering, and wet-bulb 
potential temperature north of Shenyang on April 5, 2005. ..................... 185 
Appendix II.2. Vertical profiles of RH, SO2, and wet-bulb potential temperature 
north of Shenyang on April 5, 2005. ........................................................ 185 
Appendix II.3. Vertical profiles of SO2, temperature, aerosol scattering, and 
relative humidity on April 10, 2005.......................................................... 186 
Appendix II.4. Vertical profiles of SO2, temperature, aerosol scattering, and 
relative humidity on April 11, 2005.......................................................... 187 
 
 x 
 
Appendix II.5. Vertical profiles of SO2, temperature, aerosol scattering, and 
relative humidity on April 12, 2005.......................................................... 188 
 
Appendix III.1. Polygon boxes used to sample pollution plume on April 6... 189 
Appendix III.2. Polygon boxes used to sample pollution plume on April 7... 189 
 
 1 
 
Chapter 1: Introduction 
 
1.1 Background 
Atmospheric pollutants refer to chemical or biological substances introduced 
into the atmosphere directly or indirectly by anthropogenic activities, existing in such 
amounts as to adversely impact the living environment and well-being of humans and 
other living organisms [Williamson, 1973]. Atmospheric pollutants can be classified 
into two categories: primary pollutants that are emitted directly; and secondary 
pollutants that are not emitted directly but formed in the atmosphere. Air pollution, a 
situation in which one or more atmospheric pollutants produce measurable adverse 
effects, has long been noted [Williamson, 1973]. Prior to the industrial era, scarce 
accounts in a handful of large population centers blamed coal or wood burning for 
foul odor and reduced visibility, generally the first noticed aspects of air pollution 
[Williamson, 1973]. Industrialization and urbanization that followed the Industrial 
Revolution brought irreversible changes to air quality. A series of deadly air pollution 
disasters from the 1880s to the 1950s called attention to the acute health effects of 
high concentrations of sulfur-containing pollutants. From the 1940s, expansion of the 
motor vehicle fleet led to pollution characteristic of high ozone levels in some cities 
with bright sunshine [Seinfeld and Pandis, 1998]. Later epidemiological studies 
revealed more subtle health effects of aerosols, or small condensed particles 
suspended in the air, at their common ambient abundance [e.g., Pope III et al., 1995]. 
The late 1980s saw chlorofluorocarbons (CFCs), a class of man-made substances 
 
 2 
 
previously believed harmless, undergo regulation amid worries about their role in 
depleting stratospheric ozone, which protects life from harmful solar ultraviolet 
radiation [UNEP, 2000]. More recently, carbon dioxide (CO2), safe to breathe at 
ambient concentrations, has raised concerns regarding its greenhouse effect [IPCC, 
2007]. 
Two major types of air pollution have been recognized. The first type or 
?London smog? is characterized by high concentrations of sulfur compounds, usually 
found in metropolitan areas with intense coal burning when cold and stagnant weather 
suppresses dilution of pollutants [Seinfeld and Pandis, 1998]. The second type or 
?photochemical smog? occurs in cities with heavy automobile traffic as hot and sunny 
weather favors the generation of ozone and other secondary pollutants [Seinfeld and 
Pandis, 1998]. Traditionally air pollution was viewed as a local problem for 
metropolitan or industrialized regions where strong emissions greatly exceed the 
capacity of natural cleansing mechanisms such as dry and wet deposition, chemical 
reaction, and ventilation. But it is now accepted that air pollution is not just limited to 
these ?hot spots?. The motion of the atmosphere conveys pollutants far away from 
their emission sources, in some cases all over the globe. Photochemical smog may 
span over a few hundred kilometers, with peak ozone found in suburban areas [e.g., T. 
Wang et al., 2006]. Sulfur and nitrogen compounds convert to acidic species during 
transport, and can harm forests a few thousand kilometers downwind. There has been 
accumulating observational evidence on the intercontinental transport of pollutants 
across the Atlantic and the Pacific. The Antarctic ?ozone hole? in association with 
 
 3 
 
CFCs is a perfect example that trace amounts of long-lived atmospheric pollutants 
can cause global environmental problems [Seinfeld and Pandis, 1998].  
An imminent global issue is the change of climate caused by anthropogenic 
activities, particularly emissions of greenhouse gases. With CO2 increasing from 
about 280 ppm (parts per million) pre-industrial to 379 ppm in 2005, the global 
surface temperature rose by 0.76?C from 1850-1899 to 2001-2005, and is projected to 
go up by another 0.4?C in the next two decades [IPCC, 2007]. Some aspects of air 
pollution will change or have already changed in this global warming scenario. A 
warmer climate and the associated change in precipitation can affect ecosystems and 
the emissions of biogenic hydrocarbon compounds [e.g., Steiner et al., 2006], 
important precursors in the photochemical production of ozone. At the regional scale, 
the weakening of monsoon systems or mid-latitude cyclones in some regions can 
lower surface wind speed and build up pollutants near their emission sources [e.g., 
Mickley et al., 2004]. Forest fires may increase in some regions [Westerling et al., 
2006]; so would emissions from the fires. A warmer climate can also affect 
temperature-dependent chemical processes of the atmosphere. Air pollution, in turn, 
by altering the composition of the atmosphere may also impact the climate system. 
Some atmospheric pollutants are themselves direct or indirect radiative forcing 
agents. CFCs and ozone are important greenhouse gases [IPCC, 2007]. Aerosols, 
through scattering and absorbing light [e.g., Charlson et al., 1992] and interacting 
with clouds [Twomey, 1959], can change the Earth?s radiation budget, and remain the 
one of the largest sources of uncertainties in climate studies [IPCC, 2007]. 
Atmospheric pollutants influence ecosystems and the biosphere-atmosphere exchange 
 
 4 
 
of CO2. It has been proposed that nitrogen deposited to the northern hemispheric mid-
latitude forests may have turned them into a net carbon ?sink?, assimilating CO2 from 
the atmosphere [Reay et al., 2008]. Cloud-aerosol interaction may affect precipitation 
[e.g., Rosenfeld et al., 2008], another factor regulating the land-atmosphere carbon 
flux. Export of nutrients to the oceans can possibly impact maritime ecosystems as 
well [Reay et al., 2008]. High ozone levels near ground on the other hand can damage 
plants [Chameides et al., 1994]. Air pollution may also enhance the oxidizing 
capacity of the atmosphere, affecting the budget of some reactive greenhouse gases 
(e.g., methane, [Karlsd?ttir and Isaksen, 2000]). Listed above and illustrated in 
Figure 1.1 are some examples of the possible connections between air pollution and a 
changing climate. In addition, human activities in response to climate change or air 
pollution can also affect air quality. For example, shifting energy use from fossil fuels 
to renewable energy (e.g., hydro power and wind farms) can cut the emissions of CO2 
and NOx at the same time. 
Anthropogenic emissions are closely related to the economic and social status 
of a region; so is air pollution, which could be distinctly different between developed 
and developing areas. Therefore understanding air pollution under global warming 
circumstances requires examining the emissions, distribution, and evolution of 
atmospheric pollutants on a region-by-region basis. To study air pollution in any 
given region, observations from close range are highly desired. It is based on this 
notion that the studies presented in this dissertation have been formulated. This 
observational research focuses on China, an emerging economic power and the most 
populous country in the world, and delivers a data set that lends some insights into 
 
 5 
 
China?s air pollution and its impact on the composition of the atmosphere on local 
and global scales. 
 
Figure 1.1. Some examples of the impacts of atmospheric pollutants on climate, 
ecosystems, and humans. 
 
1.2 Some Important Atmospheric Pollutants 
In this section, I briefly review the sources, sinks, and residence times (or 
lifetimes) of some atmospheric pollutants discussed in more detail in this dissertation. 
Sources of atmospheric pollutants refer to locations, facilities, or mechanisms 
releasing pollutants to the atmosphere or producing secondary pollutants. Sinks of 
atmospheric pollutants are processes that remove pollutants from the atmosphere (or a 
portion of the atmosphere) or convert pollutants to other species [Seinfeld and Pandis, 
 
 6 
 
1998]. Regarding the entire atmosphere as a reservoir for a certain pollutant with total 
abundance Q, its rate of change is determined by the rate the pollutant is introduced 
into (P) and removed from (R) the atmosphere, and can be expressed as  
RPdtdQ ?=   (1.1) 
Assuming a steady state under which P = R, the average residence time of the 
pollutant, ?, is 
P
Q
R
Q ==?   (1.2) 
The residence time can be used to estimate how far certain pollutants may travel 
before being removed from the atmosphere. Long-lived species (e.g., CO2) are 
generally well-mixed around the globe. Short-lived species are unevenly distributed, 
and are usually more abundant where strong emissions exist. 
Carbon monoxide (CO) is a poisonous trace gas [Seiler and Junge, 1970] 
limited by a National Ambient Air Quality Standard (NAAQS). CO participates in 
photochemical processes that form ozone in the air. Three major sources of CO are 
oxidation of methane (CH4), biomass burning, and combustion of fossil and domestic 
fuels, each contributing about 650-800 terragrams (1012 g) CO per year (Tg CO/yr) 
[IPCC, 2001]. Reaction with hydroxyl radicals (OH) is the most important sink of 
CO, consuming 1500-2700 Tg CO/yr. The chemical lifetime of tropospheric CO is 
around 30-90 days [Seinfeld and Pandis, 1998]. 
Nitrogen Oxides (NOx = NO + NO2) play a central role in the photochemical 
processes of the atmosphere. They are mainly from high-temperature combustion of 
fossil fuels (33 Tg N/yr). Soil emissions, lightning, and biomass burning add 
 
 7 
 
considerable but less NOx to the atmosphere (in total about 19 Tg N/yr [IPCC, 
2001]). The major sink of NOx is reaction between NO2 and OH to form nitric acid 
(HNO3), which ultimately is scavenged from the atmosphere. Compared to CO, NOx 
has a shorter lifetime of hours to few days [Seinfeld and Pandis, 1998]. The sum of 
NOx and its atmospheric oxidation products is total reactive nitrogen oxides (NOy), 
including NOx, HNO3, nitrous acid (HONO), the nitrate radical (NO3), dinitrogen 
pentoxide (N2O5), peroxynitric acid (HNO4), peroxyacetyl nitrate (PAN), alkyl 
nitrates (RONO2), and peroxyalkyl nitrates (ROONO2). 
Sulfur dioxide (SO2) arises predominantly from anthropogenic sources (~70 
Tg S/yr, [Seinfeld and Pandis, 1998; Chin et al., 2000] including combustion of 
sulfur-containing fossil fuels and industrial processes, although occasional large 
volcanic eruptions inject large amount of SO2 into the atmosphere in a short period of 
time. Dry deposition accounts for almost half of the total SO2 sink [Chin et al., 2000]. 
In-air oxidation of SO2 (by OH, R1.1) and in-cloud reaction (by hydrogen peroxide, 
H2O2, R1.2) are other important SO2 sinks; the latter mechanism is faster [Chin et al., 
1996]. Sulfate particles generated from SO2 chemical conversion scatter radiation and 
can be activated as cloud condensation nuclei (CCN). Estimates of SO2 lifetime vary, 
from around one day [Chin et al., 2000] to a few days [Koch et al., 1999]. 
K+?+ 422 )( SOHgOHSO   (R1.1) 
K+?+ 42222 )( SOHaqOHSO   (R1.2) 
Ozone (O3) is a product of complex photochemical reactions involving two 
precursors: NOx and volatile organic compounds; the two determine ozone 
concentration in a nonlinear way. Exchange between troposphere and stratosphere is 
 
 8 
 
another source of tropospheric ozone. Ozone in the stratosphere is ?good?, in terms of 
its absorption of harmful UV radiation. On the contrary, ozone near ground is ?bad?, 
with adverse effects on the health of both humans and plants. Ozone near the 
tropopause acts as a greenhouse gas [IPCC, 2007]. Photolysis of ozone generates OH, 
an important player in atmospheric chemistry (R1.3). 
OHOHDO
ODOhO
2)(
)(
2
1
2
1
3
?+
+?+ ?   (R1.3) 
Aerosols or particulate matter are known for their complicated composition 
and origins. Directly emitted or formed in the atmosphere, these condensed particles 
suspended in the air can come from both natural (e.g., sea spray, volcanoes) and 
anthropogenic sources (e.g., combustion). Dust, if mobilized from areas where human 
activities have caused desertification, may also be considered ?anthropogenic?. Of 
great concern in recent years is the aerosol radiative forcing (ARF), the biggest 
uncertainty in climate change studies. Some proposed ARF mechanisms are: 
absorption and scattering of radiation (direct forcing), change of cloud reflectivity, 
alteration of cloud fraction and lifetime (indirect forcing), change in atmospheric 
stability due to aerosol heating and cloud formation (semi-direct forcing). Another 
issue drawing much attention is the health effects of aerosols. Aerosol size ranges 
from a few nanometers (nm) to up to few hundred micrometers (?m). Smaller 
aerosols with aerodynamic diameter below 2.5 ?m (PM2.5), also referred to as fine 
particles, could be of great harm, as their size allows for deeper penetration into the 
respiratory system. Coarse mode aerosols (> 2.5 ?m in aerodynamic diameter) 
usually consist of dust, sea salt, and plant particles, and have a lifetime shorter than 
that of fine particles (days to about a week) due to gravitational sedimentation 
 
 9 
 
[Seinfeld and Pandis, 1998]. Studies also demonstrate the association between 
aerosols and reduced visibility [e.g., Dzubay et al., 1982]. Table 1.1 lists the U.S. 
NAAQS standards for some pollutants. 
 
Table 1.1. U.S. National Ambient Air Quality Standards for some atmospheric 
pollutants 
Source: EPA website: http://www.epa.gov/air/criteria.html; Primary standards protect 
people from health risk; Secondary standards prevents unacceptable effects on the 
public welfare, for example unacceptable damage to buildings and crops. 
a Parts per million. 
b Particulate matter with aerodynamic diameter below 10 ?m. 
c Particulate matter with aerodynamic diameter below 2.5 ?m. 
 
1.3 Air Pollution in China 
1.3.1 Industrialization, urbanization, and air pollution 
With her economy soaring at unprecedented pace, products swarming into 
global market, and cities expanding faster than ever, China has become arguably the 
second largest economic entity in the world, as measured by purchasing power parity 
 Primary standards Secondary standards 
Pollutant Level Averaging time Level Averaging time 
9 ppma 8-hour None CO 
35 ppm 1-hour Same as Primary 
NO2 0.053 ppm Annual Same as Primary 
PM10b 150 ?g/m3 24-hour Same as Primary 
15.0 ?g/m3 Annual Same as Primary PM
2.5
c 
35 ?g/m3 24-hour Same as Primary 
0.075 ppm 8-hour Same as Primary O
3 0.12 ppm 1-hour Same as Primary 
0.03 ppm Annual SO
2 
0.14 ppm 24-hour 
0.5 ppm 3-hour 
 
 10 
 
(PPP) [World Bank, 2006]. However, such dazzling achievement does not come 
without a cost. Among many challenges for this country is the deteriorating 
atmospheric environment, a byproduct of fast industrialization and urbanization. 
Contributing over 40% of the gross domestic product (GDP, Figure 1.2), 
industry is the locomotive driving economic growth in China (Figure 1.3), but also 
consumes about 70% of China?s energy, mostly from coal (Figure. 1.4). Annual coal 
production in China has increased almost by a factor of three since 1978 (Figure 1.4, 
[Chinese CBS, 2005]), accounting for more than 60% of total energy consumption. 
SO2 emissions, mostly from coal-burning, have recovered from a decline in mid-
1990s [Larssen et al., 2006] and been on the rise over the past several years (Figure 
1.5). NOx emissions from China have also increased [Richter et al., 2005], as power 
generation grows [Chinese CBS, 2005]. CO emissions, largely from the residential 
sector, were projected to be relatively steady, as more efficient combustion 
techniques cancel out growing energy demand [Streets and Waldhoff, 2000], although 
emissions from steel mills may continue to go up. 
 
 11 
 
 
Figure 1.2. GDP (line and dots) and percentage contribution from industry (shaded 
columns) for China during 1978-2004 (source: [Chinese CBS, 2005]).  
 
 
Figure 1.3. Correlation between annual growth rate of GDP and industry of China 
during 1989-2004 (source: [Chinese CBS, 2005]). 
 
 12 
 
 
Figure 1.4. Coal production (line and dots) and percentage contribution to total 
energy consumption (shaded columns) of China during 1978-2004 (source: [Chinese 
CBS, 2005]).  
 
Figure 1.5. Reported total SO2 emissions (line and squares), industrial SO2 emissions 
(line and triangles), and industrial output value (line and stars) of China during 1995-
2005 (source: [Chinese CBS, 2005]). 
 
 13 
 
Air pollution, previously considered a local problem for industrialized big 
cities in China [Tang et al., 1990], has developed into a regional problem [e.g., He et 
al., 2002; Shao et al., 2006]. From 1985 to 2002 the urban population in China 
doubled, and the urbanized area nearly tripled [Guo, 2004, document available online 
at http://www.drcnet.com.cn/]. Fast urbanization processes turn agricultural land into 
city clusters and mega cities [Shao et al., 2006]. Industry in rural areas is booming: in 
2004, one in every five of China?s 650 million person labor force worked for factories 
built in the countryside [Chinese CBS, 2005]. These numerous and mostly small 
factories are substantial sources of atmospheric pollutants [Streets et al., 2003]. In 
addition to industrialization and urbanization, other factors such as the 20-fold growth 
of the motor vehicle fleet since 1978 [Chinese CBS, 2005] may also have exacerbated 
the air quality issue in China. 
 
1.3.2 Air pollution in China: State of the issue 
In 2002, only about one third of the 343 monitored cities met the Class II 
National Ambient Air Quality Standard of China (Table 1.2) [Hao and Wang, 2005]. 
In cold winter, indoor heating and stagnant meteorological conditions build up heavy 
pollution in cities across northern China [e.g., He et al., 1999; Liu and Song, 2001; 
Ma et al., 2003; Wu et al., 2002]. In spring, dust storms pick up in the arid northern 
China; some particularly strong ones can have widespread effects [e.g., Sun et al., 
2005]. Higher temperature and stronger sunshine in summer are conducive to the 
formation of ozone and secondary aerosols, causing photochemical smog and haze in 
big cities [e.g., Bergin et al., 2001; Liu and Shao, 2004]. Ozone violations in the cities 
 
 14 
 
in the warmer south may occur from late spring until early fall [e.g., X. Wang et al., 
2003]. Acid deposition has become a serious problem in southern China, where the 
buffering capacity of soil is weak [e.g., Wang and Wang, 1995, 1996].  
Experiments at a number of rural or ?background? sites demonstrate the 
extent of the air pollution issue in China. Almost 700 ppb (parts per billion) CO was 
measured at a rural site in the Yangtze River Delta (YRD) region in spring [T. Wang 
et al., 2004], far higher than the normal range seen in the Northern Hemisphere 
[Seinfeld and Pandis, 1998]. At a mountainous site 50 km north of downtown 
Beijing, T. Wang et al. [2006] recorded maximum O3 of 286 ppb in the summer of 
2005. PM2.5 monitoring in winter 2002 shows comparable aerosol loadings at eight 
urban and suburban sites in the Pearl River Delta region [Cao et al., 2003]. Similarly, 
great PM2.5 levels were found both downtown and in the outskirts of Beijing [Zheng 
et al., 2005]. 
 
Table 1.2. Concentration limits for some pollutants in the National Ambient Air 
Quality Standard of China (adapted from [He et al., 2002]). (Unit: mg/m3) 
Pollutants Averaging 
Time 
Class Ia Class IIb Class IIIc 
Daily 0.12 0.3 0.5 TSP  
Annual 0.08 0.2 0.3 
Daily 0.05 0.15 (0.15) 0.25 PM10  
Annual 0.04 0.1 0.15 
Daily 0.05 0.15 (0.37 d) 0.25 SO2  
Annual 0.02 0.06 (0.08 d) 0.1 
Daily 0.08 0.08 0.12 NO2  
Annual 0.04 0.04 (0.1) 0.08 
Daily 0.1 0.1 0.15 NOx  
Annual 0.05 0.05 0.1 
CO  Daily 4 4 (10 e) 6 
O3  Hourly 0.12 0.16 (0.24f) 0.2 
Numbers in parenthesis are the U.S. EPA NAAQS values. 
a Set for natural conservation, resort, or tourist areas. 
 
 15 
 
b Set for urban residential, commercial, or rural areas. 
c Set for industrial districts or traffic centers.  
d Unit converted assuming 20?C, 101.325 kPa (NIST standard temperature and 
pressure). 
e 8-hour concentration for the U.S. standard. 
f Unit converted assuming 20?C, 101.325 kPa, hourly concentration standard 
only applied in limited areas. 
 
1.3.3 Air pollution in China: impacts 
Elevated pollutant levels lead to various problems. Ozone in rural China, 
projected to rise further in the future [Chameides et al., 1999a], may reduce the yield 
of some crops by almost 20% [Aunan et al., 2000]. High aerosol loading in eastern 
China (aerosol optical thickness, AOT: ~0.5, [Xin et al., 2007]) can diminish solar 
radiation reaching the surface, and may cost China ~10% of total crop production 
[Chameides et al., 1999b]. Acid deposition can affect ecosystems and contaminate 
water bodies [e.g., Yang et al., 2001]. As discussed earlier, high ozone and aerosol 
levels also adversely impact human health, causing increased morbidity [Wang and 
Mauzerall, 2006], and over 300,000 estimated excess deaths every year in China 
[WHO and UNDP, 2001]. Strongly absorbing aerosols (particles that absorb solar 
radiation) may influence atmospheric heating rates [Z. Li et al., 2007a] and affect the 
evolution of the planetary boundary layer (PBL) [Yu et al., 2002]. Some studies 
attribute the recently documented weakening of the East Asian summer monsoon [Xu 
et al., 2006], reduced cloudiness [Qian et al., 2006] and precipitation [Zhao et al., 
2006], and ?north drought south flood? climate anomaly [Menon et al., 2002; Xu, 
2001] over China to aerosol forcing. Changes in precipitation may further affect 
ecosystems and agriculture. In 2006, the first ever ?Green GDP? report issued by 
 
 16 
 
Chinese EPA and Chinese CBS [2006] estimated that the economic loss due to air 
pollution in China was more than 25 billion U.S. dollars in 2004, or 1.2% of GDP. 
The World Bank [1997] suggested that air pollution cost a much higher fraction of 
China?s GDP (7%).  
Once vented into the free troposphere and driven by strong westerly winds 
aloft, pollutants can potentially travel over long distances [e.g., Berntsen et al., 1999; 
Chatfield and Crutzen, 1984; Dickerson et al., 1987, Jaffe et al., 1999, 2003a; Husar 
et al., 2001; Prospero et al., 2003; Allen et al., 2004; Mari et al., 2004; Guttikunda et 
al., 2005], turning a local air pollution issue into a large scale problem. While the 
average intercontinental transport time was estimated to be 2-4 weeks [Liu and 
Mauzerall, 2005]; in some episodic transpacific events, pollution plumes from East 
Asia could reach North America in just days [e.g., Husar et al., 2001; Jaffe et al., 
1999; Stohl et al., 2003]. Pollutants from East Asia could increase surface ozone [e.g., 
Fiore et al., 2002; Hudman et al., 2004; Jaffe et al., 2003b; Lin et al., 2000] and 
particle concentrations, and reduce visibility [Park et al., 2004; Fairlie et al., 2007] in 
the U.S. Transboundary transport of pollutants from China into Korea and Japan also 
has been documented in both observational and numerical studies [e.g., Kim et al., 
2001; Arimoto et al., 2006]. Aerosols (or precursors) from China might impact the 
climate over large areas, such as the cloud system in the northern Pacific storm track 
[R. Zhang et al., 2007]. 
 
 
 17 
 
1.4 Overview of Research 
1.4.1 Emission inventory: method, uncertainties, and evaluation 
Usually constructed with ?bottom-up? methods, emission inventories estimate 
the amount of certain pollutants released into the atmosphere from a certain area 
during a certain period of time. Emissions of a specific pollutant j are calculated as: 
??=
m n
nmlkjnmlkjlkj EFAE ,,,,,,,,,,    (1.3) 
where E represents emissions, A is the activity rate (e.g., fuel consumption), and EF 
is emission factor after abatement (e.g., pollutant release from unit mass of fuel). k 
represents an area; l stands for a sector; m is fuel or activity, and n represents 
emission abatement [X. Wang et al., 2005]. Total emissions of the species j from the 
whole inventory region would be: 
??=
k l
lkjj EE ,,   (1.4) 
In practice, anthropogenic and natural contributions are often treated 
separately [e.g., Streets et al., 2003], with the former classified into three categories: 
large point sources (LPS) with strong emissions from a small fixed area (e.g., stacks 
of power plants), area sources including all stationary sources not identified as LPS, 
and mobile sources mainly consisting of on-road vehicles [X. Wang et al., 2005]. 
Activity rate (or fuel consumption) data are normally from government statistics, and 
emission factors are from in-lab or field experiments. The quality of the A and EF 
largely determines the reliability of an emission inventory. 
For a large developing country like China, collecting information to compile a 
?bottom-up? inventory takes tremendous effort. Bias and inconsistency in 
government statistics, as pointed out by Sinton [2001], Streets et al. [2003], and X. 
 
 18 
 
Wang et al. [2005], require extra caution when using the data. Coexistence of 
advanced and primitive technologies, in addition to the wide variety of fuels, adds to 
the uncertainties in emission estimates. Emission factors are derived from scarce in-
lab experiments [e.g., Chen et al., 2006], which are far from sufficient considering the 
many distinct combinations of combustion devices and fuels in China. Even existing 
emission factors deviate [Streets et al., 2003], sometimes by a factor of five (e.g., CO 
emission factors from small stoves given by Zhang et al. [1999] vs. Venkataraman 
and Rao [2001]). Field experiments can characterize emissions from some processes 
such as biomass burning [e.g., Andreae and Merlet, 2001], but can introduce bias 
when results are extrapolated to a much larger area. Yet another source of uncertainty 
for emission inventories is the abatement technique employed: some big cities like 
Beijing may have more stringent emission control measures in place [e.g., Beijing 
EPA, 2002], although a national clean air act (China?s Air Pollution Prevention Law) 
gives general guide. Additional uncertainties stem from rapid economic growth, 
which makes the air pollution issue of China a ?moving target? [Q. Zhang, personal 
communication]. 
Large uncertainties exist in current emission inventories of China, especially 
for pollutants mainly from low-efficiency combustion processes such as CO and 
black carbon (BC). Streets et al. [2003] put error bars of 150% and 400% on their 
?bottom-up? estimates of CO and BC emissions from China, respectively. This CO 
emission inventory [Streets et al., 2003], along with some others [e.g., Streets and 
Waldorf, 2000], has been tested extensively [Streets et al., 2006]. In general, 
observations in eastern China [T. Wang et al., 2002, 2004; Wen et al., 2008] and its 
 
 19 
 
downwind region [e.g., Huebert et al., 2003],  inverse [Arellano et al., 2004; Heald et 
al., 2004; Palmer et al., 2003; Tan et al., 2004; Y. Wang et al., 2004] and forward 
modeling [Allen et al., 2004; Carmichael et al., 2003; Heald et al., 2003; Kurata et 
al., 2004], and more recently four-dimensional variational data assimilation (4DVAR, 
[Yumimoto and Uno, 2006]) all suggest an underestimate of China?s CO emissions in 
these inventories. A new inventory [Streets et al., 2006], with updated activity rate 
and emission factors for some industrial sources, adjusts the CO emitted from China 
up by about 60%, but still has an uncertainty of almost 70%. Observational and 
numerical studies also help evaluate emissions of other pollutants from East Asia, 
including mercury [Jaffe et al., 2005], ammonia [Kim et al., 2006], and black carbon 
[Huebert et al., 2003]. Ma and van Aardenne [2004] reported large discrepancies in 
NOx, non-methane hydrocarbon compound (NMHC), and CO emissions from China 
between three emission inventories. 
 
1.4.2 Long-range transport of atmospheric pollutants: mechanisms 
Key factors determining the potential for long-range pollutant transport are the 
lifetime of a certain pollutant [e.g., Liu et al., 2005; Liu and Mauzerall, 2005] and 
meteorological conditions. As previously discussed, the pollutant lifetime is 
determined by the rate it is removed from atmosphere via dry deposition, wet 
removal, and photochemical reactions [Seinfeld and Pandis, 1998]. These removal 
processes are all somewhat influenced by meteorological conditions: atmospheric 
stability affects deposition; relative humidity, precipitation, and cloudiness control 
wet removal; temperature and radiation regulate photochemical processes. Pollutants 
 
 20 
 
at higher altitudes, above the boundary layer, tend to exist longer as lower 
temperature slows chemical reactions and deposition diminishes.   
Direction and speed of winds determine the destination of pollutants before 
their removal from the atmosphere. Strong winds in the middle and upper troposphere 
promote long-range transport of pollutants, in some cases driving pollutant plumes 
across ocean basins in just days [e.g., Husar et al., 2001]. In comparison, the 
timescale of transpacific transport in the lower troposphere is typically 2-3 weeks 
[e.g., Holzer et al., 2003, 2005; Liu and Mauzerall, 2005; Liu et al., 2005]. Given a 
longer lifetime and stronger advection, pollutants ventilated from the polluted PBL 
into the free troposphere (FT) can potentially be carried far away from their emission 
sources. The mechanisms lifting pollutants to higher altitudes are therefore crucial for 
long-range transport. 
Long-range transport of pollutants from mid-latitude (30-60? north and south) 
continental regions is related to eastward moving mid-latitude wave cyclones [e.g., 
Berkowitz et al., 1996; Bethan et al., 1998; Bey et al., 2001; Cooper et al., 2001; 
Esler et al., 2003; Hannan et al., 2003; Kaneyasu et al., 2000; Liang et al., 2004, 
2005; Mari et al., 2004; Prados et al., 1999], which are under the control of the large-
scale atmospheric circulation. Long-range transport events from East Asia are 
episodic [e.g., Berkowitz et al., 1996; Jaffe et al., 2003b; Mari et al., 2004; Yienger et 
al., 2000], and show seasonal variability in association with the East Asian monsoon. 
The dominant Siberian anticyclone in winter traps pollutants over East Asia at lower 
altitudes [Newell and Evans, 2000], until prevailing northwesterly winds convey them 
to the southeast, where they are lifted to the upper troposphere by convection and 
 
 21 
 
transported further in the upper branches of Hadley cell circulation [Liu et al., 2003; 
Liang et al., 2004]. With cyclone activities intensifying and convection picking up, 
spring is the peak season for eastward transport of pollutants from East Asia [e.g., 
Holzer et al., 2003; Igarashi et al., 2006; Jaffe et al., 1999; Liu et al., 2003; Liang et 
al., 2004]. Summer months see weaker mid-latitude cyclones; and surface winds from 
the ocean prevail. But strong convection over the hot land surface can sometimes vent 
pollutants into upper troposphere [e.g., Kiley and Fuelberg, 2006], although some of 
the uplifted pollutants are transported around the Tibetan anticyclone to the Middle 
East [Liu et al., 2003]. Eastward transport of Asian pollutants is not as strong in fall 
as in winter [Gregory et al., 1997], as northeasterly winds push pollutants 
southwestward [Liu et al., 2003]. Factors controlling the interannual variability of 
transpacific transport include the strength of Pacific high and Aleutian low [Liang et 
al., 2005; Liu et al., 2005], and El Ni?o and La Ni?a events, whose influence on 
atmospheric circulation may alter the frequency of mid-latitude wave cyclones [e.g., 
Fuelberg et al., 2003; Liu et al., 2003; Liu et al., 2005]. 
Liang et al. [2004, 2005] summarized the major pathways for pollutants to 
leave East Asia, and to enter North America (Figure 1.6). The outflow pathways 
shown in the figure include convective uplift (E1), slantwise ascent in the warm 
conveyor belt (E2), postfrontal boundary layer transport (E3), and low-level 
prefrontal jet transport (E4). Import mechanisms to North America include advections 
in the free troposphere westerly flow (I1), boundary layer transport (I2), large-scale 
subsidence in the Pacific high (I3), subsidence in the postfrontal dry air stream (I4), 
and subsidence due to mountain waves (I5). E2, E3 and E4 associated with mid-
 
 22 
 
latitude cyclones are of synoptic scale. Convective lifting (E1) is of relatively small 
scale, but is often induced by or embedded in frontal systems [e.g., Dickerson et al., 
2007; Kiley and Fuelberg, 2006]. A pollutant lofting mechanism not illustrated in the 
figure is orographic lifting [e.g., Hanna et al., 2003; Liu et al., 2003], which happens 
when air flows interact with uneven terrain.  
 
 
Figure 1.6. Schematic of major pathways for pollutant outflows from East Asia (E1, 
E2, E3, and E4) and inflows into North America (I1, I2, I3, I4, and I5) [Liang et al., 
2005]. 
 
As depicted in Figure 1.6, convection and the warm conveyor belt lifting are 
two major mechanisms ventilating PBL air into free troposphere. The warm conveyor 
belt (WCB) refers to large scale ascending motion ahead of the surface cold front of a 
mid-latitude cyclone system (Figure 1.7). The air stream originating from low 
altitudes in the warm sector of the cyclone sends warm, moist air northward and 
upward along isentropic surfaces, reaching the middle and upper troposphere before 
turning anticyclonically at higher altitudes over the surface warm front [Cooper et al., 
2001; Mari et al., 2004]. As condensation occurs along the way, the warm conveyor 
 
 23 
 
belt can often be identified with a synoptic scale cloud band on satellite images. 
Upward motion in WCB?s is generally slow, typically lifting air from PBL to upper 
troposphere (300 mb) in 2-3 days [Stohl et al., 2002]. The chemical properties of a 
WCB largely depend on its origin [e.g., Bethan et al., 1998]. A WCB originating 
from the PBL over a pollutant source region [e.g., Stohl et al., 2003] is more polluted 
than one initially from a clean marine boundary layer [e.g., Mari et al., 2004], and 
more likely to transport pollutants downwind. 
 Figure 1. 7. Major air streams associated with a mid-latitude cyclone: cold conveyor 
belt (CCB), warm conveyer belt (WCB), and dry intrusion (DI) [Kiely and Fuelberg, 
2006]. 
 
Convection is induced by atmospheric instability, and occurs most frequently 
in summer in mid-latitudes [e.g., Hartmann, 1994; Horowitz et al., 2003; Liang et al., 
2004]. Strong updrafts up to tens of meter per second associated with these local or 
mesoscale processes makes them very efficient in lofting pollutants (Figure 1.8). The 
role of convection in the long-range transport of pollutants has been identified in 
 
 24 
 
observational and numerical experiments [Bertram et al., 2007; Chatfield and 
Crutzen, 1984; Dickerson et al., 1987; Donnell et al., 2001; Pickering et al., 1988; 
1990]. 
  Figure 1. 8. Schematic of CO redistribution by a deep convective process [Dickerson 
et al., 1987]. 
 
Yet to be determined is the relative importance of the warm conveyor belt 
ascent and rapid convection in exporting pollutants from East Asia. Aircrafts 
measurements off the east and west coasts of the Pacific during PEM-West A/B 
[Gregory et al., 1996, 1997], TRACE-P [Jacob et al., 2003], ACE-Asia [Huebert et 
al., 2003], and ITCT 2K2 [Parrish et al., 2004] measured the chemical properties of 
air flows from Asian continent, particularly China. Many studies recognize warm 
conveyor belt lofting as the principal mechanism lifting and transporting pollutants 
emitted from East Asia [e.g., Cooper et al., 2004; Jaffe et al., 1999; Kaneyasu et al., 
 
 25 
 
2000; Martin et al., 2002; Miyazaki et al., 2003]. But these studies rely on data 
collected downwind of China. The backward trajectory models and/or chemical 
transport models employed in these studies can explicitly simulate synoptic-scale 
features like WCB, but need to parameterize convective processes. The coupling of 
the WCB and convection further complicates this issue. Deep convective processes 
embedded in a WCB could either dilute [e.g., Mari et al., 2004] or enrich [e.g., 
Donnell et al., 2001] the pollutants within the rising air stream of a WCB. 
 
1.4.3 Aerosol chemical composition in China 
In recent years, there has been a lot of effort analyzing aerosol chemical 
composition in China. Bulk [e.g., Wang, Zhuang et al., 2006] and size segregated 
[e.g., Lun et al., 2003] samples have been collected from rural and urban sites [e.g., 
Zheng et al., 2005], from the dust-laden west [e.g., Zhang et al., 2003; Xu et al., 
2004] to the industrialized east [e.g., Xu et al., 2002], and from various platforms 
including surface stations [e.g., Sun et al., 2004], towers [e.g., Chan et al., 2005], 
balloons [e.g., Xu et al., 2001], and aircraft [e.g., Hatakeyama et al., 2005]. Inorganic 
ions [e.g., Yao et al., 2003], elements [e.g., Ning et al., 1996], and carbonaceous 
compounds [e.g., Cao et al., 2003; Yang et al., 2005] are among the most commonly 
studied species. Dust storms in spring draw a lot of attention, in light of their severe 
regional environmental effects [e.g., Sun et al., 2005; Zhang et al., 2002]. Generally, 
sulfate has been found to be more abundant than nitrate [e.g., He et al., 2002; Yao et 
al., 2003; Hu et al., 2008], a phenomenon often attributed to the relatively large 
contribution from coal-burning stationary sources vs. mobile sources, although the 
 
 26 
 
latter are growing quickly [Wang, Zhuang et al., 2006]. In winter and spring in N 
China, the cold environment slows gas-phase (OH) oxidation from SO2 to sulfate, the 
proposed major wintertime SO2 conversion pathway in some studies [e.g., Yao et al., 
2003]. Interaction between dust particles and pollutants, via absorption of gases, 
surface reaction, and particle coagulation, has been suggested as another important 
mechanism producing secondary aerosols in spring, particularly in dust storms [e.g., 
Sun et al., 2005; Wang, Zhuang et al., 2007; Yuan et al., 2008]. Single particle 
analysis [e.g., Gao and Anderson, 2001; Sullivan et al., 2007] has found traces of S 
and N compounds on the surface of some airborne dust particles sampled in the 
region. 
 
1.4.4 EAST-AIRE: an international regional experiment 
East Asian Study of Tropospheric Aerosols: an International Regional 
Experiment (EAST-AIRE) is a scientific project jointly conducted and supported by 
both American and Chinese institutes and agencies. The major goals of EAST-AIRE 
are to understand better the origins, transport, physical, chemical, and optical 
properties of dominant natural and anthropogenic aerosols and gases in China, as well 
as to gain preliminary insights into the direct and indirect effects of these aerosols and 
gases on radiation, clouds, and precipitation [Z. Li et al, 2007b]. In view of the urgent 
need for high-quality observational data of aerosols and precursor gases from China, 
a special focus of EAST-AIRE is to make measurements within or near major aerosol 
source regions of China, using instruments traceable to international standards [Z. Li 
et al, 2007b]. Routine measurements of aerosols and radiation parameters are 
 
 27 
 
provided by a nationwide surface remote sensing network, comprised of more than 
twenty stations and three super sites distributed in various climate zones and source 
regions of different aerosol types in China. This long-term monitoring network is 
supplemented by intensive field experiments involving in-situ measurements of trace 
gases and aerosols, from surface sites and instrumented aircraft. 
From February to April 2005, a U.S.-China joint Intensive Field Campaign 
(IFC) was conducted under the auspices of EAST-AIRE. This dissertation benefits 
from the data set from this IFC. Surface measurements of SO2, CO, NO/NOy, O3, 
aerosol absorption, and aerosol scattering were made in Xianghe, a rural site about 70 
km ESE and generally downwind of Beijing in March. Bulk aerosol samples were 
collected during and beyond the ground experiment (results presented in Chapter 3). 
The April aircraft campaign measuring CO, SO2, O3, and aerosol scattering over 
Shenyang, an industrialized city of 6 million residents ~600 km NE of Beijing, 
determined the vertical distribution of pollutants up to 4500 m altitude under a variety 
of synoptic conditions (results given in Chapter 4).  
 
1.4.5 Outline of research 
Some of the specific questions I try to address in this dissertation include:  
1) Current estimates of pollutant emissions from China: how good are they? 
2) What are some of the important meteorological mechanisms lifting 
pollutants from their sources in China into the free troposphere, where they can 
potentially move far downwind? 
 
 28 
 
3) How do some of the pollutants evolve during transport, and can their 
evolution be tracked with state-of-the-art satellite sensors? 
I start with ground measurements of trace gases, aerosol optical properties, 
and aerosol chemical composition. I evaluate a few current emission inventories for 
China by comparing observed and inventory pollutant ratios. Emissions of gaseous 
pollutants from a local small-scale coal-fired boiler are characterized as well. I 
calculate the enrichment factors of some elements in aerosol samples, which provide 
some insights into the origins of aerosols in the region. 
I also analyze the synoptic variation of pollutants, observed from the ground 
as well as from the aircraft. Backward trajectories and weather analysis will help 
determine the sources of our sampled air parcels. Focusing on two research flights 
measuring the chemical properties of the atmosphere around a frontal system, a case 
study reveals the role of upwind dry convection in lofting and transporting pollutants. 
The conversion of SO2 to sulfate aerosols during transport is studied using 
satellite data. An analysis combining aerosol composition, backward trajectories, and 
satellite cloud product studies the conversion process as pollutants are carried into 
northern China by NW flows behind cold fronts. A second analysis makes use of 
forward trajectories, a chemical transport model, and multiple satellite sensors to 
track the evolution of a large pollution plume as it leaves China and moves into the 
northern Pacific, and estimates the decay of SO2 and change of aerosols in the plume.  
 
 
 29 
 
Chapter 2: Experimental Methods 
 
2.1 Sampling Site and Platform 
2.1.1 Xianghe site 
The Xianghe Atmospheric Observatory (39.75?N, 116.96?E; 35 m above sea 
level), a supersite of the EAST-AIRE project, is located about 70 km east-southeast 
and generally downwind of the Beijing metropolitan area. This area of plains is 
surrounded by agricultural land, densely occupied residences and light industry. 
Sitting between two mega-cities (with Beijing to the northwest and Tianjin to the 
southeast, shown in Figure 2.1) and less than 5 km west of the local township 
(population: ~50,000), the site sees pollutants of urban, rural, or mixed origins, 
depending on wind direction. A small coal-fired boiler about 50 m east of the four-
story main facility building provides heating in winter months. 
Routinely measured at the site are aerosol optical thickness, direct and diffuse 
radiation, and sky conditions by instruments installed about 15 m above ground, on 
the top of the main facility building. During the first IFC of EAST-AIRE in March, 
2005, trace gas detectors measuring CO, NO/NOy, SO2, and O3, and instruments 
detecting aerosol absorption and scattering coefficients were deployed in the same 
location, with sampling inlets mounted on the top of the building, ~10 m away from 
the radiation instruments. A high-volume sampler near the aerosol instruments 
collected bulk aerosol samples from March to April. 
 
 30 
 
Instruments on a 32-m meteorological tower about 50 m southwest of the 
main facility building measured ambient temperature, relative humidity (RH), and 
wind speed at five levels (2, 4, 8, 16, and 32 m above ground) every 10 min. A 
weather vane at the top of the tower determined the wind direction, and a barometer 
monitored local atmospheric pressure. During the ground experiment, the area was 
generally cold and dry, with average temperature of ~4.4?C, RH of ~36%, and very 
little precipitation (~0.1 mm in total). 
 
Figure 2.1. Location of the Xianghe site. 
 
2.1.2 Y-12 aircraft 
A modified Chinese Y-12 twin engine turboprop aircraft (Figure 2.2) was 
used in the aircraft experiment. The Y-12 has a cruise speed of about 200 km/hr and a 
ceiling of about 7000 m. A forward-facing isokinetic inlet mounted on the top of the 
 
 31 
 
cockpit, ahead of the engines, directs samples into aerosol instruments. Conductive 
rubber tubing minimizes loss of small aerosol particles. An aft-facing inlet at the 
same spot supplies sampling flow to trace-gas instruments. Teflon inlet line was used 
to avoid ozone loss. Due to losses of coarse aerosol particles to the aerosol sampling 
line, aerosol data collected from the aircraft represent mainly fine particles. The 
sampling efficiency of the inlet for aerosols of different sizes has not been 
quantitatively characterized.  
A RH/T probe (EIL Instruments Inc. Rustrak RR2-252, Hunt Valley, MD) 
installed in between the aerosol and trace gas inlets measured ambient temperature 
and relative humidity, with a precision of 0.5?C in temperature and 2% in RH 
[Taubman, 2004]. A Rosemount Model 2008 pressure transducer provided static 
pressure readings during flight, which were compared to surface meteorological 
records at the airport upon takeoff and landing. The corrected pressure data were then 
converted into pressure altitude (palt) using the 1976 standard atmosphere equation:  
palt = 8303.951372 ? ln(1013.25/pressure)   (2.1) 
where pressure is in mb. Flight positions were recorded with the Global Positioning 
System (Garmin GPS-90) and validated with known geographic reference points. 
During the airborne campaign, theY-12 aircraft was instrumented with CO, 
SO2, and O3 monitors, and an integrating nephelometer measuring aerosol scattering 
coefficients. Taoxian International Airport (41.64?N 123.49?E), where the aircraft was 
based, is ~20 km south-southeast of downtown Shenyang, the capital of China?s 
northeastern province of Liaoning, and a big city (population: 6 million) ~600 km NE 
 
 32 
 
of Beijing. Within 80-km radius of Shenyang are five more industrialized cities of 
400,000 or more residents (Fushun, Benxi, Liaoyang, Tieling, and Anshan). 
 
 
Figure 2.2. Y-12 twin engine turboprop research aircraft used for the aircraft 
campaign. 
 
2.2 Measurements of Trace Gases 
2.2.1 Carbon monoxide 
Carbon monoxide (CO) was measured with a modified (Figure 2.3, 
[Dickerson and Delany, 1988]) commercial non-dispersive infrared (NDIR) gas filter 
correlation (GFC) analyzer (Thermo Environmental Instruments, Model 48, Franklin, 
MA). This instrument utilizes the absorption by CO at its vibration-rotation band of 
4.67 ?m. Radiation from the IR source alternately passes through two gas filters 
before entering the sample cell of ~30 m effective optical path. The gas filter filled 
 
 33 
 
with high concentration of CO produces the reference beam, and the other filter with 
CO-free nitrogen (N2) produces the sample beam. The difference between the two 
beams is used to calculate CO concentration in the sample cell.   
Interference from carbon dioxide (CO2) and water vapor complicates CO 
measurements. The response of the detector also drifts as temperature changes, by as 
much as 200 ppb in 24 hr [Dickerson and Delany, 1988]. Beads of palladium/alumina 
(Pd/Al2O3) heated to about 250?C are used as a catalyst to convert CO to CO2, 
providing a chemical zero. A 3-way solenoid valve, manually activated or timer 
controlled, directs the sample flow to go through (zero mode) or bypass (sample 
mode) the CO zero, generating a dynamic baseline of the detector. During the field 
experiment, filters were placed ahead of trace gas instruments (except for the catalytic 
converter of the NOy detector) to remove particles. On the aircraft, no particle filters 
were used. 
During the ground experiment, the instrument was operated with a 60-minute 
cycle switching between 48 min of sample and 12 min of zero. CO concentration was 
logged with 1-minute resolution and processed into 10-min and hourly averages. 
When deployed onboard the Y-12 aircraft, the zero and sample modes were manually 
controlled, and data were recorded at 10-s resolution. The detector was calibrated 
with CO working standard gas (~1 ppm CO in nitrogen, Scott-Marrin Inc., Riverside, 
CA) traceable to a NIST (National Institute of Standards and Technology) standard 
reference material before and after the field campaign. Its calibration-determined 
detection limit is 40 ppb (signal-to-noise ratio S:N = 2:1 for ?1? noise), for 1-min 
means of 10-s data. 
 
 34 
 
 
 
Figure 2.3. Schematic diagram of the modified CO analyzer (adapted from 
[Dickerson and Delany, 1988]). 
 
2.2.2 Sulfur dioxide 
A modified (Figure 2.4, [Luke, 1997]) commercial pulse-fluorescence detector 
(Thermo Environmental Instruments, Model 43, Franklin, MA) measured sulfur 
dioxide (SO2). The instrument works on the principle that SO2 molecules, 
electronically excited after absorption of UV radiation, can decay back to the ground 
state through fluorescence emission, quenching, or dissociation. Fluorescence 
emission dominates if SO2 molecules are activated by UV radiation in the wavelength 
range of 190-230 nm. The ambient concentration of SO2, over a wide range of 
temperature and atmospheric conditions, is proportional to the fluorescence intensity. 
Aromatic hydrocarbons could interfere with SO2 measurements as they 
respond to UV radiation in a similar manner. These compounds are removed as the 
sample air flows through a hydrocarbon ?kicker? ahead of the fluorescence chamber. 
The sample flow intermittently passes through a potassium carbonate (K2CO3) filter, 
 
 35 
 
which reacts with SO2 and generates a dynamic zero to correct for the baseline drift 
of the instrument.  
During the ground experiment, the SO2 instrument worked on the same 
zero/sample cycle as the CO instrument, with data logged at 1-min resolution and 
averaged into 10-min and hourly means. In the aircraft campaign, the zero/sample 
modes of the SO2 instrument were in synchronization with CO. SO2 concentration 
was recorded with 10-s resolution. The instrument was calibrated with a NIST 
traceable standard material (~5 ppm SO2 in air, Scott-Marrin Inc., Riverside, CA) 
before and after the field deployment, and has a detection limit of 80 ppt (parts per 
trillion, signal-to-noise ratio S:N = 2:1 for ?1? noise) for 1-min means of 10-s data. 
 
 
Figure 2.4. Schematic diagram of the modified SO2 analyzer (adapted from [Luke, 
1997]). 
 
2.2.3 Ozone 
Ozone (O3) concentration was determined employing a commercial analyzer 
(Thermo Environmental Instruments, Model 49, Franklin, MA) based on the ozone 
 
 36 
 
UV absorption at 254 nm. The sample flow is divided into two streams (Figure 2.5), 
one drawn through an ozone scrubber as the reference and the other directly into the 
optical cell as the sample. Regulated by the solenoid valves, the reference and sample 
streams alternately flow into optical cells A and B. The intensity difference between 
the two cells is used to calculate ozone concentration. O3 data were stored at 1-min 
and 10-s resolutions during the ground and aircraft experiments, respectively. 
Calibration with an in-house primary standard (Thermo Environmental Instruments, 
Model 49 PS) before and after the field deployment shows the instrument has a 
detection limit of 4 ppb (signal-to-noise ratio S:N = 2:1 for ?1? noise) for 10-s data.  
 
 
Figure 2.5. Schematic diagram of the TEI O3 analyzer (adapted from [Taubman, 
2004]). 
 
 
 37 
 
2.2.4 Total reactive nitrogen oxides 
A modified [Fehsenfeld et al., 1987; Stehr et al., 2000] commercial 
chemiluminescence detector (Thermo Environmental Instruments, Model 42, 
Franklin, MA) was employed to measure nitric oxide (NO) and total reactive nitrogen 
oxides (NOy). The instrument measures NO based on the quick reaction between NO 
and O3 (generated in the instrument). A fraction (~10%) of the NO2 molecules 
produced in the reaction (NO + O3 ? NO2 + O2) are excited, and release a 
characteristic luminescence in the wavelength range of 600-3000 nm, which is then 
detected by a photomultiplier. NOy species other than NO (e.g., NO2, PAN, and 
nitrate) are first reduced to NO by a molybdenum catalyst, and then measured with 
the same detector. A solenoid valve (Figure 2.6) controls the sample flow to 
alternately go through (NOy mode, 18 min) and bypass (NO mode, 12 min) the 
converter. Very short tubing ahead of the converter minimizes the line losses of nitric 
acid and nitrate aerosols. 
The instrument was calibrated with a working standard (~4.5 ppm n-
propylnitrate in nitrogen, and NO in nitrogen; Scott-Marrin Inc., Riverside, CA) 
before the field experiment. In China the instrument was calibrated with Chinese NO 
standards and via gas-phase titration [e.g., Dickerson et al., 1984]. The NO detection 
limit of the detector is 150 ppt (signal-to-noise ratio S:N = 2:1 for ?1? noise) for 1-
min mean of 10-s data. The conversion efficiency for NO2 exceeds 95% and is 
assumed unity for all NOy compounds. This assumption probably leads to an 
underestimate of some nitrate aerosols that are more difficult to convert to NO [Chen, 
2002]. The overall uncertainty of NOy measurements using this technique is about 
 
 38 
 
?30% [e.g., Stehr et al., 2000]. NO and NOy were only measured during the ground 
experiment, with 1-min resolution and averaged into 10-min and hourly data. 
 
 
Figure 2.6. Schematic diagram of the modified NO/NOy detector (adapted from 
[Chen, 2002]). 
 
2.3 Aerosol Optical Properties 
2.3.1 Aerosol scattering coefficients 
An integrating nephelometer (TSI Model 3563, Shoreview, MN) was used for 
measurements of aerosol scattering coefficients (bsp) at three wavelengths (450, 500, 
and 700 nm). A Lambertian light source (and an orthogonal light detector allow for a 
geometrical integration of the angular distribution of scattered light, from 7? to 170? 
in scattering angle. A bypass aerosol filter is periodically switched inline with the 
inlet to supply a particle-free sample, thereby allowing the instrument to account for 
the light scattered by gases and internal surfaces of the instrument, which is then 
subtracted from the total scattering. Straight conductive tubing less than 1 m in length 
conducted sample flow into the nephelometer. The instrument was calibrated with 
CO2 and particle-free air before and after the field campaign. 5-min records were 
 
 39 
 
saved during ground deployment and processed into 10-min and hourly averages. 
When operated on the aircraft, the instrument output was logged with 30-s resolution. 
For an averaging interval of 5 min, the detection limits of the instrument at 450, 550, 
and 700 nm are 0.44 ? 10-6 m-1, 0.17 ? 10-6 m-1, and 0.26 ? 10-6 m-1, respectively 
(signal-to-noise ratio S:N = 2:1, [Anderson et al., 1996]).  The major source of 
uncertainty, especially when the sampled particles are dominated by coarse-mode 
particles, is the non-ideal forward-scattering truncation of 0-7? [Anderson et al., 
1996]. Following Anderson and Ogren [1998], wavelength-dependent correction 
factors can be calculated using the ?ngstr?m exponent (?) to represent changes of bsp 
with wavelength. A measure of the wavelength-dependence of particle scattering, ? is 
defined as: 
[ ]
)/log(
)(/)(log
21
21
??
??? spsp bb?= .  (2.2) 
where bsp(?1) and bsp(?2) are the aerosol scattering coefficients at wavelengths ?1 and 
?2, respectively. Aerosol scattering measurements with this instrument have an 
estimated uncertainty of ~10% [Anderson et al., 1996]. The heat of the instrument 
reduces the RH below ambient level. The RH in the integrating volume was 17.5% ? 
17.5% (?2?) and never exceeded 42.5% during the ground experiment. The actual 
aerosol scattering in the atmosphere changes with environmental humidity and is 
often determined with the growth factor, an empirically derived relationship between 
bsp and the environmental RH [Charlson et al., 1992; Jeong et al., 2006]. The growth 
factor depends on the aerosol composition and can be very different between different 
areas. For this study, no correction of the humidity effect was performed because 
there is little information available concerning the aerosol growth factors in this area. 
 
 40 
 
The generally dry atmospheric conditions during the ground experiment kept the 
humidity effect to less than 20% most of the time (from a rough estimate). 
 
 Figure 2.7. Schematic diagram of the TSI 3563 3-wavelength integrating 
nephelometer (adapted from [Anderson et al., 1996]). 
 
2.3.2 Aerosol absorption coefficient 
Aerosol absorption coefficient (bap) was determined with a Particle/Soot 
Absorption Photometer (PSAP, Radiance Research, Seattle, WA). This filter-based 
instrument collects aerosols to a filter, and continuously monitors the transmittance of 
the filter (Figure 2.8). The absorption coefficient, with units of m-1, is calculated with 
the following equation: 
??????= IIVAbap 0ln   (2.3) 
where A is the area of the sample spot, V is the volume of sample air that passes 
through the filter during the averaging period (one minute in this study), I0 and I are 
 
 41 
 
transmittances of the filter during the previous and current averaging periods, 
respectively. The particle-free reference flow (Figure 2.8) helps detect the constancy 
of the LED light source. Nonlinearity in instrument response and magnification of 
absorption by the filter are corrected for with manufacturer?s calibration [Bond et al., 
1999]. Routine calibration of PSAP is not applicable due to lack of standard material. 
Under its typical configuration (1-min averaging period), the instrument has a 
detection limit of 1.8 ? 10-6 m-1 (signal-to-noise ratio S:N = 2:1 for ?1? noise) [Bond 
et al., 1999]. Due to high aerosol loadings in Xianghe, I decreased the sampling flow 
from ~2 L/min to ~0.2 L/min to improve the filter-change interval to a manageable 
level (every 2-3 hr on a polluted day). The actual flow rate was calibrated with an 
electronic bubble flow meter (DC-2, Bios International Corp., Butler, NJ). The 
influences of flow rate, spot size, and the interference from non-absorbing aerosols 
were corrected following the method of Bond et al. [1999], which is valid for filter 
transmittance above 0.7. In data processing, data with filter transmittance larger than 
0.65 were kept to maintain the sample size, particularly in early morning hours. This 
introduces some uncertainties to the aerosol absorption data. The corrected aerosol 
absorption coefficient at 574 nm is then extrapolated to 550 nm [Virkkula et al., 2005] 
for comparison with scattering measurements at the same wavelength. It is also 
assumed that the ambient RH has negligible effects on the absorption measurements, 
as RH was mostly relatively low during the experiment. The uncertainty of the 
absorption measurement is about 30% at a 95% confidence level. 
 
 
 
 42 
 
 
Figure 2.8. Schematic diagram of the PSAP (adapted from [Bond et al., 1999]). 
 
2.4 Aerosol Chemical Composition 
2.4.1 Aerosol sampling 
A University of Miami high-volume aerosol sampler (Figure 2.9, [Prospero et 
al., 2003]) collects particles (total aerosols without size cut) on Whatman 41 
quantitative filter paper (8 by 10 in, or 20.32 by 25.4 cm ), at about 15 m above 
ground. A manometer measures the pressure drop as the sampling flow passes 
through an orifice plate. The recorded pressure drop is used to calculate the flow rate, 
which starts at about 800 L/min initially upon filter change and drops as the filter 
becomes loaded with particles. Depending on aerosol loading, the total volume of air 
sampled by each filter ranges from 400 to 500 m3 during the 12-hr sampling period 
(08:00-20:00 local time for daytime samples and 20:00-08:00 for nighttime samples). 
From March 1 to April 13, 85 samples were collected along with blank samples taken 
 
 43 
 
every 10 days. Both exposed and blank filters were stored in a freezer until chemical 
analysis.  
 
Figure 2.9. Left: setup of the bulk aerosol sampler during the Xianghe experiment; 
Right: schematic diagram of the bulk aerosol sampler. 
 
2.4.2 Chemical analysis 
The chemical analysis of the aerosol samplers was conducted by our 
colleagues at the Institute of Atmospheric Physics (IAP), Chinese Academy of 
Sciences (CAS), and is only briefly described here. For element analysis, one-fourth 
of the filters were first digested in concentrated HNO3 (6 mL) and HCl (2 mL), using 
a microwave sample digestion system (PerkinElmer Life and Analytical Sciences, 
Inc., Model Multiwave 3000, Waltham, MA). An inductively coupled plasma mass 
spectrometer (ICP-MS, Agilent Technologies, Inc., Model 7500a, Santa Clara, CA) 
was used to analyze elements including Be, Na, Mg, Al, K, Ca, V, Mn, Fe, Ni, Cu, 
Zn, As, Mo, Ag, Cd, Ba, Tl, Pb, Th, and U. 
One-fourth of each filter was ultrasonically extracted in 50 mL deionized 
water (resistivity: 18 M?/cm) for about 30 min. The filtrate, after passing through a 
microporous membrane (0.45 ?m in pore size), was then analyzed with an ion 
 
 44 
 
chromatograph (IC, Dionex Corporation, Model ICS-90, Sunnyvale, CA) for the 
concentrations of F-, Cl-, NO3-, SO42-, NH4+, Ca2+, Na+, Mg2+, and K+.  All ions have 
detection limits below 0.01 ?g/m3. Quality control and quality assurance procedures 
were routinely applied. More detailed descriptions are available in other studies 
employing the same analysis system [Xu et al., 2007a, b].   
 
 
 45 
 
Chapter 3: Xianghe Ground Experiment 
 
3.1 Abundance of Pollutants: Statistics and Variations 
Some results of this chapter, including the measurements of trace gases and 
aerosol optical properties, comparison to emission inventories, and estimate of black 
carbon emissions have been presented in the Journal of Geophysical Research special 
section on EAST-AIRE [C. Li et al., 2007]. Aerosol chemical composition and the 
emissions from the local small coal-fired boiler were not included in that paper. 
 
3.1.1 Concentrations of trace gases 
One prominent feature of air pollution in Xianghe is the high pollutant levels 
for a non-urban site. Hereafter, results are presented as average ? one standard 
deviation (?) unless otherwise noted. The hourly averaged SO2 (17.8 ? 15.7 ppb, 
maximum: 96.5 ppb), CO (1.09 ? 1.02 ppm, maximum: 6.59 ppm), and NOy (26.0 ? 
24.0 ppb, maximum: 184 ppb) concentrations observed in Xianghe are much higher 
than at a rural site about 50-60 km east of the urban center of Baltimore, MD, USA 
(SO2: 2.8 ? 3.2 ppb, CO: 0.219 ? 0.087 ppm, NOy: 8.1 ? 6.3 ppb; [Stehr et al., 
2000]). The springtime ambient temperature (average: 4.4?C) and solar flux (due to 
large solar zenith angle and heavy aerosol loading) in Xianghe were too low to be 
conducive to the photochemical production of O3, and its concentration was only 29.1 
? 16.5 ppb, with an hourly maximum of 73.4 ppb. 
 
 46 
 
For comparison, during the TRACE-P and ACE-Asia field campaigns in the 
spring of 2001, T. Wang et al. [2004] observed high levels of trace gases at a rural site 
in Yangtze Delta Region, about 1000 km SSE of Xianghe. Their reported CO (0.68 
ppm), SO2 (15.9 ppb), and NOy (13.8 ppb) concentrations were lower than, but 
comparable to, this study. High pollutant levels observed at rural sites in different 
areas probably reflect the large extent of the air pollution issue in China, particularly 
in the more developed and populated East.  
 
3.1.2 Aerosol optical properties and composition 
Heavy aerosol loadings were observed at the Xianghe site. The average 
aerosol scattering (bsp) and absorption coefficients (bap) at 550 nm are (358 ? 363) ? 
10-6 m-1, and (65 ? 75) ? 10-6 m-1, respectively. Aerosol optical properties measured 
with the same type of instruments in Bondville, IL, USA, a continental site at similar 
latitude, by Delene and Ogren [2002] recorded aerosol scattering ((57.7 ? 17.7) ? 10-6 
m-1) and absorption ((4.66 ? 2.27) ? 10-6 m-1) almost an order of magnitude smaller. 
Note that aerosol loadings in Bondville are lower in March than in other months 
[Delene and Ogren, 2002]. At a site NE of Beijing, farther away from the city than 
Xianghe, Yan et al. [2008] found lower but comparable aerosol scattering ((154 ? 160) 
? 10-6 m-1) and absorption ((18.3 ? 14.0) ? 10-6 m-1) coefficients in spring. Table 3.1 
summarizes the statistics of trace gases and aerosol optical properties observed during 
the ground experiment. 
 
 47 
 
Table 3.1. Statistics of trace gases and aerosol optical properties in Xianghe  
 Mean Standard 
Deviation 
Median 10th 
Percentile 
25th 
Percentile 
75th 
Percentile 
90th 
Percentile 
COa 1.09 1.02 0.766 0.252 0.350 1.43 2.34 
SO2b 17.8 15.7 13.8 2.11 4.74 26.9 37.8 
NOyb 26.0 24.0 18.8 4.09 7.04 37.4 56.3 
O3b 29.2 16.5 30.3 5.55 13.7 41.9 49.1 
Scatteringc 358 363 323 45.0 87.0 507 814 
Absorptiond 65.4 74.9 38.7 11.3 16.5 84.9 115 
a Unit: ppm; 
b Unit: ppb; 
c Unit:  10-6 m-1 or Mm-1, wavelength: 550 nm; 
d Unit: 10-6 m-1 or Mm-1. 
 
The statistics of water-soluble ions are given in Table 3.2, which shows fairly 
high concentrations of secondary aerosol species such as sulfate (SO42-, 7.2 ? 6.2 
?g/m3), nitrate (NO3-, 9.5? 9.2 ?g/m3), and ammonium (NH4+, 5.4 ? 4.2 ?g/m3). Most 
interestingly, in contrast to a number of previous studies in northern China, our 
observed NO3- mass concentration ([NO3-]) is higher than that of SO42- ([SO42-]). I 
further discuss the [NO3-]/[SO42-] ratio in Chapter 5. For comparison, Chen [2002] 
found on average 3.3 ?g/m3 of SO42-, 4.1 ?g/m3 of total nitrate (HNO3 + all 
particulate NO3-) and 1.92 ?g/m3 of total ammonium (NH3 + particulate NH4+) in 
April PM2.5 samples collected at a traffic-influenced site in Fort Meade, Maryland. 
Table 3.3 summarizes the statistics of element concentrations for aerosol 
samples from Xianghe. High loadings of crustal elements including Al (9.6 ? 5.3 
?g/m3), Mg (3.1 ? 1.6 ?g/m3), Ca (9.8 ? 4.7 ?g/m3), and Fe (6.5 ? 3.4 ?g/m3) indicate 
existence of soil particles in springtime aerosols in the region. Aerosol samples in 
New Castle, NH, a coastal site in the northeastern U.S. contain much lower Al (0.21 
?g/m3) and Fe (0.15 ?g/m3) [Pike and Moran, 2001]. Comparing the two sites also 
shows much greater concentrations of some pollutant tracer elements in Xianghe (e.g., 
 
 48 
 
Pb: 0.28 ?g/m3 vs. 0.005 ?g/m3; Cd: 4.5 ng/m3 vs. 0.17 ng/m3; Ag: 0.59 ng/m3 vs. 
0.031 ng/m3; Cu: 75 ng/m3 vs. 2.1 ng/m3). For reference, the U.S. EPA NAAQS 
standard for lead is 0.15 ?g/m3 for a rolling 3 month average and 1.5 ?g/m3 for a 
quarterly average [U.S. EPA, 2006]. The ratio between the 90th percentile (the 
concentration below which 90% of the observations fall) and the 10th percentile of 
each element is calculated to estimate its variability. Crustal elements in general have 
ratios between 3 and 5, but the ratios of some pollutant tracer elements (e.g., Cu, Ag, 
As, and Pb) exceed 10, showing large variation in their loadings. Water soluble ions 
such as sulfate, nitrate, and ammonium, aerosol scattering and absorption, as well as 
traces gases including CO, SO2 and NOy are also found to be very variable. Aerosol 
optical thickness (AOT) at the same site fluctuated over a wide range of 0.2-4 during 
the experiment [Z. Li et al, 2007a]. In sections 3.1.3 and 3.1.4, I discuss in more 
details the variations of pollutant levels observed in Xianghe. 
 
Table 3.2. Statistics of water-soluble inorganic ions in aerosol samples from Xianghe 
(?g/m3)a 
Element Mean Standard 
Deviation 
Median 10th 
Percentile 
25th 
Percentile 
75th 
Percentile 
90th 
Percentile 
F- 1.5?10-1 1.2?10-1 1.1?10-1 1.2?10-2 3.1?10-2 2.2?10-1 3.4?10-1 
Cl- 3.4 2.6 3.2 7.0?10-1 1.1 4.6 6.4 
NO3- 9.5 9.2 6.4 9.6?10-1 2.1 13.0 21.1 
SO42- 7.2 6.2 5.2 1.7 2.7 8.9 14.3 
Na+ 6.5?10-1 3.8?10-1 6.2?10-1 2.2?10-1 4.0?10-1 8.6?10-1 1.1 
NH4+ 5.4 4.2 4.4 1.2 1.8 7.4 11.5 
K+ 1.4 1.0 1.1 3.4?10-1 6.0?10-1 2.1 3.1 
Ca2+ 3.4 1.6 3.1 1.2 2.1 4.6 5.3 
Mg2+ 3.0?10-1 1.5?10-1 2.8?10-1 1.2?10-1 1.8?10-1 4.0?10-1 4.8?10-1 
a Part of the Na, Ca, K, and Mg exist in minerals that are not very water-soluble, the 
concentrations of the ions are thus generally lower than the corresponding elements 
(Table 3.3). Uncertainties in both analysis methods may also contribute to the 
differences between water-soluble ions and elements. 
 
 
 
 
 49 
 
Table 3.3. Statistics of elements in aerosol samples from Xianghe (?g/m3) 
Element Mean Standard 
Deviation 
Median 10th 
Percentile 
25th 
Percentile 
75th 
Percentile 
90th 
Percentile 
Be 3.9?10-4 1.7?10-4 3.6?10-4 1.9?10-4 2.6?10-4 4.8?10-4 6.4?10-4 
Na 1.2 5.7?10-1 1.1 4.6?10-1 0.8 1.6 2.0 
Mg 3.1 1.6 2.7 1.1 2.2 4.0 5.0 
Al 9.6 5.3 8.4 3.5 5.8 11.9 15.5 
K 3.6 2.0 3.2 1.2 2.0 5.2 6.8 
Ca 9.8 4.7 9.5 3.5 6.3 13.6 16.2 
V 1.2?10-2 6.5?10-3 1.1?10-2 3.7?10-3 6.5?10-3 1.5?10-2 2.0?10-2 
Mn 1.7?10-1 8.3?10-2 1.6?10-1 6.9?10-2 1.0?10-1 2.2?10-1 2.8?10-1 
Fe 6.5 3.4 5.8 2.3 4.1 8.9 11.4 
Ni 1.2?10-2 3.6?10-2 7.3?10-3 3.3?10-3 4.6?10-3 1.1?10-2 1.6?10-2 
Cu 7.5?10-2 1.6?10-1 2.6?10-2 6.9?10-3 1.3?10-2 7.7?10-2 1.7?10-1 
Zn 4.4?10-1 3.6?10-1 3.9?10-1 4.9?10-2 8.6?10-2 6.3?10-1 9.0?10-1 
As 1.8?10-2 1.4?10-2 1.4?10-2 3.1?10-3 5.7?10-3 2.6?10-2 3.6?10-2 
Mo 2.8?10-3 2.3?10-3 2.5?10-3 4.8?10-4 9.2?10-4 4.1?10-3 5.6?10-3 
Ag 5.9?10-4 4.8?10-4 5.6?10-4 6.6?10-5 1.6?10-4 8.5?10-4 1.2?10-3 
Cd 4.5?10-3 3.7?10-3 3.8?10-3 4.4?10-4 8.1?10-4 6.6?10-3 9.6?10-3 
Ba 9.2?10-2 5.3?10-2 7.9?10-2 3.8?10-2 5.8?10-2 1.1?10-1 1.8?10-1 
Tl 2.2?10-3 2.1?10-3 1.6?10-3 3.5?10-4 5.9?10-4 3.3?10-3 4.6?10-3 
Pb 2.8?10-1 2.4?10-1 2.3?10-1 2.9?10-2 5.4?10-2 4.5?10-1 6.6?10-1 
Th 1.4?10-3 7.6?10-4 1.3?10-3 5.4?10-4 9.0?10-4 1.8?10-3 2.4?10-3 
U 4.5?10-4 2.2?10-4 4.2?10-4 1.9?10-4 2.9?10-4 6.0?10-4 7.7?10-4 
 
3.1.3 Diurnal variation of pollutants 
Diurnal cycles of CO, NOy, SO2, and O3, aerosol scattering (550 nm) and 
absorption are given in Figure 3.1. The error bars are 10th (lower bars) and 90th (upper 
bars) percentiles of the data. With the exception of O3, all parameters show minima at 
around 3 pm local standard time (LST) and accumulate until reaching maxima during 
the rush hour (7-9 am LST) on the following day. Aerosol absorption from 11 pm to 6 
am LST may be underestimated due to filter saturation (fewer measurements on more 
polluted days as filters saturate quickly and cannot be replaced in time). Both 
meteorological conditions and emissions contribute to these diurnal cycles. A 
nocturnal inversion forms near ground on cold nights, trapping locally emitted 
pollutants near ground until around 8 am LST the next morning, when sunlight heats 
 
 50 
 
the surface and breaks up the inversion (cf. meteorological records available at: 
http://www.atmos.umd.edu/~zli/EAST-AIRE/station.htm). Emissions may be strong 
during the morning rush hour, when rural households set up stove fires for cooking 
and heating [X. Wang et al., 2005], and local factories start running. As the 
temperature increases during the day, stronger surface winds and active convection 
dilute pollutants near the surface. Needs for heating reduce, too.  
The error bars in Figure 3.1 demonstrate that CO, NOy, SO2, and aerosol 
scattering are more variable in the morning and evening than at mid-day. The sample 
size for each hour of the day is similar (e.g., 120 and 122 10-min averages during 7-8 
am and 3-4 pm LST for CO, respectively); and the diurnal change in error bars 
probably reflects the emission sources detected by the instruments. Local emission 
sources are diverse, as noticed by the author of this dissertation, who identified coal-
fired boilers, two-stroke, four-stroke and diesel vehicles, biofuel, liquefied propane, 
and agricultural burning near the site. In stagnant morning and evening hours, various 
nearby sources may dominate the pollutants captured by the instruments. Stronger 
mixing and winds during daytime transport mixed plumes from a larger region to the 
site, reducing variability. The variability of aerosol absorption cannot be properly 
determined between 11 pm and 6 am because of the filter saturation issue mentioned 
above. 
 
 51 
 
 
Figure 3.1. Diurnal cycles of CO, NOy, aerosol absorption, SO2, aerosol scattering 
(550 nm), and O3 at the Xianghe site. Lines show the median; error bars are 10th 
(lower bars) and 90th (upper bars) percentiles of the data. Note that the value and 
variability of the aerosol absorption during 11 pm-6 am LST cannot be properly 
determined due to saturation of the filter in the instrument. 
 
Table 3.4 compares some elements and water-soluble ions in the daytime and 
nighttime samples. Secondary aerosol species such as sulfate (SO42-) and nitrate 
(NO3-) are slightly more abundant in nighttime samples. The shallow nocturnal 
boundary layer may accumulate pollutants near ground at night, but photochemical 
production of these components is stronger in daylight. Pollution-related elements 
(Cu, As, Ag, Cd, and Pb, cf. section 3.3 for their enrichment factors) are all higher at 
 
 52 
 
night as well. Crustal elements including Mg, Al, K, Fe, and Ca, on the other hand, 
are less consistent, with slightly more K and Fe at night but more Mg and Al in 
daytime. Note that the so-called ?daytime? samples were mostly collected from 8 am 
to 8 pm LST, thus incorporating the valley as well as part of the peak in the diurnal 
cycles of aerosols as shown in Figure 3.1.  
 
Table 3.4. Concentrations of some elements and water-soluble ions in daytime and 
nighttime samples (average ? standard deviation, ?g/m3) 
 Daytime Nighttime 
Cl- 2.7 ? 2.0 4.1 ? 3.0 
NO3- 8.6 ? 9.2 10.4 ? 9.2 
SO42- 6.9 ? 7.1 7.4 ? 5.4 
Na+ (5.6 ? 3.1) ? 10-1 (7.4 ? 4.3) ? 10-1 
NH4+ 4.8 ? 4.5 5.9 ? 4.0 
K+ 1.3 ? 1.0 1.5 ? 1.0 
Ca2+ 3.4 ? 1.6 3.3 ? 1.6 
Mg 3.1 ? 1.6 3.0 ? 1.6 
Al 10.0 ? 5.8 9.2 ? 4.8 
K 3.4 ? 2.0 3.8 ? 2.1 
Ca 9.8 ? 4.6 9.8 ? 4.9 
Fe 6.3 ? 3.3 6.7 ? 3.4 
Cu (6.1 ? 10.9) ? 10-2 (8.8 ? 19.2) ? 10-2 
As (1.5 ? 1.5) ? 10-2  (2.0 ? 1.4) ? 10-2  
Ag (5.2 ? 4.4) ? 10-4 (6.7 ? 5.2) ? 10-4 
Cd (4.0 ? 3.8) ? 10-3 (5.0 ? 3.6) ? 10-3 
Pb (2.3 ? 2.2) ? 10-1 (3.3 ? 2.4) ? 10-1 
 
3.1.4 Synoptic variation of pollutants 
Hourly averaged aerosol absorption coefficient and scattering coefficients at 
450, 550, and 700 nm are shown in Figure 3.2, along with locally-recorded 
temperature, relative humidity, surface pressure, wind direction, and wind speed. 
Hourly mean trace gas (CO, SO2, NOy, and O3) concentrations are given in Figure 3.3, 
as is the AOT measured with the co-located sun photometer [Z. Li et al, 2007a]. The 
 
 53 
 
gaps in both plots are due to a power failure (on March 7, day of year (DOY) 66), 
routine maintenance of instruments, and instrument malfunctions. PSAP filter 
saturation due to heavy aerosol loadings causes most of the gaps in PSAP data in 
early morning hours (11 pm -6 am LST).  
Fast transitions between highly polluted and relatively clean episodes are seen 
in both figures. Surface weather analysis (Korean Meteorological Administration: 
available at http://abc-gosan.snu.ac.kr/sfc.html) identifies five cold-front passages on 
March 5-6 (DOY 64-65), March 10-11 (DOY 69-70), March 16-17 (DOY 75-76), 
March 19-20 (DOY 78-79), and March 21-22 (DOY 80-81). These frontal passages 
led to an abrupt change in wind direction (from generally south to generally north), an 
increase in atmospheric pressure, and a decrease of temperature and relative 
humidity.  
Shaded areas roughly representing the analyzed cold front passages in Figures 
3.2 and 3.3, by no coincidence, also mark dramatic changes in air quality. Pollutants 
accumulate before cold front passages and quickly disperse afterward, as prefrontal 
continental tropical (cT) air mass is replaced by postfrontal continental polar (cP) air 
mass. The high wind speed behind cold fronts facilitates ventilation and vertical 
mixing, reducing surface pollution. Descending air from less populated northern 
regions (Mongolia and Siberia) also brings in relatively clean air. Postfrontal morning 
ozone, on the other hand, is higher than under average conditions (~40 ppb vs. few 
ppb), while CO (a few hundred ppb), SO2 (a few ppb) and NOy (a few ppb) are lower. 
The high wind speeds behind cold fronts dilute primary pollutants and increase the 
thickness of the nocturnal PBL, resulting in reduced titration of O3 by local pollutants. 
 
 54 
 
 
Figure 3.2. Hourly averaged aerosol scattering (450, 550, and 700 nm), aerosol 
absorption (550 nm), relative humidity (RH, %), temperature (T, ?C), surface 
pressure (kPa), wind direction (WD, degree), and wind speed (m/s) observed at the 
Xianghe site. Time periods with analyzed cold front passages are indicated 
approximately with hatched areas. 
 
 
 55 
 
 Figure 3.3. Hourly averaged concentrations of SO
2, CO, NOy, and O3 observed at the 
Xianghe site. Time periods with analyzed cold front passages are indicated 
approximately with hatched areas. The top panel shows the aerosol optical thickness 
(at 500 nm) determined with the co-located Cimel sun photometer. 
 
Figures 3.4 and 3.5 present the time series of some water soluble ions (SO42-, 
NO3-, NH4+, Cl-, and K+) and pollutant tracer elements (Cu, Zn, As, Cd, and Pb). CO 
 
 56 
 
is shown in the two figures as a reference. The good correlation between these aerosol 
compounds and CO (R2: CO and NH4+, 0.43; CO and Cl-, 0.66; CO and SO42-, 0.47; 
CO and NO3-, 0.48; CO and K+, 0.32; CO and Cu, 0.35; CO and As, 0.52; CO and Zn, 
0.50; CO and Cd, 0.33; CO and Pb, 0.50) suggests that they are likely mainly related 
to anthropogenic sources, and largely controlled by synoptic conditions. The 
exceptionally large peaks in Cl- and Cu on DOY 100-101 remain unaccounted for, but 
are probably from certain special aerosol sources. The synoptic variation in the 
relative ratio between sulfate and nitrate is another interesting feature and will be 
discussed in Chapter 5. 
Given in Figure 3.6 are time series of some crustal elements as well as CO and 
wind speed. Red shaded areas in the plot mark periods of high wind speed or 
relatively low CO level (DOY 66). With a few exceptions, an outbreak of strong 
surface winds is usually accompanied by a peak in Al and a drop in CO, as strong 
winds mobilize dust, either locally or from upwind desert regions (average wind 
speed for Al above 75th percentile: 3.5 m/s, for Al below 25th percentile: 3.2 m/s, for 5 
highest Al concentration: 5.6 m/s). Note that the high Al peak sometimes lasts only 1-
2 aerosol samples or 12-24 hr, shorter than the entire period during which CO is 
maintained low by high wind speed. High dust levels are sometimes but not always 
observed under windy postfrontal conditions. Mg has a variation similar to Al 
(correlation coefficient R2 = 0.77). Strong winds sometimes also boost the loadings of 
Ca, Fe, and K, but overall these elements are more abundant under calm and polluted 
conditions (weak winds and high CO, average wind speed for Ca above 75th 
percentile: 2.9 m/s, for Ca below 25th percentile: 3.1 m/s; for K above 75th percentile: 
 
 57 
 
2.8 m/s, for K below 25th percentile: 3.4 m/s; for Fe above 75th percentile: 2.8 m/s, for 
Fe below 25th percentile: 3.7 m/s ), likely due to substantial local sources related to 
anthropogenic activities. A fraction of K is probably from biomass burning, as 
evidenced by the correlation between CO and water-soluble K+. Construction sites 
could release calcium-containing particles to the atmosphere [Wang, Zhuang et al., 
2005]. Local sources of crustal elements in airborne particles also include traffic, 
agriculture, and some other activities. Figure 3.6 appears to suggest that Al would 
serve as a better tracer for dust transported into the region than other crustal elements. 
 
 58 
 
 
Figure 3.4. Time series of some water-soluble ions (NH4+, Cl-, SO42-, NO3-, and K+) 
in aerosol samples from Xianghe. 
 
 
 59 
 
 
Figure 3.5. Time series of some pollutant tracer metal elements (Cu, As, Zn, Cd, and 
Pb) in aerosol samples from Xianghe. Cu concentration for the samples from DOY 
101-101.5 is 1.26 and 0.55 ?g/m3, respectively. 
 
 
 60 
 
 
Figure 3.6. Time series of some crustal elements (Fe, Ca, K, Al, and Mg) in aerosol 
samples from Xianghe. 10-m wind speed averaged to sampling time is shown in the 
top panel. Shaded areas indicate outbreaks of strong winds or drop in CO 
concentration (DOY 67). 
 
 61 
 
3.2 Pollutant Ratios and Implications for Regional Emissions 
3.2.1 Pollutant ratios from correlation analysis 
Different types of emissions sources often have distinct emission profiles. For 
instance, a boiler used in electricity generation has a high flame temperature and 
excess air supply, emitting a lot of NOx but little CO. On the other hand, a coal-
burning stove for household heating usually has a lower flame temperature and 
limited ventilation, releasing substantial CO but less NOx. Simultaneous 
measurements of NOx and CO can help characterize the emission sources. Once 
emitted into atmosphere, NO reacts rather quickly, and in this study the ratio between 
CO and NOy is used.  
The ratios of CO/NOy, CO/SO2, bsp/SO2, and bsp/bap during the morning rush 
hour (7-9 am LST) and mid-day (2-4 pm LST) are obtained via correlation analysis of 
simultaneous measurements (10-min means). As shown in Figure 3.7, the correlation 
coefficients (R2) for CO/NOy, CO/SO2, bsp/SO2, and bsp/bap are higher in the 
afternoon than in the morning. This supports the argument that the afternoon data 
draw information from better-mixed and more ?regional? plumes, while the early 
morning data contain information from less-mixed and more ?local? plumes. Table 
3.5 summarizes the results of the correlation analysis.  
Table 3.5. Results of the pollutant concentration correlation analysisa. 
 CO/NOy CO/SO2 bsp/SO2b bsp/bap 
 AM PM AM PM AM PM AM PM 
slope 35.4 27.2 50.0 35.8 1.4?10-5 2.3?10-5 4.4 5.8 
y-intercept 304.0 180.7 330.4 181.6 9.0?10-5 -1.6?10-5 5.1?10-6 -1.7?10-5 
R2 0.44 0.79 0.47 0.83 0.54 0.88 0.85 0.92 
aAbout 100 data points (10-min means) for each correlation analysis;  bUnit: m-1/ppb. 
 
 62 
 
 
Figure 3.7. Correlations of CO and NOy, CO and SO2, aerosol scattering (550 nm) 
and SO2, aerosol scattering (550 nm) and aerosol absorption (550 nm) in the morning 
rush hour (7-9 am LST, left panels) and in the mid-afternoon (2-4 pm LST, right 
panels) at the Xianghe site. All data points are 10-min averages. The equations shown 
in the plots are the best-fit lines through the data. 
 
The CO/NOy ratio is 35.4 in the morning and 27.2 in the afternoon. 
Industrialized regions in China have more of high-efficiency combustion sources 
 
 63 
 
releasing little CO but much NOy (e.g., large point sources), whereas rural areas have 
more of low-efficiency sources emitting CO but little NOy (e.g., residential stoves). 
The CO/NOy ratio in the more ?regional? afternoon data is thus smaller. Small-scale 
boilers (~1 MW) operate during daytime in local factories, with a high flame 
temperature and excess air supply to optimize burning efficiency [D. Zhou, personal 
communication], may also emit flue gas with a relatively low CO/NOy ratio. Later in 
this chapter (section 3.4), I determine the emission factors for the small coal-fired 
boiler that heats the site.  
CO and SO2 have some emission sources in common. Coal is the dominant 
energy source in China and is widely used in the north for household heating and 
cooking. He and Chen [2002] reported that the average sulfur content of coal in 
China is 1.1-1.2% by weight, but can be as high as 3%. Low-efficiency stoves 
burning sulfur-containing coal emit both SO2 and CO in substantial amounts. Most 
thermal power plants in China run on coal, and are likely important sources of SO2 
but not CO. The CO/SO2 ratio in more ?regional? data (35.8) is lower than that in 
more ?local? morning data (50.0), reflecting the difference in source composition.   
Compared to previous studies, the y-intercepts of the afternoon CO/NOy and 
CO/SO2 scatter plots are about 180 ppb, close to the mid-latitude continental 
background [Novelli et al., 1998]. The slopes (CO/NOy and CO/SO2 ratios) of the 
afternoon scatter plots are similar to those observed in rural eastern China (CO/NOy: 
35 vs. 37; CO/SO2: 36 vs. 29, the latter values are from T. Wang et al., [2002; 2004]).  
During the TRACE-P campaign in spring 2001, downwind aircraft measurements  
found a similar CO/SOx (SO2 + sulfate) ratio (~27) but much higher CO/NOy ratio 
 
 64 
 
(~55) in plumes from the Beijing area (altitude < 2 km, air mass age < 1 day) 
[Carmichael et al., 2003]. 
Aerosol scattering is fairly well-correlated with SO2. As shown in section 3.1 
and earlier chemical composition researches [e.g., He et al., 2001; Zheng et al., 2005], 
sulfate is one of the major aerosol compounds in the region. The bsp/SO2 ratio is 
~50% higher in the afternoon than in the morning, probably showing the aging of 
primary pollutants. Also, the relative role of organic aerosols is likely higher in the 
morning when emissions from vehicles and biofuel burning are more important. 
Additionally, strong mixing within the daytime PBL may bring ?yesterday?s 
pollutants? in the residual layer (the layer above the shallow nocturnal boundary 
layer) down to the ground level. The pollutants from the residual layer are often of 
greater chemical age [e.g., Taubman et al., 2004] and more thoroughly reacted. Lack 
of vertical profiles of pollutants prevents further investigation of this mechanism. 
The bsp/bap ratio is slightly higher in the afternoon than in the morning. This 
could be explained with BC emitted from numerous local low-efficiency stoves. 
Photochemical conversion from precursor gases to non-absorbing aerosols (e.g., 
sulfate and nitrate) can also improve bsp/bap ratio over time. Aged aerosols in the 
residual layer, possibly mixed down to ground level during daytime as discussed 
above, are often ?brighter? (more scattering). The single scattering albedo (SSA), a 
critical parameter in aerosol radiative forcing research, is defined as the ratio between 
aerosol scattering and aerosol extinction (scattering + absorption). Based on the 
bsp/bap correlation, the SSA at 550 nm is near 0.81 in the morning and around 0.85 in 
the afternoon. SSA retrieved with the co-located Cimel sun photometer was about 0.9 
 
 65 
 
during the same period [Z. Li et al, 2007a]. The difference originates presumably 
from the aerosols actually detected by the two methods, since the Cimel retrieval is 
sensitive to aerosols in the whole atmospheric column, while the PSAP and 
nephelometer measure surface values. Compared to the other experiments in China, 
SSA derived from this experiment is lower than in the Yangtze delta region (0.93, [Xu 
et al., 2002]), but similar to that in Guangzhou (0.86, [Andreae et al., 2005]) and 
Beijing (0.81, [Bergin et al., 2001]). Overall, measurements conducted in recent years 
agree that aerosols near the surface in the eastern part of China are light-absorbing. 
 
3.2.2 Implications for CO and SO2 emissions 
As discussed in Chapter 1, emission inventories can be evaluated by 
comparing observed and inventory pollutant ratios [e.g., T. Wang et al., 2002, 2004]. 
Here I further examine some contemporary emission inventories for China, by 
contrasting the observed CO/SO2 and CO/NOy ratios from this study to inventory 
values. 
 
Table 3.6. Observed and national inventory pollutant ratios (ppbv/ppbv)  
 This study Wen et al. 
[2008] 
Streets et 
al. [2003] 
Streets et 
al. [2006]a 
X. Wang et 
al. [2005]b 
CO/SO2 35.8 72.1 13.0 17.5 7.7 
CO/NOx 27.2 25.4 16.7 22.7 8.6 
a Using SO2 and NOx emission data for 2000, from Streets et al. [2003]. The base 
year for CO emissions is 2001. 
b Emission inventory is regional. 
 
Table 3.6 lists the CO/SO2 and CO/NOx (CO/NOy) ratios from this study, an 
experiment carried out at street level in downtown Beijing [Wen et al., 2008], an 
 
 66 
 
inventory study for East Asia [Streets et al., 2003], an updated CO emission inventory 
for China after TRACE-P field campaign [Streets et al., 2006], and a high-resolution 
regional inventory covering eastern China [X. Wang et al., 2005]. Afternoon ratios 
from this experiment, presumably reflecting more ?regional? emission characteristics, 
are used in the comparison. The different ratios between the two experimental studies 
probably reflect the difference between rural and urban emission sources. The lower 
CO/NOx ratio and higher CO/SO2 ratio observed by Wen et al. [2008] in downtown 
Beijing are attributed to mobile sources, which emit CO and NOx  (but with a lower 
CO/NOx ratio than residential sources [X. Wang et al., 2005]), and little SO2. Also, 
high-sulfur coal has been gradually replaced by low-sulfur fuels in Beijing since 
1998, as part of the government?s efforts to improve air quality [Brajer and Mead, 
2003], but not in rural areas like Xianghe and smaller cities/towns. Despite these 
differences, rural and urban sites both record higher CO/SO2 ratios than the 
inventories. The CO/SO2 ratio derived from our experiment is a factor of three higher 
than Streets et al. [2003] and a factor of four higher than X. Wang et al. [2005], 
implying an underestimate of CO emissions by these earlier inventory studies. SO2 
emissions in inventories are considered more robust than CO [Streets et al., 2003], 
but can also be biased, as the NOy/SO2 ratio from this experiment (1.18) is about 40% 
higher than the inventories (0.78 [Streets et al., 2003] and 0.89 [X. Wang et al., 
2005]). Another possibility is that SO2 emissions in this region went down from 2000 
to 2005. However, this is in contrast with government statistical data (Chapter 1). 
Given a 36% boost in their estimated CO from China, the new inventory by Streets et 
al. [2006] has a greater CO/SO2 ratio, but is still only half of our observed value. 
 
 67 
 
Note that quantitatively evaluating emission inventories with our current observations 
is subject to uncertainties. Mismatch in spatial coverage between bottom-up and in-
situ methods is one such uncertainty, and will be discussed in section 3.2.3. Some 
other uncertainty sources are addressed in section 3.2.4.   
 
3.2.3 Cluster analysis and implications for regional emission inventory 
Figure 3.8 presents the inventory ratios of CO/SO2 and CO/NOy for March 
2001, calculated from the updated CO inventory by Streets et al. [2006] and their 
earlier estimates of SO2 and NOx emissions [Streets et al., 2003]. Estimated CO 
emissions from each 0.5?0.5? grid cell shown in the figure exceed 10000 tonne for 
the month (The Xianghe cell, 39.5-40?N, 116.5-117?E, emits ~30000 tonne CO). The 
blue and red lines in the upper panel of the figure are 24-hr backward trajectories 
calculated using the NOAA Air Resource Lab (ARL) HYSPLIT model [Draxler and 
Rolph, 2003] and NCEP reanalysis meteorological data. Trajectories are initiated 
daily at 3 pm LST, 500 m above ground level over Xianghe; and could be classified 
into two groups: one from west and southwest and often relatively polluted (red lines: 
DOY 65, 68, 73-75, 78, 80, and 85), and the other from north and northwest and 
sometimes relatively clean (blue lines: DOY 61-64, 66-67, 69-72, 76-77, 79, and 81-
84). A few trajectories passing over the grid cell of downtown Beijing (39.5-40?N, 
116-116.5?E) are marked in green in the lower panel (DOY 65, 74-75, 78, and 80). 
 
 
 68 
 
 
Figure 3.8. Emission inventory ratios of CO/SO2 (upper panel) and CO/NOx (lower 
panel) in March, 2001 [Streets et al., 2003, 2006]. Only grid cells with CO emissions 
exceeding 10000 tonne/month are shown. The upper limit of color scale is set at 50 
for clearer view of the spatial patterns, although some grid cells have ratios higher 
than 50. Blue, red and green lines are 24-hr backward trajectories started at 3 pm LST 
each day, 500 m above ground over Xianghe, calculated with the HYSPLIT model, 
using NCEP global reanalysis data. Blue lines, red lines (upper panel), and green lines 
(lower panel) represent trajectories passing over different areas (area to the N & NW 
 
 69 
 
of Xianghe with high CO/SO2 inventory ratios, area to the W & SW of Xianghe with 
low CO/SO2 inventory ratios, and the grid cell containing Beijing). 
 
The CO/SO2 inventory ratios demonstrate an obvious spatial pattern over the 
region around Xianghe, with the ratios in the north (higher than 45) more than twice 
as high as in the south (around 20). The CO/SO2 ratio for the Beijing grid cell is even 
lower at about 12.5, while the same ratio for the Xianghe grid cell is 22.4. CO/NOy 
(NOx) inventory ratios have a different spatial pattern, lower in the Beijing grid cell 
and the area immediately to the southeast of Xianghe. Table 3.7 lists the pollutant 
inventory ratios for areas north and northwest of Xianghe, west and southwest of 
Xianghe, the Beijing grid cell, and the Xianghe grid cell. Results of correlation 
analysis for different trajectory groups are summarized in Table 3.8.   
 
Table 3.7. Regional inventory pollutant ratios within different areas around Xianghe 
(ppb/ppb) 
Streets et al. 
[2003] 
Streets et al. 
[2006]a Area Latitude Range Longitude Range 
CO/SO2 CO/NOx CO/SO2 CO/NOx 
N & 
NW 40 ? 41 ?N 
114.5 ? 117.5 
?E 45.7 28.2 57.2 35.3 
W & 
SW 37.5 ? 40 ?N 114 ? 117 ?E 10.1 14.7 16.5 24.0 
Beijing 39.5 ? 40?N 116 ? 116.5 ?E 10.9 16.3 12.5 18.7 
Xianghe 39.5 ? 40 ?N 116.5 ? 117 ?E 15.0 19.1 22.4 28.6 
a Use SO2 and NOx emission data from Streets et al. [2003]. 
 
Table 3.8. Results of correlation analysis for different trajectory groups  
North and Northwest Southwest and West Beijing y/x 
a b R2 a b R2 a b R2 
CO/SO2 33.1 193.6 0.469 36.7 167.5 0.847 41.4 48.4 0.842 
CO/NOx 40.6 64.4 0.525 21.8 335.4 0.758 36.5 -91.6 0.908 
NOx/SO2 0.71 3.7 0.665 1.47 -3.8 0.832 1.05 5.65 0.775 
a: slope (ppb/ppb), b: y-intercept 
 
 70 
 
 Air parcels from N & NW of Xianghe can have very different chemical 
characteristics at times; SO2, NOy, and CO from that sector (blue lines, upper panel 
Figure 3.8) are only moderately correlated. Brisk descending flows from north or 
northwest bring clean air with background level CO (~200 ppb) and very little SO2 
and NOy. Weak winds  from the same direction (e.g., on March 20), on the other hand, 
can build up ~1 ppm CO and ~10 ppb SO2 and NOy, probably emitted by local 
agricultural burning. Such diverse conditions complicate direct comparison between 
emission inventories and measurements. The measurement-derived CO/SO2 ratio of 
33 is lower than the inventory ratio (57), but the CO/NOy ratios (observation: 41, 
inventory: 35) are close. The high CO/SO2 inventory ratios in the region are due to 
traffic sources, estimated to contribute more than 70% of total CO emissions in the 
area [Streets et al., 2006], even for the mountainous areas of 40.5-41?N, 116-117?E. 
Total estimated traffic CO emissions from these two grid cells are 310000 tonne (in 
2001), only slightly lower than from downtown Beijing (386000 tonne) [Streets et al., 
2006]. Reexamining mobile sources in China is called for to reduce the uncertainties 
revealed here. 
 CO is well-correlated with both SO2 and NOy in air from west and southwest 
of Xianghe (red lines, upper panel Figure 3.8). The observed CO/SO2 ratio associated 
with these trajectories is 36.7, more than a factor of two higher than the inventory 
value (W & SW, Table 3.7). The y-intercept of ~170 ppb is close to the background 
CO concentration. As for the CO/NOy ratio, the observed value (21.8) is not far from 
the inventory value of 24.0, but the y-intercept is more than 300 ppb of CO. The 
CO/SO2 ratio under windy conditions (wind speed > 3 m/s) is also calculated to 
 
 71 
 
minimize influence from nearby sources; the resulting R2 (0.656) and y-intercept 
(138.5 ppb) are lower, but the CO/SO2 ratio remains about the same (38.6). A couple 
of trajectories might have passed over a grid cell about 600 km west of Xianghe, 
where CO emitted from the large steel mills in the city of Baotou could affect our 
measurements. But excluding the data of these two trajectories does not significantly 
change the observational CO/SO2 ratio (35.6, R2: 0.752, y-intercept: 192.9 ppb). In 
short, our CO/SO2 correlation analysis for air from west and southwest of Xianghe 
consistently derives CO/SO2 ratio over 100% higher than the inventory, lending us 
some confidence concluding that CO emissions might be underestimated for the area. 
Or SO2 emissions could be overestimate or SO2 emissions have decreased since 2000. 
 For the trajectories passing over the Beijing grid cell, the CO/SO2 ratio (41.4) 
from our measurements is not as high as what is directly observed downtown (Table 
3.6, [Wen et al., 2008]), but still a factor of three higher than the inventory value 
(12.5). The grid cell west of downtown Beijing (39.5-40?N, 115.5-116?E) has a higher 
CO/SO2 ratio, but only emits one fifth as much CO. Our CO/NOy ratio from this 
trajectory group (36.5) is higher than Wen et al. [2008], and is double the 
corresponding inventory ratio (18.7). During 2000-2004, CO/SO2 and CO/NOx ratios 
based on Beijing EPA?s yearly average data from their urban monitoring network 
were 89 and 56, respectively [T. Wen, personal communication]. These ratios could 
be biased high, however, as SO2 and NOx would be removed through fast 
photochemical processes in summer months. The CO/SO2 inventory ratio for the grid 
cell immediately north of Beijing (40-40.5?N, 116-116.5?E) is greater (Figure 3.8), 
with strong traffic CO emissions. This is probably true as the city has expanded 
 
 72 
 
rapidly northward in recent years. Even after combining inventory emissions from 
this northern grid cell with the downtown grid cell (i.e. consider the area of 39.5-
40.5?N, 116-116.5?E as the urbanized area of Beijing), the resulting inventory 
CO/SO2 (17.0) and CO/NOy (22.3) ratios are still considerably smaller than 
observations. Thus I conclude that CO emissions from Beijing are likely 
underestimated (or SO2 emissions are overestimated) by inventory studies. Detailed 
analysis of pollutant emissions in downtown Beijing is given by Wen et al. [2008]. 
Worth noting is that current emission inventories [Streets et al., 2003, 2006] tend to 
have rather low CO/SO2 (NOx) ratios in Chinese mega cities such as Beijing, Tianjin 
(Figure 3.8, 39-39.5?N, 117-117.5?E), Shanghai, and Guangzhou. Characterizing 
emissions from mobile sources and/or small-scale boilers [Streets et al., 2006] may 
help reduce the uncertainties in emission inventories for big cities in China. In section 
3.4, I present a case study on the trace gas emissions from a small coal-fired boiler. 
 
3.2.4 Some other issues of inventory evaluation 
 Besides the spatial issue discussed above, some other issues could also 
complicate the comparison between inventories and measurements. Measurements 
are normally only available for a limited period of time, whereas emission inventories 
are often based on annual activity data. The vertical distribution of pollutants could be 
another issue for pollutants with different release heights. Losses of pollutants during 
transport need to be taken into account as well. 
 As discussed in Chapter 1, CO has a lifetime much longer (30-90 days) than 
SO2 and NOx (hours to days), and as a result the observed CO/SO2 and CO/NOy ratios 
 
 73 
 
are expected to be greater than those of actual emissions. However, the effects of 
chemical reactions are likely small in this study, due to the proximity of our site to 
heavily populated areas and emission sources. Trajectory analysis (Figure 3.8) 
suggests that in some cases plumes from Beijing could reach Xianghe in hours. In 
addition, chemical conversion of locally emitted SO2 and NOx during our experiment 
is slowed by the relatively cold and dry environment, and limited sunshine, as 
evidenced by the observed ozone level. As will be detailed in Chapter 5, the mole 
ratio between sulfate and SOx (sulfate + SO2) is less than 10% most of the time. 
Taking sulfate into consideration, the CO/SOx ratio is above 30 for W & SW 
trajectories, still almost double the inventory value. 
 Surface measurements may not exactly mirror all emissions in an area if 
pollutants are emitted from different heights and have distinct vertical distributions. 
In China, a large fraction of CO is from area sources near the surface, though steel 
mills inject CO at relatively high altitudes. SO2 is different; large point sources emit a 
substantial part of total SO2. Therefore the CO/SO2 ratios measured at the surface 
would be higher than inventory values. This sampling bias is partly avoided in this 
study, as I select measurements from local afternoon for comparisons, when vertical 
mixing in the PBL is most vigorous. As a test, I calculated the regional CO/SO2 
inventory ratio, excluding large point sources of CO and SO2. The resulted surface-
source-only-ratio for the area W & SW of Xianghe (37.5-40?N, 114-117?E) is about 
20.8 [Streets et al., 2006], 25% larger than the all-source-ratio (16.5), but still 70% 
smaller than our value from afternoon observations. 
 
 74 
 
 The inventory studies compared here are based on statistics for the whole 
years of 2000-2001, while this experiment only lasted one month in 2005. This 
mismatch can also hinder direct comparison between observations and inventories, 
although the inventory emissions of trace gases in March are close to annual averages 
[Streets et al., 2003], and no discernable change occurred in the monitored CO/SO2 
ratio in Beijing from 2000 to 2004 [T. Wen, personal communication]. In summary, 
different spatial and temporal coverage, losses of pollutants, and uncertainties in 
measurements and emission estimates all contribute to the discrepancies discussed in 
this section. Reducing the gap between observations and bottom-up estimates helps 
improve emission inventories. The better agreement between the new CO inventory 
for China [Streets et al., 2006] and our measurements reflects an important 
achievement of the TRACE-P experiment. More extensive field experiments and 
direct measurements of emission factors should help improve understanding of 
pollutant emissions from China. 
 
3.2.5 Black carbon emissions 
 The correlation between CO and BC, both products of inefficient combustion, 
can be used to estimate BC emissions based on known CO emissions. Data (Figure 
3.9) from the afternoon hours (2-4 pm LST) are used to avoid sampling gaps at night 
and to get a more regionally representative ratio. The mass concentration of BC is 
calculated using PSAP aerosol absorption data, assuming that BC is the dominant 
absorbing compound at 550 nm. The specific absorption coefficient (10 m2/g) 
determined by Barnard et al. [2005] is adopted for the calculation. The mass 
 
 75 
 
concentration of CO is computed with locally measured ambient pressure and 
temperature. 
 
 
Figure 3.9. Correlation between 10-min average BC and CO during the afternoon 
hours (2-4 pm LST) at the Xianghe site. Data points with relatively low BC/CO ratio 
probably reflect emissions from agricultural burnings (spotted during the experiment) 
during the smoldering stage. 
 
 The BC/CO ratio from this study is 8.05 ? 10-3 (?g/m3 BC vs. ?g/m3 CO), 
lower than observed over the Indian Ocean during the INDOEX experiment (12.5 ? 
10-3, [Dickerson et al., 2002]) but higher than that for the eastern U.S. (3.4 ? 10-3, 
[Chen et al., 2001]). Despite the uncertainties in sampling and measurements, the 
different ratios likely reflect differences in energy consumption of these regions. 
Among the three regions, India probably has the most intense emissions from 
biomass burning, followed by China. The inefficient combustion of other fuels and 
high-emission vehicle fleets in India and China could also contribute to the high 
BC/CO ratios observed. Using the ratio in Figure 3.9 and the CO emissions estimated 
 
 76 
 
by Streets et al. [2006], BC emitted from China in 2001 is estimated at ~1300 Gg/yr, 
in agreement with the inventory (1049 Gg/yr, [Streets et al., 2003]). If as suggested in 
previous sections, the actual CO emissions could be twice as much as estimated, the 
resulted BC emissions could be up to 2600 Gg/yr. The relative uncertainty of this 
estimate is large (at least a factor of two) and ill-constrained, due to sizeable 
uncertainties in BC measurements, CO inventories, and spatial and temporal coverage 
of the experiment (BC/CO ratio from a single location used for a large region and 
assumed constant in all seasons). Nonetheless, bottom-up estimates and observational 
results are in good agreement in this study, compared to a similar study for India 
[Dickerson et al., 2002]. 
 
3.3 Enrichment of Elements and Ions in Aerosols 
3.3.1 Enrichment factors and enrichment of lead 
 Using Al as the reference element, the enrichment factor of an element X in 
aerosols can be calculated with the following equation: 
CrustAerosol
X Al
X
Al
XEF ?
?
??
?
??
?
??
?
?=   (3.1) 
where 
AerosolAl
X ?
?
??
?
?  is the mass ratio between X and Al in aerosols, while 
CrustAl
X ?
?
??
?
? is 
the same ratio in the crustal material, taken from Lide [1998]. The concentrations of 
elements in different types of soil vary, but normally by less than a factor of 5 [e.g., 
Chen et al., 1991]. If EFx is close to one, X in aerosols mainly comes from soil 
sources. On the other hand if EFx is much higher than unity, X in aerosols may have 
 
 77 
 
important sources other than soil particles. Elements with high EF have been used as 
tracers of anthropogenic aerosols [e.g., Sun et al., 2005]. 
 Shown in Figure 3.10 are EF?s of different elements in aerosol samples from 
Xianghe. Elements abundant in the Earth?s crust like Ca, Fe, and Mg generally have 
EF below 5, and are mostly from soil sources, although as discussed in section 3.2, 
substantial sources of Ca and Fe are related to human activities. Compared to other 
crustal elements, the EF of K is more variable, probably due to emissions from 
biomass burning. The EF of V is close to one, showing little evidence of oil 
combustion [Rahn, 1981]. As, Ag, Pb, and Cd have EF?s above 100, and are largely 
from anthropogenic pollution sources. For example, As may be from coal burning, 
incinerators, and smelters; Cd may be emitted from incinerators, smelters, and during 
manufacture of pigments [Pacyna, 1998]. EF?s of pollution elements are generally 
more variable than crustal elements, due to large synoptic fluctuations in pollutant 
levels discussed earlier. 
 Lead (Pb) is highly enriched, with an EF over 100. Our average Pb level (0.28 
?g/m3) is comparable to the loadings near busy roads of downtown Beijing (0.24-0.31 
?g/m3, spring 2002-2004) [Sun et al., 2006] and higher than the spring average 
concentration of 0.0036 ?g/m3 found in Fort Meade, MD [Chen et al., 2002]. During 
the 1970s leaded gasoline was the major source of Pb in particulate matter in the U.S., 
although point sources such as incinerators, battery manufacturing, and glass 
manufacturing also emit Pb [e.g., Pekney et al., 2006]. In recent years after the 
phasing out of leaded gasoline, resuspension of lead-containing particles (due to 
deposition of previously emitted Pb) has been suggested to be an important source of 
 
 78 
 
lead in aerosols in the U.S. [e.g., Ehrman et al., 1992; Harris and Davidson, 2005]. 
Leaded gasoline was phased out in China in 2000 [Chinese EPA, 1999]. But the good 
correlation between the EF of Pb and CO in our data (Figure 3.11) implies that 
substantial anthropogenic sources may still exist in the region. The concentration of 
Pb is correlated with CO (R2 > 0.5, and higher under stagnant conditions) but not with 
Al (tracer of soil particles). Our site is ~500 m away from the nearest major road with 
limited traffic. Traffic in Chinese cities grows fast recently, but in the mid 1990s the 
total lead added to gasoline in China was only ~900 tonne/yr, about 1/3 of that in 
Australia [Thomas, 1995]. The mean concentration of Pb in soil in parks in Beijing 
[Chen et al., 2005] is 66.2 ?g/m3, only a factor of 3-4 higher than the background 
crustal concentration. Therefore resuspension of Pb-enriched soil particles may 
merely be a minor source of Pb in this study. Pb can poison the catalytic converter of 
cars [Thomas, 1995] resulting in higher CO emissions; old cars used to run on leaded 
gasoline may emit both CO and Pb in substantial amounts, and this might explain the 
observed correlation between CO and Pb. Considering that the phasing out of 
gasoline took a long period in the U.S. [Nadim et al., 2001], it is also possible that 
leaded gasoline may still be in use in China. Regardless of its origin, the high 
concentrations of Pb in China exceed the new U.S. EPA standard [U.S. EPA, 2006] 
and require more detailed analysis. 
 
 79 
 
 
Figure 3.10. Box and whisker chart of the enrichment factors. The center horizontal 
line, lower edge, and upper edge of the box represent the median, the 25th and 75th 
percentiles, respectively. Lower and upper whiskers are the 10th and 90th percentiles. 
Crosses stand for the 1st (lower) and 99th (upper) percentiles. Maximum (upper) and 
minimum (lower) EF?s are marked with horizontal lines outside the box body. 
 
 
Figure 3.11. Enrichment factor of Pb (red solid line) and hourly averaged CO 
concentration (black dots). 
 
 80 
 
3.3.2 Enrichment of chloride ion 
 Figure 3.12 is a scatter plot of the concentrations of Na+ (0.65 ? 0.38 ?g/m3) 
and Cl- (3.4 ? 2.4 ?g/m3) ions in aerosol samples, as a function of ambient SO2. It 
appears that the samples could be divided into two groups, one with high SO2 
concentration (>10 ppb) and enriched in Cl- with respect to sea water, and the other 
with low SO2 (<10 ppb) and a Cl-/Na+ ratio close to that of sea water (blue line in 
Figure 3.12, slope ~1.8, [Prospero et al., 2003]), although the two ions are only 
weakly correlated in both groups. Coal burning, the dominant energy source in China, 
emits both SO2 [Streets et al., 2003] and HCl. The latter could enrich Cl- in aerosols 
through interactions with soil particles [Wang, Zhuang et al., 2005]. The Cl/S mass 
ratio in coal used in six power plants in China ranges from 3.5?10-2 to 0.12, but only 
a fraction of Cl is emitted during combustion [Chen et al., 2007]. Cl--enriched 
aerosols could be a potential chlorine source for atmospheric chemical processes in 
the region.   
 
 81 
 
 
Figure 3.12. Scatter plot of Cl- and Na+ concentrations. The color of the dots 
represents the average SO2 for the corresponding samples. Blue (slope = 1.8) and red 
(slope = 5) solid lines are for reference only. 
 
3.4 Characterize Emissions from a Small Coal-Fired Boiler 
3.4.1 The boiler 
 The boiler studied (Figure 3.13), as mentioned in Chapter 2, is located ~50 m 
east of the major facility building where all trace gas instruments were deployed 
during the Xianghe experiment. This hot water boiler made in 1996 has a rating of 0.7 
MW and a working pressure of 0.2 MPa. At its full capacity, the boiler burns 220 kg 
coal every hour; during the experiment, the actual coal consumption was about 100 
kg/hr. This may influence the emissions (e.g., less CO) from the boiler due to possible 
enhancement in the air-fuel ratio. But the boiler is equipped with induced-draft and 
forced-draft fans, so lean combustion conditions are probably achieved even when the 
boiler runs at its maximum capacity. Without detailed information about the 
 
 82 
 
configuration and conditions of the boiler, I will not further discuss this issue in this 
dissertation. A cyclone separator with unknown efficiency removes particles in the 
flue gas, which is emitted through a 15-m chimney. There is no control on emissions 
of trace gases. In winter, it provides heating for the whole Xianghe observatory, 
roughly 3000 m2 in four buildings. The boiler normally runs 2-3 times in the morning 
and evening a day, each time for 1-3 hr, depending on outdoor temperature. A 
detailed running record was kept during the experiment. No element analysis was 
performed for the raw coal in chunks (0-50 mm in size) used; this type of coal from 
Datong, China is usually bituminous, consisting of 70-90% carbon and 0.5-3% sulfur 
[Chen et al., 2006, Ge et al., 2001]. I picked out the plumes from the boiler by 
looking for pollutant peaks that were observed when the boiler was running under 
weak east winds (wind speed < 2 m/s), and when maximum SO2 exceeded 80 ppb. In 
total, six such events were recorded in the 4-week experiment. 
 
Figure 3.13. The coal-fired boiler at the Xianghe site. 
 
 83 
 
3.4.2 Method 
 CO2 was measured with a Thermal Environment Instruments, Model 41C 
analyzer (Franklin, MA) operated by IAP, CAS. As an example, Figure 3.14 shows a 
plume from the boiler observed near midnight, March 14, 2005. Peaks of CO2 (450 
ppm), NOy (90 ppb), SO2 (500 ppb), and CO (4500 ppb) appeared almost 
simultaneously as the boiler started running at 9:35 pm LST. I derive the 
?background? through linearly interpolating the concentrations before and after the 
peaks (shown as red lines in the plot), assuming that the background pollutant levels 
(controlled by all other local emission sources) change slowly and linearly in the 
short period, and that the peaks appear as fresh boiler emissions are superimposed on 
the background. ?CO2, ?CO, ?SO2, and ?NOy are calculated by subtracting 
background from the peaks. CO2 is relatively smooth, as the instrument has a longer 
averaging time (5 min) than the others (1 min). Because the CO peak lasted for a 
shorter period of time than other trace gases (~15 min into the boiler run, or 9:50 pm 
LST), I define the first 15 min of each boiler run as a ?cold start?, marked by the area 
between two vertical dashed lines in Figure 3.14. As the boiler just starts running, 
unstable or less-than-optimized combustion conditions may result in emission 
characteristics different from normal combustion conditions.  
 
 84 
 
 
Figure 3.14. One plume from the boiler detected on March 14, 2005. Red lines are 
baselines derived through linear interpolation of measurements before and after the 
peak. The ?cold start? period of the boiler run (from the starting time to 15 min into 
the run) is marked with two vertical dashed lines. 
 
3.4.3 Emission characteristics of the boiler and implications 
 Results of correlation analysis for data of all six boiler plumes and the subset 
of cold-start only data are given in Figures 3.15 and 3.16, respectively. 1-min data are 
used for the correlation of ?CO/?SO2 and ?NOy/?SO2; 5-min data are used for CO2. 
Too few simultaneous NOy and CO2 data points are available for a meaningful 
correlation analysis.   
 
 
 85 
 
 
Figure 3.15. Correlations of ?CO, ?SO2, ?NOy, and ?CO2 for all plume data. Dotted 
lines mark the 95% confidence interval for the correlation analysis. Note the slopes 
for ?SO2/?CO2 and ?CO/?CO2 are in ppbv/ppmv. 
 
 86 
 
 
Figure 3.16. Same as Figure 3.15, but for cold start data. 
 
 ?SO2 and ?CO2 are well-correlated, with R2 of 0.88 for all plume data and 
0.91 for cold start data. The ?SO2/?CO2 ratio does not change much in the two 
figures (3.5 (ppb/ppm) vs. 3.8), suggesting that SO2 emissions from the boiler are not 
affected by burning conditions. The corresponding ratio in the emission inventory 
ranges from 1.1 to 20 (ppb/ppm) [Streets et al., 2003]. In comparison, ?CO and 
?CO2 are only weakly correlated in all plume data (Figure 3.15), probably due to 
dependence of CO emissions on combustion conditions. For the few cold start data 
points, ?CO and ?CO2 are not significantly correlated. Note that the inventory 
[Streets et al., 2003] CO/CO2 emission ratio for domestic coal burning in China is 
much higher at about 60 (ppb/ppm), compared to the ratio from this study (17 
(ppb/ppm)). In 1-min data, ?CO and ?SO2 are correlated in both all plume data and 
 
 87 
 
cold start data, as are ?NOy and ?SO2. Comparing Figures 3.15 and 3.16, one can see 
that the cold start data have a greater ?CO/?SO2 ratio (7.4 vs. 5.0) but slightly 
smaller ?NOy/?SO2 ratio (0.16 vs. 0.18) than all plume data. As SO2 emissions vary 
little under different burning conditions, the change in ?CO/?SO2 ratio suggests that 
less-than-optimal conditions during the cold start stage may cause stronger CO 
emissions. NOy emissions, on the other hand, remain about the same. 
 Assuming carbon makes up 80% of the mass of the coal and is all converted 
to CO2 in combustion, burning 1 kg coal will emit 2.93 kg CO2. From the 
?SO2/?CO2 ratio, I estimate that roughly 15.4 g SO2 is emitted at the same time. If all 
the sulfur were released as SO2, the sulfur content of the coal used in this boiler 
would be about 0.8%. Different assumptions concerning the carbon content of the 
coal and the percentage of sulfur released as SO2 may lead to slightly different 
results. The overall relative uncertainty is about 30%; the SO2 emission factor is 15.4 
? 4.6 g SO2/kg coal (hereafter the units for emission factors are presented as g 
(species)/kg (fuel) unless otherwise noted), and the sulfur content of the coal is 0.8% 
? 0.24%. The SO2 emission factor from this analysis is lower than the value (46.7 
g/kg) given by Ge et al. [2001], who used raw coal with sulfur content of 2.81%, but 
within the range of the factors used in emission inventories (3.29-54.8 g/kg for the 
domestic sector, [Streets et al., 2003]). The average sulfur content of Chinese coal is 
1.1-1.2% by weight [He and Chen, 2002]. According to the Beijing EPA?s data, 
replacing small coal boilers in downtown Beijing has reduced coal usage by about 6 
million tonne, and decreased SO2 emissions by 56,000 tonne. This translates into an 
average SO2 emission factor of about 9.3 g/kg for similar small coal-burning emission 
 
 88 
 
sources in Beijing.  If I assume, according to Streets et al. [2003], that the emission 
factor of CO2 for domestic coal combustion in China is about 2 kg CO2/kg coal, the 
SO2 emission factor would be 10.5 g/kg. 
 Estimating the CO emission factor from the ?CO/?CO2 ratio is subject to 
greater uncertainty, as CO emissions are more variable, depending on combustion 
conditions. The CO emission factor is about 31.7 g CO/kg coal from this analysis, 
higher than the factor determined by Ge et al. [2001] for a larger automatic stoker (15 
g/kg), but much lower than the values used in two emission inventories (74 g/kg, 
[Streets et al., 2003]; EPA AP-42, 124 g/kg, [Streets et al., 2006]) for the domestic 
sector of China. Emission factors can also be derived with the correlations of 
?CO/?SO2 and ?NOy/?SO2. The CO emission factor calculated this way is 33.7 ? 
13.5 g/kg for all plume data (49.6 ? 19.8 g/kg for cold start), consistent with the result 
from the ?CO/?CO2 ratio. Using the CO2 emission factor from Streets et al. [2003] 
would give an even lower CO emission factor (23.6 g/kg for all plume, 34.7 g/kg for 
cold start). The NOy emission factor is 2.0 ? 0.8 g/kg for all plume data and 1.8 ? 0.7 
g/kg for cold start (assuming all NO2), on the higher end of the range employed by 
Streets et al. [2003] (1.19-2.24 g/kg). 
 From March 2 to March 22, 2005, the boiler completed 66 runs for a total run 
time of about 88 hr (accumulated cold-start run time: 16.5 hr). The overall coal 
consumption was about 8800 kg, emitting roughly 135 kg SO2, 300 kg CO, and 18 kg 
NOy (assuming all NO2). The heating season in this area usually lasts from mid-
November through mid-March the following year. The actual running time of a 
heating unit depends on weather. A reasonable guess is that the boiler run time in 
 
 89 
 
November is the same as in March, and December, January, and February each has 
twice as much run time. The total boiler running time would sum to about 350 hr for 
the entire heating season. Coal consumption of the season adds up to 35000 kg, 
emitting about 540 kg SO2, 1200 kg CO, and 72 kg NOy. Streets et al. [2006] 
estimated that in China, domestic boilers burnt about 80 million tonne of coal in 
2001. If 25% or 20 million tonne were used by hand-fed boilers for heating, 
emissions from these small units in 2001 would include 308000 tonne SO2, 674000 
tonne CO, and 40000 tonne NOy, applying the emission factors from this study. This 
is but a small fraction of the total emissions from China (1.5% of SO2, 0.4% of CO, 
and 0.3% of NOy, [Streets et al., 2003, 2006]). But these small boilers are often 
located in densely populated communities (according to the author?s experience 
living in the area); pollutants emitted to limited height by them are often trapped near 
ground as they run on cold winter nights. These factors may lead to heavy local 
pollution at times as observed in this study, and may cause severe adverse health 
effects. 
 
3.5 Summary 
Measurements of CO, SO2, NOy, O3, aerosol scattering, aerosol absorption, 
and aerosol chemical composition have been made in Xianghe, a rural site about 70 
km east-southeast and generally downwind of Beijing. Located in a rural area 
undergoing fast industrialization and in proximity to two mega cities, the Xianghe site 
allows us to look into China?s air pollution issue from close distance. 
 
 90 
 
During the experiment in March 2005, air pollution at the Xianghe site was 
characterized by high pollutant levels and fast transitions between clean and heavily 
polluted periods. Cold front passages, repeating every 4-5 days, are the major 
mechanism driving these transitions, as they are associated with strong vertical 
motions and wind shear accompanying the change of air mass. The observed response 
of air quality to weather is similar to what has been found in the eastern U.S. What is 
different is that northern China is much drier in spring, and wet deposition is 
inefficient in removing pollutants at this time of the year. Some pollutants are 
probably lofted away from the surface during cold front passages, but pollutant 
transport and evolution afterwards need further investigation. The next chapter will 
analyze aircraft-measured pollutant profiles under a variety of weather conditions, 
and discuss the role of mid-latitude wave cyclones in pollutant lofting and transport. 
Synoptic variation of chemical processes involved in the production of secondary 
aerosols (sulfate and nitrate) is another interesting phenomenon observed in Xianghe, 
and will be discussed in more detail in Chapter 5.  
The observed diurnal cycle of pollutants discloses the diverse nature of the 
emission sources near the Xianghe site. Pollutant levels are more variable in the 
normally stagnant morning, (7-9 am LST), as the instruments detect more ?local? and 
fresh pollutants. Under strong winds in the afternoon hours (2-4 pm LST), the 
instruments see more ?regional? and well-mixed aged pollutants. Low single 
scattering albedo (morning: 0.81, afternoon: 0.85) suggests that aerosols near the 
surface are light absorbing, as seen in previous experiments in eastern China [Bergin 
et al., 2001; Xu et al., 2002; Andreae et al., 2005]. The ground SSA is much lower 
 
 91 
 
than the column integrated value (0.9) [Z. Li et al, 2007a]. Stronger aerosol 
absorption at lower altitudes could have impacts on the local radiation budget as well 
as atmospheric stability. Further work is needed to quantitatively address these 
effects. 
Using the CO emission data from Streets et al. [2006] and the BC/CO ratio 
from this experiment, black carbon emitted from China is estimated at 1300-2600 
Gg/yr. This is in agreement with bottom-up estimates [Streets et al., 2003], but both 
have substantial uncertainties. Enrichment factor analysis confirms that Al, Ca, Fe, 
Mg, and K are mainly from soil sources, although Ca, Fe, and K might also be related 
to anthropogenic activities in the region such as construction (Ca) and biomass 
burning (K). Pb is still highly enriched several years after leaded gasoline was 
supposed to be phased out in China, and likely comes from human sources as shown 
by the good correlation between Pb and CO. The enrichment of Cl- ion with respect to 
sea water is correlated with ambient SO2 level, probably a result of HCl released from 
extensive coal burning in the area.  
Afternoon gas concentration ratios (CO/SO2, CO/NOx (NOy)) were compared 
to inventories. With a 36% increase in estimated CO emissions from China, the 
updated CO inventory by Streets et al. [2006] agrees better with our observations than 
earlier estimates [Streets et al., 2003; X. Wang et al., 2005]; but a difference of a 
factor of two in CO/SO2 ratios still exists. The CO/NOx ratio derived from this 
experiment is close to Streets et al. [2003; 2006], but a factor of four higher than X. 
Wang et al. [2005]. A number of issues could contribute to the difference between 
observations and inventories, including but not limited to temporal and spatial 
 
 92 
 
coverage of measurements, release heights of pollutants, chemical and physical losses 
of pollutants, and inventory errors. But for air from the populated area to the W and 
SW, the observation-derived CO/SO2 ratios are consistently more than 100% higher 
than the inventory values [Streets et al., 2003, 2006]. Even larger gaps between 
observations and inventories are found for plumes from downtown Beijing, where in-
situ measurements also record a CO/SO2 ratio much higher than inventories [Wen et 
al., 2008]. It is also worth mentioning that CO/SOx (~27) and CO/NOy (~55) ratios 
from earlier measurements downwind of the Beijing area [Carmichael et al., 2003] 
indicated a possible low bias in inventory CO emissions for the region as well. 
Neither chemical conversion of SO2 nor different release heights of SO2 and CO 
would explain the revealed discrepancy. While quantitatively evaluating emissions 
for a vast country like China is difficult with measurements from a single location, it 
is plausible that CO emitted from China could still be underestimated (or SO2 might 
be overestimated or SO2 emissions might have decreased in recent years) [Streets et 
al., 2006], since the inventories have difficulty matching the observed CO/SO2 ratios 
for more polluted areas. The inventory ratios of CO/SO2 are generally low (around 10) 
in some biggest cities of China [Streets et al., 2003, 2006]; emissions from mobile 
sources, biofuel, and small coal-fired boilers in China need to be further measured to 
reduce the uncertainties [Streets et al., 2006].  
Our conclusion that CO emissions from China might be underestimated, to a 
large extent, relies on the SO2 inventory, and it is worthwhile to discuss the recent 
trend of emissions from China. Since 2001, a substantial fraction of coal-fired power 
plants in China have been equipped with flue-gas desulfurization (FGD) devices. 
 
 93 
 
From 2001 to 2006, the penetration of the FGD technique increased from 3% to 30%, 
reducing the average SO2 emission factor for Chinese coal-fired power plants by 15% 
[Q. Zhang, personal communication]. Despite this improvement in the emission 
controls, SO2 emissions from China during the same period increased by ~30-50% 
[Chinese CBS, 2006; Q. Zhang, personal communication]. The uncertainty of SO2 
emission inventory for China is relatively small (within 20%, [Streets et al., 2003; Q. 
Zhang, personal communication]). Therefore it is unlikely that SO2 is overestimated 
or has decreased (in fact SO2 emissions from China in 2005 are likely greater than in 
2000, the base year of Streets et al. [2003]); and it is very plausible that CO emissions 
from China in 2005 are much higher than in 2000 and 2001 (base years of Streets et 
al. [2003, 2006]). Indeed a new emission inventory that is just becoming available [Q. 
Zhang et al., A new inventory of anthropogenic emissions in Asia for the year 2006, 
Atmos. Chem. Phys. Discuss., manuscript in preparation, 2008] estimates that CO 
(166889 Gg/yr), SO2 (31020 Gg/yr), and NOx (20830 Gg/yr) emissions from China in 
2006 are all higher than in 2000 and 2001. The national inventory ratios of CO/SO2 
and CO/NOx for China based on this new inventory are 12.3 and 12.6, respectively, 
still lower than our measurement-derived values. The CO/SO2 and CO/NOx inventory 
ratios for Beijing and Hebei (the province with strong emissions to the south and 
southwest of Xianghe) combined are 16.4 and 18.2, respectively. These are higher 
than the national values, but lower than our observations. It appears that although this 
updated emission inventory has attempted to account for some recent changes, it may 
still underestimate CO emissions from China. 
 
 94 
 
Trace gas emissions from a small coal-fired boiler at the Xianghe site were 
characterized with ambient data. Our estimated SO2 (15 g/kg) and NOy (2 g/kg) 
emission factors are in line with inventories. But the CO emission factor (31 g/kg all 
data, 50 g/kg cold start) is only one third of the values used in inventories [Streets et 
al., 2003, 2006]. If the boiler in this study is representative, similar small boilers in 
China may be more efficient than previously thought, and would not account for the 
possible low bias in inventory CO. However, this conclusion needs to be verified by 
measuring emissions from more boilers under different running conditions. Mobile 
sources, on the other hand, are becoming more important in Chinese cities, and 
experiments characterizing their emissions are called for. Note that small coal boilers 
may have severe local environmental effects, as they are often in residential areas, run 
under stagnant weather conditions, and have low stacks with insufficient emission 
control measures. 
 In summary, in this chapter I have presented results from the Xianghe ground 
experiment. I evaluated some current emission inventories of China, and found 
substantial uncertainties in inventory CO emissions, which are likely biased low. 
Examining the plumes from a local coal-fired boiler suggested that emissions of CO 
from similar boilers in China are less intense than earlier estimates, and that 
emissions of mobile sources and biofuel need further studies. The observed day-to-
day variations in air quality are closely related to frontal systems. The interaction 
between atmospheric chemistry and synoptic systems will be the focus of later 
chapters. 
 
 95 
 
Chapter 4: Shenyang Aircraft Campaign 
 
4.1 Overview of the Aircraft Campaign 
 Many of the results presented in this chapter have been published in the 
Journal of Geophysical Research special section on EAST-AIRE [Dickerson et al., 
2007]. Additional information on the meteorology and flights as well as day-to-day 
variation of air quality during the aircraft campaign is given in this section. 
 
4.1.1 Weather 
 Weather in Shenyang area during the aircraft campaign was characterized by 
frequent passages of mid-latitude wave cyclones. During the 12-day experiment from 
April 1 to 12, 2005, four low pressure systems visited the region, two of which 
brought precipitation (April 6 and 9) including thunderstorms on April 6. 
Accompanying these low pressure centers and the associated fronts, as discussed in 
Chapter 3, were changes in pollutant levels and meteorological conditions such as 
wind speed, wind direction, temperature and RH. Ahead of cold fronts, winds were 
usually from the relatively warm and moist south; behind cold fronts, winds from the 
cold and dry north were more common. Figure 4.1 and Table 4.1 summarize weather 
conditions recorded during the experiment. Daily surface weather maps at 00 UTC (8 
am LST) over East Asia, generated by the Korean Meteorological Administration 
(KMA), are given in Appendix I.  
 
 96 
 
 
Figure 4.1. Meteorological records from the airport near Shenyang where the Y-12 
research aircraft was deployed during the experiment. Blue lines with arrows roughly 
mark the time of flights. Red and black arrows in the top panel indicate rain and 
?blowing dust? observed by the meteorologist at the airport, respectively. 
 
 
 
 
 97 
 
 
Table 4.1. Summary of weather conditions during the Shenyang aircraft campaign. 
Date Weather Conditions 
April 1, 2005 
(DOY 91) 
A high over the region, a weak low to the SW but moving E, 
variable wind direction, low surface wind speed (< 1 m/s). 
April 2, 2005 
(DOY 92) 
A low E of Shenyang over Korea, a high controls the region. Strong 
N winds in daytime (surface wind speed up to 10 m/s). 
April 3, 2005 
(DOY 93) 
High over the region but moving E. A strong low sits NNW of the 
region, with its trough W of Shenyang near China-Mongolia border. 
Variable winds. 
April 4, 2005 
(DOY 94) 
Two lows (N, and WNW of Shenyang) with a ridge sitting in 
between. Weak surface winds in daytime generally from S. 
April 5, 2005 
(DOY 95) 
A strong low with its center to the NW, and its trough extending 
southward, W of Shenyang. Strong S winds up to 8 m/s. Local dust 
emissions and very low visibility. 
April 6, 2005 
(DOY 96) 
The strong low passed the region on the day, bringing rain in the 
morning and strong thunderstorms in the afternoon. 
April 7, 2005 
(DOY 97) 
The low moved to E of Shenyang; its cold front was over the Sea of 
Japan. Strong surface winds from NW (up to 6 m/s). 
April 8, 2005 
(DOY 98) 
A high over the region. Winds mostly from NE up to 4 m/s. A weak 
low was NW of the region, just S of Lake Baikal. 
April 9, 2005 
(DOY 99) 
The low on the previous day moved SE to the China-Mongolia 
border NW of Shenyang region. Another low about 1000 km to the 
S, near Shanghai. A high W of the two lows. Winds variable, with 
changing from generally N to generally S. Wind speed 2-5 m/s. 
Light rain during the day. 
April 10, 2005 
(DOY 100) 
The low E of the region, over the Sea of Japan. High pressure 
system over the area. Winds from N during the day, with surface 
wind speed up to 6 m/s. 
April 11, 2005 
(DOY 101) 
A weak low with its trough W of Shenyang. Variable winds from S, 
E, and W, surface wind speed no more than 2 m/s. 
April 12, 2005 
(DOY 102) 
A high to the W dominated the region. Surface winds as strong as  7 
m/s generally from N. 
 
4.1.2 Flight routes 
 Different flight routes were chosen under various weather conditions. Spiral 
ascents were made over the same locations but on different days, as mid-latitude lows 
passed the region, to measure prefrontal and postfrontal pollutant profiles. In these 
profiling flights (April 5, 7, 10, and 11), the aircraft went up in spirals from ~300 m 
 
 98 
 
to ~4000 m altitude (above sea level) over two locations roughly 100 km from 
Shenyang, one to the south and the other to the north. In two cloud seeding flights on 
April 7 and 9, the aircraft mostly flew through clouds at constant altitudes of about 
3000-4000 m. Additionally there was a transit flight from the service facility to the 
airport on April 1, and a survey flight on April 12. Table 4.2 gives information for 
each flight. Maps of flight routes are available from the EAST-AIRE website at: 
http://www.meto.umd.edu/~zli/EAST-AIRE/station.htm. Flight height information 
can be found in Figure 4.2. 
 
Table 4.2. Summary of the flights during the Shenyang aircraft campaign. 
Date Takeoff 
(LST) 
Landing 
(LST) 
Flight Description 
April 1, 
2005  
13:45 15:51 Transit flight from the service facility in Harbin 
(45.71?N, 126.65?E) to Shenyang airport (41.64?N, 
123.49?E). Flight altitude about 3000 m. 
April 5, 
2005 
10:36 13:40 Research flight: one spiral of 300-4000 m at 11:30 
over Liaozhong, S of Shenyang (41.37 ?N 122.61 
?E), and one spiral of 400-4000 m at 13:00 over 
Xiaoming, N of Shenyang (42.48?N 123.61?E) 
April 6, 
2005 
07:28 09:34 Cloud seeding flight in the triangular area between 
Haicheng (40.85?N, 122.70?E), Benxi (41.30?N, 
123.83?E), and Tieling (42.28?N, 123.83?E). 
Flight altitude about 3000 m. 
April 7, 
2005 
14:43 18:13 Research flight: one spiral of 400-4000 m at 15:40 
over Xiaoming, and one spiral of 300-4000 m at 
17:10, over Liaozhong. 
April 9, 
2005 
04:17 07:47 Cloud seeding flight E of Shenyang. Flight altitude 
about 3700 m. 
April 10, 
2005 
10:05 13:30 Research flight: one spiral of 500-4000 m at 11:10 
over Xiaoming, and one spiral of 300-4000 m at 
12:30, over Liaozhong. 
April 11, 
2005 
09:53 14:04 Research flight: one spiral of 80-4000 m at 11:30 
over Liaozhong, and one spiral of 400-4000 m at 
13:00 over Xiaoming. 
April 12, 
2005 
11:00 12:54 Research flight surveying the region around 
Shenyang. Flight altitude about 1000-1500 m. 
 
 
 99 
 
4.1.3 Synoptic variation of air pollution during the aircraft campaign  
 As can be seen from Figure 4.2, air pollution over Shenyang region features 
strong day-to-day variation. Below 1000-1500 m altitude, or generally within the 
daytime PBL, flights on April 1, 5, and 11 found SO2 as much as 20 ppb near ground 
and up to 10 ppb around 1000 m, as stagnant conditions prevented dispersion (April 1 
and 11) or strong surface winds blew from the populated and industrialized region to 
the south of Shenyang (April 5). Meteorological records (Table 4.1) suggest that these 
flights were conducted with a low pressure system approaching the area from W, and 
in general represent prefrontal atmosphere. Flights on April 7, 10, and 12 were under 
fairly strong N winds and high pressure systems. These postfrontal flights recorded 
little SO2 near ground, mostly less than 5 ppb. Small legs of April 12 flight with high 
SO2 were probably influenced by large point sources the aircraft flew over. Small SO2 
concentrations measured over the airport on April 6 and 9 were likely related to wet 
removal processes on the two rainy days. Above the PBL, SO2 concentration is 
generally low (cf. section 4.2 for more detailed results). But substantial SO2 was 
measured at 2000-3000 m altitude on April 5. Propelled by strong westerly winds in 
the free troposphere (FT), the observed SO2 and other pollutants can potentially travel 
far downwind, and will be the major topic of discussion in section 4.2. Column 
integrated SO2 varied from as little as 0.1-0.2 Dobson Unit (DU) on April 7 and 10 to 
as much as 1-2  DU on April 1 and 5 [Krotkov et al., 2008], demonstrating a factor of 
10 change in just days. 
 
 100 
 
 
Figure 4.2. Flight altitudes and SO2 concentrations (in color, unit: ppb) along the 
flight tracks during the Shenyang aircraft campaign. Gaps in data are mainly due to 
instrument zero mode. 
 
 
 101 
 
 Aerosol scattering data (Figure 4.3) show somewhat different day-to-day 
variability compared to SO2. High aerosol loadings on April 5 (DOY 95), as will be 
shown later in this chapter and in Chapter 5, are mainly due to local and transported 
dust. Relatively abundant aerosols at around 2000 m altitude on April 7 are likely 
dust particles carried into the area by NW winds from arid desert regions. I attribute 
exceptionally high scattering at 1000-3000 m altitude on April 10 to possible cloud 
interference as the aircraft flew through a deck of fair weather cumulus. Slightly high 
aerosol levels near ground during the April 11 flight probably reflect the 
accumulation of anthropogenic pollutants under stagnant weather conditions. Aerosol 
optical thickness (AOT) at 500 nm, a measure of the column abundance of aerosol 
particles, observed from the surface using a hand-held (black squares, Figure 4.4, [Xin 
et al., 2007]) and a Cimel sun photometers (red dots, Figure 4.4, [Xia et al., 2007]) in 
the area also was also very variable during the aircraft campaign, ranging from 0.1-
0.2 to above 1.0. The difference between the two sun photometers probably originates 
from different calibrations. 
 
 
 102 
 
 
Figure 4.3. Flight altitudes and aerosol scattering at 550 nm (color, unit: m-1) along 
the flight tracks during the Shenyang aircraft campaign. 
 
 103 
 
 
Figure 4.4. Hand-held (black squares) and Cimel (red dots) sun photometer measured 
aerosol optical thickness at 500 nm over Shenyang area, from April 1 to 15, 2005. 
 
4.1.4 Vertical distribution of pollutants: comparison to the Northeastern 
U.S. 
 Here I compare the results from the Shenyang aircraft campaign to the 
chemical climatology amassed over the past decade through hundreds of flights over 
the northeastern U.S., with essentially the same bundle of instruments [Hains et al., 
2008; Prados et al., 1999; Taubman et al., 2006, 2007]. Detailed analyses on vertical 
profiles of pollutants from flights on April 5 and 7 are given in section 4.2. Vertical 
profiles from the other research flights (April 10, 11, and 12) are presented in 
Appendix II. Our data set from a single season in China is much smaller, but shows 
that concentrations of trace gases and aerosols over China are markedly greater than 
are typically measured over N America (Figure 4.5). Both data sets show that SO2 is 
 
 104 
 
more abundant in the lower PBL than at higher altitudes, but SO2 in the lowest 500 m 
over the U.S. [Taubman et al., 2006] is only about one third of that over China. The 
boundary layer in summer is deeper, and SO2 over the U.S. around 1000 m altitude, 
measured mostly in summer, is comparable to the spring results over China from this 
study. But above the PBL, the mean SO2 concentration over China is again sizably 
larger. For example, at 2500 m altitude over the U.S., the mean mixing ratio of SO2 is 
0.2 ppb (90th percentile: 0.6 ppb) and over China the mean is greater at 0.6 ppb (90th 
percentile: 2.3 ppb). The aerosol scattering over China in spring is greater than that 
under smoggy summer conditions over N America, and an order of magnitude larger 
than over the U.S. in spring. At 2500 m, the average bsp over N America in summer is 
about 50 Mm?1 (10-6 m-1); the average over China is about 250 Mm?1. These results 
underscore the potential large-scale impacts emissions from China may have, but 
further flights are required to build a comprehensive chemical climatology over the 
region. Comparison of CO profiles is given in Figure 4.11, section 4.2. 
 Another interesting comparison that can be made between the U.S. and China 
data is about the ?ngstr?m exponents (AE), determined with the same nephelometer 
(Chapter 2) and sometimes used as a proxy of the overall particle size (Figure 4.6). 
The ?ngstr?m exponents over China are smaller than over the eastern U.S. at all 
levels. With exceptions at few altitudes, the 75th percentiles of China data are smaller 
or comparable to the 25th percentiles of the U.S. data. This suggests that using the 
same instrument, the overall size of aerosols sampled over N China is greater than 
over the NE U.S. Note there could be substantial sampling line losses of larger 
particles in both data sets, but particularly so over the dust-laden N China. Surface 
 
 105 
 
measurements at three AERONET sites generally in the flight regions of China 
(Liaozhong and Xianghe) and the U.S. (Goddard Space Flight Center, GSFC, 
Greenbelt, MD), show similar differences in ?ngstr?m Exponents (Table 4.3). This 
comparison corroborates the importance of dust in aerosol loadings over N China in 
the dry spring months, as discussed in Chapter 3 and earlier in this chapter.  
 
Figure 4.5. a) Mean vertical profiles of SO2 for flights ahead of the front (dashed 
line) and all eight flights during the Shenyang experiment (solid line). For 
comparison, the mean SO2 profile measured on smoggy summer days over the NE 
U.S. [Taubman et al., 2006] is shown as a dotted line. b) Mean vertical profiles of 
aerosol scattering (bsp) at 550 nm ahead of the front (dashed line) and for all eight 
flights conducted during the Shenyang experiment (solid line). For comparison, the 
mean bsp profile measured over the NE U.S. in spring 1996 is shown as a dot-dash 
 
 106 
 
line and the mean for smoggy summer days is shown as a dotted line [Taubman et al., 
2006, 2007]. 
 
 
 
Figure 4.6. Vertical profiles of ?ngstr?m Exponents observed from aircraft flights 
over N China in spring and over the NE U.S. in summer with the same instrument. 
Note that overall the aerosols over China are larger and more variable in size. The 
aircraft measurements mainly represent fine particles due to line losses of coarse 
particles. 
 
Table 4.3. Statistics of ?ngstr?m exponents (440-675 nm) in 2005 at three 
AERONET sites in China and the U.S. 
Site Period (DOY) 25
th 
Percentile Median 
75th 
Percentile Mean 
Standard 
deviation 
GSFCa 1.39 1.75 1.86 1.62 0.30 
Liaozhongb 0.43 0.82 1.20 0.81 0.42 
Xianghec 
97-120 
0.60 0.86 0.97 0.78 0.28 
GSFC 1.45 1.75 1.88 1.68 0.28 
Liaozhong 0.78 0.97 1.23 0.96 0.32 
Xianghe 
97-186 
0.82 1.06 1.22 1.00 0.30 
 
 107 
 
a GSFC: 38.99?N, 76.84?W, AOT (500 nm) = 0.27 ? 0.17 (mean ? std. dev.) during 
DOY 97-186 in 2005; 
b Liaozhong: 41.51?N, 122.70?E, AOT (500 nm) = 0.73 ? 0.62 (mean ? std. dev.) 
during DOY 97-186 in 2005; 
c Xianghe: 39.75?N, 116.96?E, AOT (500 nm) = 0.76 ? 0.63 (mean ? std. dev.) 
during DOY 97-186 in 2005. 
 
4.2 Upwind Dry Convection and Long-Range Transport: A Case Study 
4.2.1 Prefrontal flight on April 5, 2005 
 The surface weather map and flight route on April 5 are given in Figure 4.7. 
The weather conditions for this prefrontal flight involved low-level flow from the 
southwest, some of the highly populated and polluted areas of China. Winds veered 
with altitude becoming westerly in the mid troposphere. In the area of the flights and 
to the south, few clouds were observed on that day; the system brought little rain to 
the flight region until April 6, and by April 7 only cirrus clouds were spotted. Strong 
surface winds (cf. section 4.1) from the polluted warm sector of the low pressure 
system brought heavy loadings of pollutants and dust (Figures 4.8 - 4.11). Sizable 
concentrations of trace gases and aerosols were observed not only in the PBL, but 
also in the free troposphere (Figures 4.8 ? 4.11), where pollutants have relatively long 
lifetime, and stronger winds could potentially carry them far downwind, leading to 
possible large-scale impacts. 
 Aircraft measured temperature and RH profiles (Figure 4.8) indicate a strong 
inversion (7 ?C at 11 am LST) layer at about 1000 m altitude, marking the top of the 
PBL. The ozone mixing ratios given in the same plot reflect moderate photochemical 
smog formation, and show a peak just above the inversion. Aerosol scattering 
coefficients (bsp) also have a maximum at that altitude (Figure 4.9); for reference bsp 
 
 108 
 
of 4x10-4 m-1 corresponds to ~100 ?g/m3 particulate matter for a specific scattering 
coefficient of 4.2 m2/g (3 m2/g for dry ammonium sulfate and ammonium nitrate in 
the fine mode, 4 m2/g for particulate organic matter, [Boylan and Russell, 2006]). 
Multiphase reactions in the aerosol rich PBL may suppress ozone formation [e.g., 
Lary et al., 1997], while a layer of UV-scattering aerosols can inhibit photochemical 
production of ozone at low altitudes but accelerate it toward the top of the aerosol 
layer [Dickerson et al., 1997]. These processes may contribute to the increase in 
ozone from the surface to 1000 m altitude. Concentrations of aerosols and trace gases 
fell off sharply above the strong inversion, but remained well above background 
values. Substantial pollutants and possibly mineral dust persisted to the highest 
altitudes sampled, ~4000 m. As will be discussed in Chapter 5, the heavy loadings of 
pollutants and dust detected during this flight were widespread and impacted large 
areas in China, Japan, and northern Pacific as they traveled eastward in the next few 
days. 
 
 109 
 
 
Figure 4.7. Surface analysis (Korean Meteorological Administration) for 00 UTC 
April 5, 2005 showing a low-pressure system (cyclone) moving towards Shenyang. 
The approximate location of the flights is shown by the symbol of an aircraft and the 
actual flight pattern is shown in the insert. Profiling spirals were conducted north and 
south of Shenyang (locations marked with aircraft symbols). Note little cloud cover 
or precipitation was reported in the vicinity of the flight and several hundred km to 
the south ahead of the surface cold front. Yellow circles indicate occurrence of fog. 
 
 110 
 
 
Figure 4.8. Vertical profiles of ozone, RH, and temperature measured south of 
Shenyang at 11:30 LST on April 5, 2005. Note maximum in RH and ozone near the 
strong temperature inversion (7?C) around 1000 m altitude. In the lower free 
troposphere the air is dry and ozone concentrations are lower. See Appendix II for 
vertical profiles measured N of Shenyang on the same flight. 
 
 
Figure 4.9. Vertical profile of aerosol scattering coefficients, bsp, for the morning 
(11:30 LST) flight south of Shenyang on 5 April 2005.  
 
 111 
 
 Figure 4.10.Vertical profile of SO
2, for the morning (~11:30 LST) flight S of 
Shenyang on April 5, 2005. Note the high SO2 level in PBL. Above the inversion 
layer SO2 drops off quickly but remains almost 3 ppb between 2000 and 3000 m. The 
profile north of Shenyang (Appendix II) shows less but still around 1 ppb of SO2 at 
the same altitudes. 
 
 
Figure 4.11. Vertical distribution of carbon monoxide from the flight of April 5, 2005 
(squares). Mixing ratios over China drop off sharply above the inversion near 1000 
m, but remain well above the background. The median CO profile from ~400 flights 
made on smoggy summer days during 2000-2004 over the NE U.S. [Taubman et al., 
2006] (dotted line) and the median for spring flights ahead of fronts over the U.S. 
[Taubman et al., 2007] (solid line) are shown for reference. 
 
 112 
 
 To help identify the meteorological mechanism lifting the high concentrations 
of aerosols and trace gases above PBL, consider the backward trajectories (Figure 
4.12) calculated with the HYSPLIT (Hybrid Single-Particle Lagrangian Integrated 
Trajectory) model and NCEP global reanalysis meteorological data. Flow at low 
levels, in agreement with the surface weather analysis (Figure 4.7), was from the 
southwest while flow at higher altitudes originated from the west. Little evidence of 
upward motion is seen in the pre-frontal zone along the east coast of China; in fact, 
back trajectories only indicate some weak upward motion, suggesting that in this case 
the WCB was not effective in lofting pollutants over the continent. The back-
trajectory model and meteorological data used here cannot resolve small-scale 
vertical motions such as those caused by convective clouds. I examined GOES-9 
satellite infrared images for April 3?5 2005 for clouds, and confirmed that in the 
warm sector where the low-level air originated, the sky was generally clear. 
Convection ahead of the cold front (associated with clouds) does not appear to have 
been the mechanism responsible for pollutant lofting in this event.   
 
 113 
 
 
Figure 4.12. Backward trajectories calculated with the HYSPLIT model for air 
arriving at location of the southern spiral (41.4?N, 122.7?E, Figure 4.7) in the PBL 
(500 m above ground level) and in the FT (1500 and 3500 m) on April 5. The lower 
panel depicts vertical motions and shows little evidence of WCB lofting in the 
prefrontal zone for this episode. Back trajectories for the second spiral point, 200 km 
to the north and 2 hr later are similar. 
 
 Convective transport from west (upstream at higher altitudes) of the flight 
area likewise would not be captured by the backward trajectories, but deep clouds 
could lift pollutants and dust to high altitudes where they could travel in strong 
westerly winds to the Shenyang area. Satellite images (Figure 4.13) show strong 
convection over arid NW Mongolia (in the top left corner of the image). This part of 
the world is sparsely populated and unlikely the source of CO and SO2 observed in 
the flight, but could be the source of dust. A smaller convective system is seen along 
the back trajectory (Figure 4.13) corresponding to the origin of the air 27 hr prior to 
 
 114 
 
the flight. The MODIS IR cloud height analysis indicates that these cloud bands 
reached maximum heights of 6-12 km. They formed near the industrialized cities of 
Baotou and Hohhot with large steel mills, and could have detrained pollutants and 
dust to high altitudes, and caused the high concentrations of pollutants seen in the free 
troposphere during the prefrontal flight on April 5, ~ 1 day later (Figure 4.14). 
 
 
Figure 4.13. GOES-9 IR image for 00 UTC April 4, 2005, 27 hr prior to the flight 
depicted in Figure 4.7. Back trajectories originating at 500 m (blue line) and 3000 m 
(red line) are superimposed; the transparent red and blue disks show the approximate 
area of origin for air sampled on 5 April.  Note deep convective clouds to the W and 
NW and clear skies to the SW. 
  
 
 115 
 
 
Figure 4.14. MODIS true color image for 04 UTC April 4, 2005 with 2500 m back 
trajectory superimposed. Heavy convection to NW was over sparsely populated, arid 
regions of Mongolia. The transparent red disk indicates the approximate origin of the 
air sampled ~24 h later on the flight of April 5. Baotou and Hohhot are industrial 
cities with substantial coal combustion and steel production. 
 
4.2.2 Postfrontal flight on April 7, 2005 
 Following the flight of April 5, 2005, the cold front passed the Shenyang 
region; the dry, relatively warm continental tropical (cT) air mass moved off the coast 
and a continental polar (cP) air mass brought cooler, drier air (Figures 4.15 and 4.16). 
The flight route on April 7 was similar to that on April 5, but the composition of the 
atmosphere in this cP air mass was largely different (Figures 4.17 and 4.18). Pollutant 
levels were low, with SO2 mixing ratios well below 1 ppb. Aerosol loading, however, 
was lower than on April 5 but still considerable with scattering coefficients of ~8 ? 
 
 116 
 
10-5 m-1 (or ~80 Mm-1). Bear in mind that the inlet employed for this study is 
inefficient at sampling coarse mode aerosols. The NASA OMI Aerosol Index shows a 
broad maximum over Shenyang on April 7, 2005. Ozone concentrations increased 
above the PBL, reaching a maximum of about 65 ppb, higher than expected for air 
with low concentrations of ozone precursors [e.g., Logan, 1999]. 
 Backward trajectories (Figure 4.16) show rapid subsidence in the free 
troposphere and brisk flow from the NW ? an arid and sparsely populated region of 
China and Mongolia. The high wind speed and low concentrations of pollutant gases 
suggest that this could be a dust event with surface material carried aloft from over 
the Gobi desert and other arid dust source regions. Ozone was likely transported 
downward from the upper troposphere/lower stratosphere (UT/LS) rather than locally 
produced in photochemical processes.   
 
 117 
 
 
Figure 4.15. Surface weather analysis (Korean Meteorological Administration) for 00 
UTC April 7, 2005 showing a cyclone NE of Shenyang. The flight was conducted 
behind the cold front in northerly flow. The approximate location of the flights is 
shown by the symbol of an aircraft and the actual flight pattern is shown in the insert. 
Profiling spirals were conducted north and south of Shenyang (locations marked with 
aircraft symbols). Yellow circles indicate occurrence of fog. 
 
 118 
 
 
Figure 4.16. Backward trajectories calculated with the HYSPLIT model for air 
arriving at location of northern spiral (42.5?N, 123.5?E) on April 7 (Figure 4.18). The 
lower panel depicting vertical motions of the trajectories shows subsidence in this 
postfrontal air. 
 
Figure 4.17. Profiles of temperature, trace gases and aerosol scattering north of 
Shenyang from the flight behind the cold front on April 7, 2005. 
 
 119 
 
 Figure 4.18. Profiles of temperature, trace gases and aerosol scattering south of 
Shenyang from the flight behind the cold front on April 7, 2005. 
 
4.2.3 Discussion 
 Lofting of pollutants over North America and the resulting inter-continental 
transport have been studied more thoroughly than over China, and a comparison of 
the two regions can be enlightening. Cold front passage is prevalent over both 
continents, but spring in northern China is dry compared to North America (Figure 
4.19). Over North America, cold fronts often demark the collision of maritime 
tropical (mT) and cP air masses, resulting in convective clouds and precipitation. 
Over northern China, the air mass in the warm sector generally has continental rather 
than maritime character; and the low humidities inhibit precipitation. Most of the 
precipitation in N China falls in the summer months (June to August), while 
precipitation in the NE U.S. falls more uniformly throughout the year. The area near 
 
 120 
 
Beijing receives less than 20 mm of rain per month in March and April, while 
Baltimore, for example, receives about 80 mm per month in the two months. The total 
annual precipitation over N China is about half that over the NE U.S. (Figure 4.19). 
 
 
Figure 4.19. Climatology of precipitation near Baltimore and Beijing [Arakawa, 
1969; Bryson and Hare, 1969]. Comparison between Baltimore and several other 
cities in the NE U.S. (e.g., New York, Philadelphia, and Boston) shows generally the 
same monthly change in precipitation (data source: National Climatic Data Center). 
Note that most of the rainfall over N China comes in the summer while over the NE 
U.S. it is more evenly distributed. In the spring, frontal passage and convection are 
common over both continents, but little rain is produced over NE Asia, thus dry 
convection may play a larger role in pollutant transport. N China has experienced 
severe drought recently, making the differences even greater [Zou et al, 2005]. 
 
 
 In contrast to precipitation, the average cloud cover over the two continents is 
not as dramatically different. Over mid-latitudes (35 to 45?N) the average total cloud 
cover in spring (MAM) is about 50% near and upwind of Beijing (115 to 130?E) and 
64% over the American Northeast (70 to 85?W) [Arakawa, 1969; Bryson and Hare, 
1969]. Cumulus clouds are reported about 7% of the time over N China and 9% over 
the NE U.S. The peak in the diurnal cycle in both places appears at 13:00-14:00 local 
time; cumulonimbus clouds are present in about 1% of the observations over both 
 
 121 
 
continents. Aerosols might have reduced cloud cover over China in the past few 
decades, but not enough to change the basic climatology [Qian et al., 2006; U.S. DOE, 
1986]. In short, convective clouds form over China in spring, but little rain falls. 
 Pollutants over China even if emitted in great amounts, would have little 
impact on the global scale unless they are lofted above the PBL ? rapid vertical 
transport of aerosols, NOx and SO2 increases their lifetimes and range of influence. 
As mentioned in Chapter 1, both WCB and convection play a role in pollutant lifting. 
The cyclone described here generated little precipitation on April 5, and over the 
continent, vertical transport associated with it appears to have been driven by dry 
(non-precipitating) convection well upwind. Dry convection, with minimal wet 
removal, may be especially effective at vertical transport of dust and pollutants. 
 The results from this April 5 flight, when combined with the climatology 
presented above, may have general implications for chemical transport models. 
Global-scale dynamical models can handle explicitly the transport along isentropes, 
such as in the WCB; for the case study presented here, this process was important 
after the cyclone moved eastward off the coast. Convection, in contrast, must be 
parameterized in large-scale models, and accurate representation of the aerosol and 
trace gas transport in dry convection poses a challenge to numerical models. Careful 
evaluation against observations and cloud resolving models is called for.  
 
4.3 Summary 
 Both convection and WCB lofting associated with wave cyclones can play an 
important role in the springtime long-range transport of pollutants from East Asia. 
 
 122 
 
Previous aircraft measurements in this region, mainly conducted off the Chinese coast, 
downwind of China?s populated areas generally identify WCB lofting as the 
dominant meteorological mechanism driving transpacific transport events. Few 
observational studies have dealt with the convective lofting of pollutants over the E 
Asian continent. In this chapter, I investigate both mechanisms by analyzing a series 
of EAST-AIRE research flights in NE China.  
 In the PBL, aircraft-measured air quality responded to mid-latitude cyclones 
in a way similar to that observed on ground (Chapter 3). Transitions between polluted 
prefrontal (~10 ppb SO2, ~1 ? 10-4 m-1 aerosol scattering) and clean postfrontal (~2 
ppb SO2) conditions were recorded, but aerosol loading remained substantial behind 
the cold front probably due to strong dust emissions.  
 In the free troposphere at 1000-4000 m altitude, concentrations of trace gases 
and aerosols were much smaller than in the PBL. During the prefrontal flight on April 
5, substantial pollution with ~300 ppb CO, ~2 ppb SO2, and ~8 ? 10-5 m-1 aerosol 
scattering was measured. Back trajectories show little or no WCB lifting over the 
continent; on the other hand, convective clouds formed near the industrial areas of 
Hohhot and Baotou, ~1000 km upwind, and a day before the flight might have lofted 
pollutants into the free troposphere. Pollutants uplifted can be carried in the westerly 
winds to the locations of aircraft spirals. Stronger convection to the northwest 48 hr 
prior to the flight occurred over sparsely populated regions of Mongolia, which may 
have contributed dust but likely not pollutant gases to the free troposphere. Up until 
this point, dry convection appears to have been the dominant mechanism of vertical 
pollutant transport in this case. 
 
 123 
 
 Behind the cold front on April 7, little SO2 but abundant aerosols (likely 
mineral dust) and ozone were observed above the PBL. High winds in the cold sector 
of the cyclone (cP air mass) might have lofted dust from the surface and subsidence 
or tropopause folding might have brought ozone from the UT/LS to the middle 
troposphere.   
 Comparing the results from the Shenyang aircraft campaign to the chemical 
climatology over the NE U.S. shows that pollutant levels are greater at almost all 
altitudes over N China. Overall, the particle size over N China is larger, probably due 
to prevalence of dust in spring. The springtime meteorological mechanisms leading to 
long-range transport of pollutants are similar over eastern North America and East 
Asia ? both involve the passage of wave cyclones and associated fronts. But one 
fundamental difference is that in the arid spring in N China, convection produces little 
rain ? this makes convective lofting more efficient for pollutant transport and makes 
convection more difficult to simulate. Future studies of the global effects of emissions 
from China must account for dry convection; convective schemes of chemical 
transport models must be evaluated not only for their ability to match observed 
precipitation but also for their ability to simulate deep clouds that do not rain. Further 
aircraft flights in the region will help quantify the relative roles of dry convection and 
the WCB in large-scale atmospheric composition and climate. 
 The aircraft measurements discussed in this chapter were conducted over a 
relatively small region no more than 100 km from the city of Shenyang. In Chapter 5 
I use satellite sensors to further study the April 5 pollution episode on a much larger 
 
 124 
 
scale, and track the pollution plume as it moved eastward with the cyclone over the 
next few days. 
 
 125 
 
Chapter 5:  Evolution of Pollutants during Transport: A 
Perspective from Space 
 
5.1 In-Cloud Processing of Upwind Sulfur Dioxide 
5.1.1 The relative nitrate-to-sulfate ratio in Xianghe aerosol samples 
 One seemingly surprising result from the analysis of Xianghe aerosol samples 
(cf. Chapter 3) is that the average mass concentration of nitrate ([NO3-]) is higher than 
that of sulfate ([SO42-]). Table 5.1 compares this study to some previous studies in the 
region near Beijing. A number of factors might affect the [NO3-]/[SO42-] ratio, for 
example sampling size range, sampling loss, interference of gaseous species, analysis 
method, and sampling time and location. PM10 or bulk aerosol samples often have 
higher [NO3-]/[SO42-] ratio than PM2.5 or PM1 samples, as the reaction between nitric 
acid and soil particles [Seinfeld and Pandis, 1998] may form nitrate compounds in the 
coarse mode. In N China, the winter [NO3-]/[SO42-] ratio (e.g., 0.5, Sun et al. [2004]; 
0.2, Hu et al. [2002]) can be lower than that in springtime samples (e.g., 0.9, Wang, 
Zhuang et al. [2005]; 0.7, He et al. [2001]), due to extensive coal burning for heating. 
Growing emissions from vehicles and large point sources in the region [Q. Zhang et 
al., 2007] may cause a larger [NO3-]/[SO42-] ratio in more recent samples [e.g., Zhang 
et al., 2004]. If no denuder is employed in the sampling process, gaseous nitric acid 
can be absorbed by the aerosol filters, leading to a positive artifact in the nitrate 
concentration (e.g., this study).   
 
 126 
 
 Given the similarities in sampling season (spring), year (2005 vs. 2001-04),  
size range (bulk), and method (no denuder), I compare this study to Wang, Zhuang et 
al. [2006] to differentiate aerosol composition between rural (Xianghe) and urban 
(Beijing) locations in the region. An earlier study sampled both urban and rural 
aerosols in the area in autumn [Huebert et al., 1988]. 
 Higher average [NO3-] (17.9 ?g/m3) and [SO42-] (20.8 ?g/m3) are found in 
more polluted downtown Beijing [Wang, Zhuang et al., 2006]. The [NO3-]/[SO42-] 
ratio in Xianghe, on the other hand, is greater than in Beijing. Both sites are 
influenced by nearby sources, but probably more so for Beijing, as in general cities 
have stronger local emissions and weaker surface winds due to friction. NOx reacts 
faster to form NO3- than SO2 does to from SO42-; and therefore [NO3-]/[SO42-] ratio at 
the downstream site in Xianghe may be greater than in upwind metropolitan Beijing 
when W or SW winds prevail. Additionally, the availability of NH3 and soil particles, 
which neutralize HNO3, differs between urban and rural atmospheres. Uncertainties 
in chemical analysis of the two studies may influence the ratio as well. Huebert et al. 
[1988] show a similar difference in the [NO3-]/[SO42-] ratio between urban and rural 
sites under hazy conditions (Table 5.1).  
 
Table 5.1. Average nitrate and sulfate concentrations from this study and some 
previous studies (?g/m3). 
Study Time Location Size NO3- SO42- Remark 
This 
study 
Spring 
2005 
Xianghe Bulk 9.5 7.2 No denuder, rural site near 
Beijing; Whatman 41 filters. 
Hu et al. 
[2002] 
Dec. 
2000 
Beijing Not 
specified 
5.6 28.0 Online ion chromatograph with 
gas measurements, urban site. 
Sun et al. 
[2004] 
Winter 
2002 & 
2003 
Beijing PM10 18.7 34.5 Medium volume, Whatman 41 
filters, urban site, usage of 
denuder not specified. 
Yao et 
al. 
[2003] 
1999-
2000 
Beijing PM2.5 10.3 18.4 Sampling during the whole year, 
with denuder, urban site. 
 
 127 
 
Zhang et 
al. 
[2004] 
1994-
1995 
Beijing  PM2.5 5.2 15.2 Details of sampling and analysis 
not given. 
 2000 Beijing PM2.5 5.6 9.9  
Wang, 
Zhuang 
et al. 
[2005] 
Spring 
2001-
2003 
Beijing PM2.5 11.9 13.5 Five urban and suburban sites, 
similar sampling and analysis as 
Sun et al. [2004] 
He et al. 
[2001] 
Spring 
2000 
Beijing PM2.5 7.3 10.2 Low-flow sampler with denuder, 
urban sites 
Zheng et 
al. 
[2004] 
Spring 
2004 
Beijing PM2.5 ~20 ~20 Sampling at 6-day interval, 5 
urban and suburban sites in 
Beijing area, concentrations 
roughly read from the figure in 
paper, usage of denuder not 
specified. 
Wang, 
Zhuang 
et al. 
[2006] 
Spring 
2001-
2004 
Beijing Bulk 7.2 19.6 Similar sampling system as Sun 
et al. [2004], urban sites, 40 
?dust? samples. 
    47.1 48.9 Same study as above, 32 ?haze? 
samples. 
    13.6 13.6 Same study as above, 117 
?clear? samples. 
Huebert 
et al. 
[1988] 
Oct. 
1985 
Beijing Bulk 1450a  
3.7b 
2050a  
8.2b 
6 samples collected on two days 
under ?hazy? and ?slightly 
hazy? conditions. 
  Xinglong Bulk 50a 
0.1b 
150a 
0.6b 
Same study as above, 100 km 
NE of Beijing, N winds, clear 
  Xinglong Bulk 1440a 
3.7b 
1540a 
6.2b 
Same study as above, W winds, 
hazy. 
a Unit: parts per trillion (ppt). 
b Unit converted to ?g/m3 assuming 20?C, 101.325 kPa. 
 
5.1.2 Synoptic change in [NO3-]/[SO42-] 
 Figure 5.1 revisits the time series of nitrate and sulfate; the two secondary 
aerosol compounds are closely related to anthropogenic emissions and in correlation 
with CO (Figure 3.3). In Chapter 3 I have described in detail the synoptic variation of 
air quality seen in Xianghe. Briefly, weak S winds ahead of cold fronts allow 
pollutants to accumulate; strong winds from the less populated N/NW behind cold 
fronts ventilate local emissions, bringing relatively clean air. 
 
 128 
 
 In Figure 5.1, nitrate is more abundant than sulfate under polluted conditions, 
and vice versa under relatively clean conditions. This change of [NO3-]/[SO42-] ratio 
with pollutant level and synoptic conditions is qualitatively consistent with Wang, 
Zhuang et al. [2006]. Their ?dust? samples are associated with strong winds (average 
wind speed: 8 m/s) and dry air (average RH: 20%) under postfrontal conditions (cf. 
Chapter 3). The corresponding nitrate concentration (7.2 ?g/m3) under these 
conditions is less than half of sulfate (19.6 ?g/m3). In contrast, the same ratio is 
almost one for ?haze? samples, which comprise on average 47.1 ?g/m3 of nitrate and 
48.9 ?g/m3 of sulfate. These ?haze? samples were mostly collected under stagnant 
(average wind speed: 2.5 m/s) and relatively moist (average RH: 59%) conditions, 
usually found ahead of cold fronts. The ?clear? samples [Wang, Zhuang et al., 2006] 
may reflect both prefrontal and postfrontal conditions, and also the transition in 
between. Similarly, Huebert et al. [1988] found that as winds switched from 
generally N to generally W and brought pollutants to their site NE of Beijing, the 
observed nitrate-to-sulfate mole ratio went up from 0.33 to 0.93.  
 
 129 
 
 
Figure 5.1. Concentrations of nitrate (squares) and sulfate (circles) in bulk aerosol 
samples from Xianghe, March 1 to April 13, 2005. 
 
5.1.3 SO42-/SOx and NO3-/NOy mole ratios 
 The mole ratio of SO42-/SOx, where SOx = SO42- + SO2 (units: ?mol/m3), is 
plotted against the NO3-/NOy mole ratio in Figure 5.2 as a function of CO 
concentration. Considering the positive artifact due to absorption of HNO3 to filters, 
the calculated NO3-/NOy ratio is better viewed as a proxy for the fraction of NOx 
converted to secondary species including HNO3. Relatively clean samples (CO < 500 
ppb) mostly have higher SO42-/SOx ratio than NO3-/NOy ratio, and vice versa for 
polluted samples (CO > 1000 ppb). The relative rates of sulfate and nitrate production 
may change with weather; and consequently so does the relative abundance of the 
two species. 
 
 130 
 
 
Figure 5.2. SO42-/SOx and NO3-/NOy mole ratios as a function of CO concentration. 
Under relatively clean conditions, SO42-/SOx ratio is generally higher than NO3-/NOy 
ratio, and vice versa under polluted conditions. 
 
 Scatter plots of SO42-/SOx and NO3-/NOy mole ratios with various 
meteorological parameters in Figure 5.3 provide some insight into the conversion 
mechanisms of SO2 and NOx. The SO42-/SOx mole ratio shows little correlation with 
temperature (Figure 5.3b), implying an additional SO2 conversion pathway besides 
the temperature-dependent gas-phase oxidation of SO2. A fraction of sulfate may also 
form during transport (other than locally generated). The relationship between SO42-
/SOx mole ratio and RH in Figure 5.3a largely reflects the distinct properties of air 
masses; SO42-/SOx is greater in the very dry but relatively clean postfrontal arctic air. 
Likewise, samples of high SO42-/SOx ratios are mostly related to strong northerly 
winds behind cold fronts (Figures 5.3c and 5.3d). The conversion of NOx, in contrast, 
is temperature dependent and correlated with solar radiation (overall, daytime 
 
 131 
 
samples more reacted) as shown in Figure 5.3f. The NO3-/NOy ratio is low when it is 
cold and dark (lower ratio for night time samples shown as circles). The only 
nighttime sample with a relatively high NO3-/NOy ratio had temperature of ~15?C. 
The formation of nitrate is probably more through gas-phase photochemical process 
than through conversion from N2O5 to nitrate at night. 
 
 132 
 
 
Figure 5.3. Scatter plots of SO42-/SOx mole ratio with a) relative humidity, with b) 
temperature, with c) wind speed, with d) wind direction, with e) Al concentration, and 
f) NO3-/NOy mole ratio with temperature. Squares and circles represent daytime and 
nighttime samples, respectively. Solid lines are results of correlation analysis between 
SO42-/SOx mole ratio and wind speed (c), and between NO3-/NOy mole ratio and 
temperature (f) for all samples. 
 
 
 133 
 
 Excluding locally emitted SO2, which reacts slowly in the cold postfrontal 
environment (Chapter 3), I can approximate the SO2 conversion ratio for 
?transported? sulfur species from the north. The abundance of local SO2 is estimated 
assuming that its mixing ratio is 1/35.8 that of locally emitted CO. The value 35.8 
comes from the correlation analysis between CO and SO2 (Chapter 3), and 
presumably reflects the CO/SO2 emission ratio from sources in the region. Local CO 
is calculated by subtracting the intercept of the same correlation analysis from the 
measured CO. This intercept (180 ppb) is also roughly the lowest CO recorded during 
the experiment, and is comparable to the ?background? values at mid-latitude 
continental sites in northern hemisphere [Novelli et al, 1998]. Under windy conditions 
(mean surface wind speed > 4 m/s), 8 of 12 samples have ?transported? SO42-/SOx 
mole ratio greater than 0.5 (2 with ratio of about 0.4). Note that this analysis is 
qualitative only, given our assumption of neglecting locally generated sulfate and the 
uncertainties in CO/SO2 emission ratio that have been discussed in Chapter 3. The 
above analysis seems to indicate that the mixing of aged SOx from distant sources 
may be the non-temperature dependent source of sulfate observed in Xianghe behind 
cold front. 
 One question that remains unanswered is why the SO42-/SOx ratio would be 
higher when winds blow from the cold north, where gas-phase photochemical 
reactions oxidizing SO2 are likely slow? The interaction between dust particles and 
pollutants, particularly in dust storms, may accelerate sulfate formation, and has been 
the subject of numerous studies, [e.g., Sun et al., 2005; Wang, Zhuang et al., 2007; 
Yuan et al., 2008], although some find little evidence of sulfate forming on dust [e.g., 
 
 134 
 
Zhang and Iwasaka, 1999]. Figure 5.3e is a scatter plot of the SO42-/SOx ratio and 
aluminum (Al), used here as a tracer for soil/dust particles. Recall that in Chapter 3 it 
was found that aluminum is a good tracer for transported dust, whereas some other 
crustal elements (e.g., calcium, iron, and potassium) are more related to local 
anthropogenic activities. Another possibility is that CaSO4 in East Asian dust may 
contribute to sulfate in aerosol samples, although it is much less abundant than 
calcium carbonate (CaCO3) [Gao et al., 2007]. As can be seen in Figure 5.3e, several 
samples have a relatively high SO42-/SOx mole ratio but low Al or dust level; sulfate 
in these samples is not likely from CaSO4 in dust or generated by interaction between 
SOx and dust particles. Aqueous-phase oxidation of SO2, for example through 
reaction with hydrogen peroxide (H2O2), is more efficient than gas-phase reactions 
[Seinfeld and Pandis, 1998]. In comparison, the aqueous-phase conversion from NOx 
to nitrate probably contributes negligibly in this study [Seinfeld and Pandis, 1998]. 
Therefore in-cloud processing of SO2, if occurs on the transport route from the north, 
can probably cause the relatively high SO42-/SOx but low NO3-/SO42- ratios observed 
in the postfrontal air, and is probably the source of non-temperature dependent 
sulfate.  
 
5.1.4 In-cloud processing of SO2 
 To find out whether in-cloud processing may have occurred during postfrontal 
transport, I analyze backward trajectories and satellite data. Backward trajectories 
calculated using the HYSPLIT model and MODIS cloud properties are synchronized 
in time and space: I overlap each trajectory point (with time, latitude, longitude, and 
 
 135 
 
height in mb from the model) with MODIS-retrieved cloud top pressure (CTP) and 
cloud optical thickness (COT) within ? 15 min, and ? 2.5? latitude/longitude. Thick 
clouds with top above the trajectory height, and clouds with top near trajectory height 
are more likely to interact with SOx traveling alongside the trajectory. MODIS CTP 
and COT are used to identify such clouds. Given the uncertainties in satellite 
retrievals, particularly for multi-layer clouds (e.g., cirrus over low level clouds), 
additionally I use NCEP reanalysis RH along the trajectory to judge the existence of 
clouds. 
 As an example, the upper and middle panels of Figure 5.4 present the 
synchronized trajectory-cloud plot for the daytime sample on March 17, 2005. Over 
Lake Baikal, where the trajectories were at the level of 700-800 mb, MODIS 
retrieved thick clouds with COT ? 15 and CTP ~600-700 mb, suggesting possible 
interaction between clouds and air parcels. Note the high RH (over 80%) along the 
trajectory. The lower panels of Figure 5.4 are histograms of MODIS-retrieved CTP 
and COT around the position of the 500-m trajectory (55.312?N, 108.117?E, 794.8 
mb, and RH ~87%), 48-hr prior to its arrival in Xianghe. To account for the 
uncertainties of trajectory computation, which is proportional to the distance traveled 
from the trajectory end point [Stohl, 1998], the histograms incorporate MODIS data 
within ?15 min of the trajectory time and the spatial uncertainty range of the 
trajectory (20% of the traveled distance from Xianghe). More than half of the clouds 
have CTP between 600 and 800 mb, close to the trajectory height. Some well-
developed clouds exist, although relatively thin clouds with COT < 10 dominate. 
Based on Figure 5.4, I conclude that the air sampled during daytime on March 17 
 
 136 
 
likely has a history of cloud-interaction. The SO42-/SOx mole ratio of the sample is 
relatively high at 0.3, likely due to oxidation of SO2 by H2O2 in the condensed phase.  
 
 
Figure 5.4. Upper: 72-hr backward trajectories starting at 200 m, 500 m, and 1000 m 
above ground from Xianghe at 04 UTC on March 17, 2005, with colors representing 
trajectory height (left) and NCEP reanalysis RH along the trajectories (right).  
Middle: MODIS-retrieved cloud top pressure (left) and cloud optical thickness (right) 
 
 137 
 
along the trajectories, synchronized with trajectory waypoints in time (? 15 min) and 
space (? 2.5 ? latitude/longitude). Lower: histograms of MODIS cloud top pressure 
(left) and cloud optical thickness (right) near the trajectory (500 m), 48 hr prior to its 
arrival. Position of the air parcel: 55.312?N, 108.117?E, 794.8 mb; RH: 86.9%. 
 
 Having examined similar plots on a case-by-case basis, I find that 16 out of 23 
samples with NO3-/SO42- mole ratio below 1.33 are related to trajectories with a 
history of possible cloud interaction, and 13 of the 23 samples have lower-than-
average Al concentration and are relatively ?dust-free?. Among the 23 samples, 9 are 
possibly cloud-influenced and ?dust free?, and are defined as ?cloud?. Two with 
abundant dust are called ?dust?, 7 have both high dust levels and probable cloud 
interaction, and are named ?cloud and dust?, and 5 remain unaccounted for. Similarly, 
out of the 15 samples with SO42-/SOx mole ratio > 0.1, 6 are ?cloud?, 4 are ?dust?, 
and 4 are ?cloud and dust?. Additionally, 10 of 11 samples with NO3-/SO42- < 0.8, and 
5 of 6 samples with SO42-/SOx > 0.15 are ?cloud? or ?cloud and dust?. Note that the 
fraction of cloud-related samples could be underestimated, as MODIS provides 
measurements only once a day (the daytime COT product is used). Figure 5.5 
summarizes the results. 
 
 
 138 
 
 
Figure 5.5. Percentage of ?cloud?, ?dust?, ?cloud and dust?, and ?unexplained? 
events for samples with NO3-/SO42- mole ratio below 1.33 (left), and samples with 
SO42-/SOx mole ratio over 0.1 (right). 
 
5.1.5 Implications for regional atmospheric chemistry 
 As discussed in Chapter 3, SO2 is the major SOx species in N China in spring; 
its high abundance relative to other SOx species is a combined result of strong local 
SO2 emissions and slow photochemical processes. When a mid-latitude wave cyclone 
moves eastward, air flow behind its cold front carries pollutants away from northern 
China, usually within the PBL (cf. Chapter 1). Several such PBL transport episodes 
featuring high SOx concentration were recorded in Cape Hedo, Japan, ~1000 km E of 
Xianghe, and a site of the Asian Brown Cloud (ABC) field campaign in coordination 
with EAST-AIRE [Takami et al., 2007]. Takami et al. [2007] observed that the SOx in 
these peaks was almost completely sulfate and proposed that only in-cloud processing 
would be fast enough for nearly complete conversion of SO2 from China. Together, 
this study and Takami et al. [2007] show that under postfrontal conditions, aqueous 
oxidation may play an important role in the formation of sulfate aerosols, both 
 
 139 
 
upwind and downwind of the strong emission sources in N China. Sulfate particles 
from the region N of China, although low in concentration, can be highly efficient 
cloud condensation nuclei (CCN), and may impact cloud systems over downwind 
areas [e.g., Zhang et al., 2007] (and subsequently aqueous chemistry there). Chemical 
transport models need to properly treat both clouds and dust to accurately represent 
the regional sulfur chemistry in East Asia. In-cloud processing of SO2 can also impact 
the distribution of sulfate aerosols and the regional radiation budget [e.g., Takemura 
et al., 2003]. 
 
5.2 Transport and Evolution of a Pollution Plume from Northern China: A Satellite-
Based Case Study 
5.2.1 Introduction 
As discussed in Chapters 3 and 4, mid-latitude wave cyclones play a central 
part in building up and venting pollution plumes over N China. These plumes with 
high concentrations of pollutants could travel far downwind off the Chinese coast, 
and potentially influence large areas (cf. Chapter 1 and the references therein). A key 
factor determining the large-scale impacts of a pollution plume is how the plume 
evolves over time and space. The SO2-to-suflate conversion can supply cloud 
condensation nuclei and may greatly affect the aerosol-cloud system over the 
northern Pacific [R. Zhang et al., 2007]. Diffusion and advection can dilute a plume, 
lessening but spreading its effects. Eulerian [e.g., Carmichael et al., 2003] and 
Lagrangian [e.g., Stohl et al., 2003] models can simulate the evolution of pollution 
plumes, but need observations for evaluation. By intercepting a plume a few times on 
 
 140 
 
its path, Lagrangian or semi-Lagrangian airborne experiments can sample the plume 
at its different stages [e.g., Methven et al., 2006; Ramanathan et al., 
http://www.eol.ucar.edu/projects/pacdex/], but aircraft observations are costly and 
scarce. 
The past several years have witnessed great progress in satellite remote 
sensing of pollutants in the troposphere. Satellite measurements have been used to 
track dust transport [e.g., Hursar et al., 2001], characterize the chemical signature of 
frontal systems [e.g., Liu et al., 2006], monitor long- [e.g., Richter et al., 2005] and 
short-term [e.g., Y. Wang et al., 2007] change in pollutant levels, evaluate emission 
inventories [e.g., Akimoto et al., 2006], and constrain chemical transport models [e.g., 
Allen et al., 2004]. With daily global coverage, finer resolution, and almost 
simultaneous observations of multiple pollutants, a series of satellite sensors in the 
?A-train? constellation provide unprecedented capacity in measuring air pollution 
from space. 
This section can be viewed as a follow-up on the analysis on the April 5 
research flight over N China given in Chapter 4. Taking advantage of the ?A-train?, I 
look at the air pollution episode at a regional scale, and track the evolution of the 
pollution plume as it moves to the east over the next few days, where and when in-
situ measurements were not available. 
 
5.2.2 Data and models 
The Ozone Mapping Instrument (OMI) onboard NASA?s Earth Observing 
System (EOS) Aura satellite retrieves SO2 column content using UV spectral fitting 
 
 141 
 
in the wavelength range of 310-380 nm [Krotkov et al., 2006; 2008]. In the 
operational OMI PBL SO2 product, the retrieved slant column SO2 density (SCD) is 
converted to total SO2 vertical column density (VCD) with a constant air mass factor 
(AMF, defined as the ratio between the slant column to the vertical column) of 0.36: 
total SO2 VCD = AMFSCD   (5.1) 
This operational AMF assumes cloud- and aerosol-free conditions, solar zenith angle 
of 30?, surface albedo of 0.05, surface pressure of 1013.13 mb, mid-latitude ozone 
profile of 325 DU, and a vertical SO2 profile typically observed over the eastern U.S. 
in summer, with most of the SO2 below 900 mb [Taubman et al., 2006]. The PBL 
OMI SO2 product has been validated against aircraft measurements during the EAST-
AIRE campaign (Chapter 4), and could distinguish between polluted and background 
conditions over N China on a daily basis [Krotkov et al., 2008]. The noise of the PBL 
SO2 product is ~1.5 DU at the instrument?s instantaneous field of view (FOV, 13 ? 24 
km at nadir), but can be greatly reduced through temporal and spatial averaging. 
Section 5.2.3 details the AMF corrections made in this case study. 
Aerosol optical depth (AOD) retrieved from the MODerate resolution 
Imaging Spectrometer (MODIS) instrument aboard NASA?s EOS/Aqua satellite 
provides aerosol information in this study, with a resolution of 10 ? 10 km. The 
MODIS AOD retrieval algorithm derives aerosol properties over dark surfaces (cloud, 
ice, snow, sun-glint, and desert free), and is based on a ?lookup table? approach, i.e., 
the observed radiation field is compared against the pre-computed radiative transfer 
calculations with a set of assumed aerosol and surface parameters [Remer et al., 
2005]. The best fit found through the comparison is the solution to the inversion. 
 
 142 
 
Over oceans, the MODIS algorithm assumes that one fine and one coarse lognormal 
mode, if properly weighted, can combine to represent ambient aerosol properties. 
Over land, the surface reflectance can be derived from longer wavelength MODIS 
channels, as atmospheric extinction due to aerosols is generally low in the 2.12 ?m 
band. An earlier version of the MODIS land algorithm assumed constant ratios 
between the surface reflectance at 0.47 and 0.66 ?m and at 2.12 ?m. The recently 
updated (collection 5: C005-L) MODIS land algorithm [Levy et al., 2007] 
parameterizes this spectral surface reflectance relationship as a function of viewing 
geometry and surface type (NDVI: normalized difference vegetation index). The 
algorithm also employs new aerosol models derived from surface sun photometer 
measurements (AErosol RObotic NETwork: AERONET); and simultaneously 
retrieves surface reflectance, AOD, and fine mode aerosol weighting. The MODIS 
AOD product has been validated against sun photometer measurements over China 
[Mi et al., 2007], and the collection 5 data show much improved agreement with 
surface observation. Both Aura and Aqua satellites are part of the ?A-train? satellite 
constellation and pass over Shenyang region ~15 min apart at about 1 pm local time 
(05 UTC).  
The HYSPLIT model along with NCEP global reanalysis data are used to 
calculate forward trajectories in this study. 72-hr forward trajectories from eight 
layers (every 500 m from 250 to 3750 m above ground level) are initiated from 
0.5??0.5? grid cells in the region of 35-49?N, 117-134?E (red rectangular boxes in 
Figure 5.11a), at 5 UTC, April 5, 2005. Each trajectory represents an air parcel 
0.5??0.5??500m in size, and is tagged with AOD and SO2 retrieved by MODIS and 
 
 143 
 
OMI and weighted with aircraft profiles. AOD retrievals are missing in some grid 
cells; and trajectories originating from these grid cells are labeled with the mean 
AOD of adjacent grid cells. Grid cells with SO2 below a certain threshold value (cf. 
section 5.2.5) are considered ?non-detectable? by OMI. Trajectories from these grid 
cells are assigned half of the threshold SO2. There is no special treatment for parcels 
with relatively low AOD, as MODIS appears to have good sensitivity even for low 
aerosol loadings [Levy et al., 2007a]. Assuming that pollutants within each parcel are 
conserved, these ?tagged? air parcels together can project and map the spatial 
distribution of SO2 and AOD in the plume. 
The aerosol extinction coefficient profile at 532 nm and depolarization ratio 
were measured in Sapporo, Japan (43.1 ?N, 141.3?E), using a NIES (National 
Institute for Environmental Studies) lidar system. The extinction coefficients of non-
spherical aerosols (dust) and spherical particles are computed following Shimizu et al. 
[2004]. Aerosol optical depth over the flight region was determined from the surface 
with a hand-held sun photometer on April 5 (Figure 4.4, measurement location: 
41.5?N, 123.6?E). 
This transport event was also reproduced in a Goddard Chemistry Aerosol 
Radiation and Transport (GOCART) model simulation. Driven by the assimilated 
meteorological fields of the Goddard Earth Observing System Data Assimilation 
System (GEOS DAS), GOCART is a global model with a horizontal resolution of 2 
(latitude) by 2.5 (longitude) degrees and 20-55 vertical sigma layers [Chin et al. 
2000]. The output of the GOCART model includes SO2 as well as concentrations and 
 
 144 
 
aerosol optical depth of sea salt, sulfate, organic and elemental carbonaceous 
compounds, and dust.  
 
5.2.3 AMF correction for OMI operational SO2 product 
AMF is a function of SO2 profile, surface albedo (Rs), viewing geometry 
(viewing angle: ?, solar zenith angle: ?0, and relative solar azimuth angle: ?), total 
column ozone (?), aerosols, and clouds [Krotkov et al., 2008]: 
( ) ( )?? ?=
0
0 2,,,,, dzznRzmAMF SOs ???   (5.2) 
( ) ( )( )zIzm
SO
TOA
2
ln
??
?=   (5.3) 
where m(z, Rs, ?, ?, ?0, ?) is the vertically resolved OMI SO2 sensitivity, ITOA is the 
OMI measured reflectance at TOA (top of the atmosphere), ?SO2(z) is the SO2 
absorption optical thickness at the layer z (km), and nSO2(z) is normalized SO2 
vertical profile. AMF can be obtained from a forward radiative transfer calculation; 
and the corrected SO2 column amount would be: 
( ) )()( 36.0 22 loperationaSOcorrectedAMFcorrectedSO =   (5.4) 
Radiative transfer calculation of AMF for each pixel or grid cell is so 
computationally demanding that it is impractical. Krotkov et al. [2008] show that the 
viewing geometry correction of AMF can be parameterized through linear regression 
with respect to the slant column ozone (SCO): 
( ) SCOrrcorrectedAMF ??= 10   (5.5) 
 
 145 
 
( ) ( )( )0secsec ?? +??=SCO    (5.6) 
where ? is the total column ozone measured by OMI. The slope (r1) and the intercept 
(r0) of the linear regression depend on the normalized SO2 vertical profile, the surface 
albedo (Rs), relative solar azimuth angle (?), aerosols, and clouds. This linear 
regression could be further simplified, as Rs is normally small in the UV band for SO2 
retrieval, causing an error less than 15%, and changes of ? would result in typically 
?10%, and in extreme cases no more than 20% error in AMF. In this study, AMF-
SCO regression coefficients are derived under various assumptions concerning SO2 
profile, aerosols, and clouds. For example, on a given day, the clear-sky AMF-SCO 
relationship is derived with normalized SO2 and aerosol vertical profiles and the 
average MODIS AOD around the core part of the plume, for two different aerosol 
types (industrial and dust): 
SCOrrAMF INDUCINDUCINDUC ??= ,,1,,0,   (5.7) 
SCOrrAMF DUSTCDUSTCDUSTC ??= ,,1,,0,   (5.8) 
where AMFC,INDU and AMFC,DUST stand for the corrected clear-sky AMF for 
industrial and dust aerosols, respectively. The retrieval assumes the same single 
scattering albedo (0.9) at 550 nm for both aerosol types, but each has a different size 
distribution and spectral dependence of absorption. AMFC,INDU and AMFC,DUST are 
then weighted with GOCART simulated dust (AODS,DUST) and total AOD (AODS) to 
obtain the clear-sky AMF (AMFC): 
S
DUSTS
AOD
AODW ,=   (5.9) 
( )[ ] ( )[ ] SCOrWrWrWrWAMF DUSTCINDUCDUSTCINDUCoC ??+????+??= ,,1,,1,,0,, 11   
 
 146 
 
          (5.10) 
 The AMF under cloudy conditions can be calculated similarly for the same 
two aerosol types, using OMI measured reflectance (331 nm) and MODIS retrieved 
cloud top pressure. An assumption is that high clouds are mostly thin and have little 
influence on SO2 retrieval. Low clouds on the other hand enhance OMI sensitivity to 
SO2 above them. As a first-order approximation, in radiative transfer calculations for 
the two cloudy days (April 6 and 7), overcast conditions are assumed: a homogenous 
cloud ?floor? with the mean reflectance of the region is placed at the average cloud 
top height of MODIS low clouds (Figure 5.6). The resulting AMF?s are then 
combined using the weighting factor derived with Eq. 5.9. No correction was made 
for relatively clear sky conditions on April 5. 
 
 147 
 
 
Figure 5.6. Histograms of MODIS cloud top pressure near the main body of the 
plume on April 6 (left) and 7 (right). Low clouds refer to those with cloud top 
pressure > 500 mb. 
 
In this study, aircraft profiles of SO2 and aerosols are used in the AMF 
calculation for April 5 (Figures 4.9 and 4.10). I calculate the trajectory-projected SO2 
and aerosol loadings at different heights (every 500 m, 0-6500 m) over the center of 
the plume (40-45?N, 138-144?E on April 6; 35-40 ?N, 160-165?E on April 7). The 
normalized vertical distributions of SO2 and aerosols are referred to as trajectory-
projected profiles (Figures 5.7, 5.8, and 5.10). Trajectory-projected SO2 profiles are 
employed for the AMF calculations for April 6 and 7, as aircraft measurements are 
not available. Lidar measurements over Sapporo, Japan (Figure 5.9) and trajectory 
 
 148 
 
projection (Figure 5.10) provide aerosol profiling information for April 6 and 7, 
respectively. 
 
 
Figure 5.7. Normalized trajectory-projected and GOCART SO2 vertical profile over 
N Japan on April 6, 2005. 
 
 
Figure 5.8. Normalized trajectory-projected and GOCART SO2 vertical profile over 
N Pacific on April 7, 2005. 
 
 149 
 
 Figure 5.9. Upper: aerosol extinction coefficients (m-1) in Sapporo, Japan on April 6, 
2005; Lower: vertical profile of aerosol extinction by spherical (red), non-spherical 
(dust, green), and total (blue) aerosols, averaged from 0330 UTC to 0430 UTC. Note 
that using the wind speed at 2000 m (~20 m/s from local sounding data) and time the 
plume passed over the region (~5 hr), the size of the plume is estimated to be ~400 
km along the wind direction (approximately from west to east). This is roughly the 
same as the satellite observed instantaneous plume size shown in Figures 5.11 and 
5.12.  
 
 150 
 
 Figure 5.10. Normalized trajectory-projected and GOCART aerosol vertical profile 
over N Pacific on April 7, 2005. 
 
5.2.4 Transport and evolution of the plume 
The weather conditions and air quality measured by aircraft near Shenyang on 
April 5, 2005 have been discussed in detail in Chapter 4, and are only briefly 
reviewed here. Weather during this prefrontal flight featured strong surface winds 
from the south, local dust emissions and low visibility (4-5 km). Winds veered with 
altitude and became westerly in the middle troposphere. The aircraft detected ~20 ppb 
SO2 within the PBL, and 1-3 ppb SO2 in the FT, which translates into a vertical 
column amount of ~2 DU. The vertical distribution of aerosols was similar to that of 
SO2, with ~500 Mm-1 scattering in the PBL and ~100 Mm-1 scattering in the FT. The 
aircraft-integrated and surface observed AOD (550 nm) were comparable at ~1. The 
?ngstr?m Exponent (AE) from the surface AOD measurements was about 0.8, 
suggesting large overall size of particles. Single scattering albedo retrieved for the 
day is 0.87 [K-H. Lee, personal communication], indicating light-absorbing aerosols. 
 
 151 
 
Satellite observations, in agreement with aircraft and surface measurements, 
detected high levels of SO2 and aerosols on April 5, but over a much larger area. The 
average OMI SO2 column content in the region near Shenyang (rectangular box, 
Figure 5.11a) was about 1.5 DU, with some ?hot spots? exceeding 3 DU. The mean 
AOD (Fig 5.12a) in the same region was ~0.7, and higher values (~1.8) were found to 
the southwest. Fairly high OMI-observed Aerosol Index (AI, ~1.5, Figure 5.12d) in 
the region implied the existence of absorbing aerosols. Dust was likely the dominant 
aerosol compound according to high AI, low AE, meteorological records, and 
GOCART output (Figure 5.13). 
 
 
Figure 5.11. OMI-retrieved (a, b, and c) and trajectory-calculated (d and e) SO2 from 
April 5 to 7 near the main part of the pollution plume. 
 
 152 
 
 
Figure 5.12. MODIS-retrieved AOD (a, b, c), OMI aerosol index (AI, d, e, f), and 
trajectory projected AOD (g, h, f) from April 5 to 7, 2005.  The rectangular box in 
panel a (also d and g) indicates the area from which forward trajectories are initiated. 
Polygon boxes in b (e, h) and c (f, i) are examples of the boxes used to represent the 
core part of the plume, and to calculate the total SO2 mass on April 6 and 7. Missing 
MODIS AOD retrievals in upper panels are mainly due to clouds, sun glint, or high 
surface reflectance (April 5). These grid cells are not included in the calculation of 
the average MODIS AOD. 
 
 
 153 
 
 
Figure 5.13. GOCART simulated SO2 column amount (DU), and AOD of black 
carbon (BC), dust (DU), organic carbon (OC), sulfate (SU) and sea salt (SS) at 06 
UTC, April 5, 2005. Red, green, and yellow boxes in the figure indicate domains for 
averaging GOCART output (results shown in Figure 5.20). 
 
 
 154 
 
As shown in Figures 5.11 and 5.12, trajectory calculations indicate that the 
pollution plume in general traveled eastward, reaching N Japan on April 6 and the 
NW Pacific on April 7. OMI and MODIS successfully captured the pollutants on both 
days, taking ?snapshots? of the plume (Figures 5.11d, e and 5.12h, i). The core part of 
the plume depicted by the OMI aerosol index (Figure 5.12e, f) resembles that mapped 
with trajectories (Figure 5.12h, i), suggesting that the forecast plume agrees with 
satellite observations and that AI could be a useful tool for tracking aerosol transport. 
GOCART simulated aerosol and SO2 loadings over N Japan on April 6 (Figure 5.14) 
and over the NW Pacific (Figure 5.15) on April 7 also confirm the trajectory-
projected movement of the plume. Based on Figure 5.9, the aerosol plume passed 
over the lidar site in ~5 hr on April 6. Using the wind speed at the plume center 
height (~20 m/s at 2000 m from sounding data), the size of the plume was estimated 
to be ~400 km along the wind direction (approximately from west to east). This is 
roughly the same as the satellite observed instantaneous plume size (Figure 5.12).    
Bear in mind that the trajectory calculations use SO2 and aerosols as ?inert 
tracers? and therefore only account for the advection and dilution of the plume 
(Figures 5.11 and 5.12). OMI retrieved SO2, on the other hand, reflects the combined 
effect of transport, chemical conversion, and change in satellite sensitivity and noise; 
as does MODIS retrieved AOD. Examining Figures 5.11 and 5.12 shows that near the 
major part of the plume, OMI SO2 is considerably lower than trajectory SO2 (e.g., 
Figure 5.11b vs. d), while MODIS AOD is greater than trajectory projection (e.g., 
Figure 5.12b vs. h). This seems to indicate that as the pollution plume moved 
downwind, SO2, a precursor of sulfate aerosols, was lost whereas aerosols were 
 
 155 
 
introduced (or the aerosol extinction was enhanced). Dust remained the dominant 
aerosol species throughout the three-day period, as evidenced by OMI AI (Figure 
5.12d, e, f) and model simulated AOD (Figures 5.13-15), but sulfate became 
increasingly important over time (Figures 5.13-15). Is it possible that the satellite 
sensors detected signal of primary trace gas (SO2) converting to secondary aerosols 
(sulfate)? Next I further discuss the evolution of the chemical properties of the plume.  
 
 156 
 
 
Figure 5.14. Same as Figure 5.13 but for 03 UTC, April 6, 2005 
 
 157 
 
 
Figure 5.15. Same as Figure 5.13 but for 03 UTC, April 7, 2005 
. 
 
 158 
 
5.2.5 Changes in AOD and total mass of SO2 
Clouds, satellite orbit, surface reflectance, and the signal to noise ratio are 
some restraining factors limiting observation of pollution plumes from space. 
Dispersion and chemical/physical losses dilute SO2, leading to the disappearance of 
the plume initially seen by OMI. On the other hand, OMI is more sensitive to SO2 
lofted to higher altitudes. Low-level clouds may enhance the OMI sensitivity to SO2 
above them, but can block the signal below them. In short, satellite sensors can 
picture the plume daily, but often only partially. To correct for the part of the plume 
not detected by the satellites, I define the dispersion weighting factor (DWF). Taking 
April 6 as an example, I first select a polygon box (P1) covering a good part of the 
plume with strong trajectory-projected SO2 signal (e.g., Figure 5.11d). The DWF for 
this box (DWFP1) is the ratio between trajectory-projected SO2 mass within P1 (MP1, 
TRAJ) and the initial total SO2 of the plume (M1), which is calculated from the OMI 
retrievals on April 5 (rectangular box, Figure 5.11a) and kept constant in trajectory 
calculations. Weighting the OMI-retrieved SO2 mass within P1 (MP1,SAT) with DWFP1 
gives an estimate of the total SO2 mass on April 6 (M2, P1): 
1
,1
1,2
P
SATP
P DWF
MM =   (5.11) 
1
,1
1 M
MDWF TRAJP
P =   (5.12) 
In light of the uncertainties in trajectories and possible mismatch with satellite 
observations, five polygon boxes (P1 ? P5) of different size and shape are selected 
each day for April 6 and 7 (see Appendix III for the polygon box selections). Each 
 
 159 
 
has a corresponding estimated total SO2 mass (e.g., M2, P1 ? M2, P5). The average is 
taken as the estimated total SO2 mass, for example on April 6: 
5/
5
1
,22 ?
=
=
i
PiMM     (5.13) 
Pixels or grid cells with retrieved SO2 under a certain threshold (presumed 
detection limit) are considered undetectable by OMI, and are assumed, on average, to 
contain half of the threshold column SO2. Three threshold values (0.2, 0.5 and 1.0 DU) 
are chosen to approximate the variety of factors that might influence the OMI 
sensitivity to SO2. Shown in Figure 5.16 are the estimated total SO2 of the plume 
from April 5 to April 7, as well as the average AOD near the plume core (rectangular 
box in Figure 5.11a on April 5, various polygon boxes in Appendix III on April 6 and 
7). The ~100000 tonne of SO2 originally measured over the rectangular box in Figure 
5.11a (35-49?N, 117-134?E) represents roughly 0.05% of the annual global emissions 
[IPCC, 2001]. As the SO2 mass decays with time, the average AOD near the main 
body of the plume increases. The polygon boxes on April 6 and 7 mainly cover the 
core part of the plume while the rectangular box on April 5 encircles both polluted 
and clean areas. This sampling difference can explain part of the AOD increase, but 
not all, as suggested by the rising AOD/SO2 ratio in our sampling areas (~0.3 to ~0.5 
in the first two days). In fact, the average column SO2 in two of the five polygon 
boxes on April 6 is greater than the plume-wide mean on April 5, reflecting the 
sampling difference. At least part of the growth in AOD with time is likely caused by 
introduction of aerosols and/or enhancement of aerosol extinction in the plume. 
 
 160 
 
 
Figure 5.16. Total SO2 mass of the plume and average AOD near the main body of 
the plume (within polygon boxes in Figure 5.12). Error bars represent the standard 
deviation of estimates using different polygon boxes. Note that the estimates of SO2 
mass with different assumptions are diverse on April 7, but much less so on April 6.  
 
5.2.6 Discussion  
Aerosols and SO2 outside of the rectangular box on April 5 (initial plume area) 
might be carried into the polygon boxes on April 6 and 7, making an ?outside 
contribution? to the estimated SO2 mass and average AOD. The hygroscopic growth 
of aerosols may also enhance AOD. I conducted computational experiments to 
address these two factors. Forward trajectories similar to those described in section 
5.2.2 are calculated over a larger area (30-49?N, 110-144?E). The hygroscopic growth 
 
 161 
 
effect on AOD is estimated with NCEP reanalysis RH along trajectories and various 
growth factors (?): 
??
?????? ??= 21 11
1
2
RH
RH
AOD
AOD
RH
RH     (5.14) 
where RH1 and RH2 are the relative humidity at two different moments along a 
trajectory. Growth factor ? when equal to zero represents aerosols with no 
hygroscopic growth; ? of 0.5 corresponds to extremely hygroscopic aerosols. For 
reference, ? is about 0.2-0.3 in summer months over the NE U.S. [Taubman et al., 
2004], where aerosols mainly consist of hygroscopic fine particles. Dust-dominated 
particles in this study are likely less hygroscopic. Figure 5.17 shows the trajectory 
projected AOD on April 6, with outside contribution and hygroscopic growth 
included. Compared to Figure 5.12h, aerosols initially outside of the plume area can 
enhance AOD in polygon boxes on April 6, particularly over the area E of Japan 
(around 40?N, 150?E, Figure 5.17). The effects of hygroscopic growth, in contrast, 
are relatively small except over the land W of the Sea of Japan. The trajectory 
projected AOD including outside contribution and aerosol swelling is compared to 
MODIS retrievals in Figure 5.18. 
 
 162 
 
 
Figure 5.17. Trajectory-projected AOD on April 6 with contributions from outside of 
the initial plume area (rectangular box in Figure 5.12) and aerosol hygroscopic 
growth accounted for. Hygroscopic growth is calculated with NCEP reanalysis RH 
along forward trajectories and ? of 0-0.5 (cf. Eq. 5.14). 
 
 
 163 
 
 
Figure 5.18. Total SO2 mass of the plume with (SO2 all) and without (SO2 ini) 
outside contribution corrected for, and MODIS (blue triangle) and trajectory 
projected average AOD including outside contribution and effects of hygroscopic 
growth (red dot) near the core part of the plume.  
 
  To correct for the outside contribution to SO2 mass, I first derive a different 
dispersion weighting factor, for example on April 6 for polygon P1: 
1
,,1
,1 M
MDWF allTRAJP
allP =   (5.15) 
where M1 is the same as in Eq. 5.12, MP1,TRAJ,all is the projected SO2 mass using 
trajectories initiated from the larger area. Satellite determined SO2 mass in P1 
(MP1,SAT) is then weighted with the new DWF to estimate the total SO2 mass on this 
day (M2,P1,all): 
 
 164 
 
allP
SATP
allP DWF
MM
,1
,1
,1,2 =    (5.16) 
The assumption here is that SO2 in the whole region (in and outside of the rectangular 
box in Figure 5.11a) decays at the same rate. The resulting total SO2 mass of the 
plume (?SO2 all?, Figure 5.18) is generally smaller than that without the correction 
(?SO2 ini?, Figure 5.18), but the general decreasing trend in SO2 remains largely 
unchanged. 
From the decay in SO2 mass, I can approximate the overall SO2 lifetime in this 
plume. Assuming first order loss, the slope of the linear fit in Figure 5.19 gives the e-
folding time or lifetime of SO2. Our estimated SO2 lifetime varies from 1.4 to 3.6 
days, and is in the range of prior modeling studies [Berglen et al., 2004; Chin et al., 
2000; Koch et al., 1999]. Note that the calculated SO2 mass on April 7 is far more 
diverse due to diluted signal. Excluding this day, the ?best-guess? SO2 lifetime is 2-3 
days.  
 
 165 
 
 
Figure 5.19. Decay plot of SO2 in the plume from April 5 to 7, 2005. The slope of 
these two linear fits gives first-order removal rate of 0.42-0.61/day, which 
corresponds to an SO2 overall lifetime of 1.6-2.4 days. Decay plots of SO2 loading 
with other threshold values give slightly different estimates of SO2 lifetime (1-4 days). 
 
Supposing that all lost SO2 converts to ammonium sulfate (assumed specific 
scattering coefficient: 4.2 g/m2), the newly generated sulfate aerosols may add to 
~0.1-0.2 average AOD over the polygon boxes on April 6, and ~0.2-0.4 in AOD on 
April 7. This is probably an overestimate, since a fraction of SO2 can be removed by 
dry deposition; and the interaction with dust may produce larger particles that scatter 
light less efficiently. Little difference is found between the observed and projected 
AOD on April 6 (Figure 5.18); loss of dust from the plume probably cancels out the 
introduction of secondary aerosols. On the 7th, the difference between trajectory-
projected and observed AOD is about 0.3. Satellites cannot discern different aerosol 
species; the GOCART simulation (Figure 5.20) suggests that removal of dust and 
increase of sulfate are the most important processes controlling the change of aerosol 
 
 166 
 
loading in this plume. Note that GOCART has a coarse horizontal resolution as 
compared to satellite observations; the boxes in Figures 5.13-15 indicate the domains 
for averaging GOCART AOD. They correspond roughly to the sampling areas 
(rectangular and polygon boxes) in Figures 5.12, but are not exact matches. In 
summary, in this case study two different satellite sensors OMI and MODIS 
successfully tracked this April 5 pollutant plume on its transport pathway, and 
measured the changes in its chemical properties. The loss of SO2 and generation of 
sulfate appear to be detectable from space, although the uncertainty is high. A 
chemical transport model run could reduce the uncertainty. The method presented 
here could be employed to study other pollution transport episodes if SO2 and aerosol 
loadings are heavy enough, but for each episode AMF correction to satellite product 
is needed for qualitative application of the data. 
 
Figure 5.20. AOD of different aerosol species near the main body of the plume 
(boxes in Figures 5.13-5.15) in GOCART output from April 5 to 7, 2005. 
 
 
 167 
 
Clouds may contaminate MODIS AOD retrievals, and might produce more 
marked errors on April 6 and 7, when cloud cover was greater. In a simple test, I 
discard MODIS pixels with high AOD (> 95th percentile) but small AE (< 0.2) to 
remove possible cloud-contaminated data. This results in lower average AOD of 0.61, 
0.88, and 1.03 on the three days, but does not change the overall trend of AOD. 
Adjusting the threshold of AOD and AE to filter out cloud-contaminated pixels gives 
slightly different but qualitatively consistent results. 
 
5.2.7 Summary  
In this section I present a case study tracking and characterizing a pollutant 
plume from East Asia using satellite measurements. MODIS and OMI, two sensors of 
the ?A-train? constellation successfully captured the regional-scale plume on three 
consecutive days along its transport pathway. The estimated SO2 overall lifetime (e-
folding time), derived from the decay of the SO2 loading of the plume assuming first-
order loss, is about 1-4 day and in line with previous modeling studies. Conversion 
from SO2 to sulfate aerosols can generate AOD signal up to a few tenths, strong 
enough to be detected from space. Satellite retrieved average AOD near the core part 
of the plume is very close to the trajectory projected value (hygroscopic growth and 
outside contribution accounted for, but no loss of aerosols) over N Japan on April 6; 
to this point the introduction of secondary aerosols and loss of primary aerosols (dust) 
may have canceled out. MODIS AOD is greater than the trajectory AOD over NW 
Pacific on April 7. The GOCART model demonstrates that dust AOD decreases while 
sulfate AOD grows as the plume moves away from its source in N China.  
 
 168 
 
Substantial uncertainties are associated with the quantitative application of 
satellite data. Errors in satellite retrievals stem from a priori assumptions and 
interferences. Cloud fraction changes quite a bit, and the plume becomes under-
sampled, particularly for aerosols over the NW Pacific on April 7. Satellite sensitivity 
to SO2 (and other trace gases) is controlled by various factors such as vertical 
distribution and surface reflectance, which would change as the plume evolves. In 
addition, SO2 decays with time. Quantification of the temporal change of the SO2 
mass is therefore subject to sampling inconsistency. In this case study, errors of 
trajectory calculation grow over time with the distance traveled by the trajectories. 
Applying a few aircraft profiles to a large region is also probably an 
oversimplification. Nonetheless, as a first step, this case study provides a cost-
effective alternative supplementing intensive semi-Lagrangian aircraft campaigns, 
and showcases the great potential in the fusion of data from multiple satellite sensors. 
With CO, NO2, and cloud information from satellites, more can be done in the future 
to investigate the transport and evolution of pollution plumes and their interactions 
with weather and climate. 
 
 169 
 
Chapter 6:  Conclusions 
 
6.1 Summary of Results 
Measurements of trace gases (CO, NOy, SO2, and O3), aerosol optical 
properties (scattering and absorption), and aerosol chemical composition were made 
in Xianghe, a rural site in N China about 70 km ESE of Beijing, in March 2005 under 
the auspices of the U.S.-China joint research project EAST-AIRE. Located in an area 
of mixed agriculture and light industry, in proximity to two mega cities (Beijing and 
Tianjin), this non-urban site featured high but variable pollutant loadings (CO: ~1 
ppm, SO2: ~20 ppb, NOy: ~20 ppb, aerosol scattering: ~ 400 Mm-1, aerosol absorption: 
~50 Mm-1, sulfate: ~8 ?g/m3, nitrate: ~10 ?g/m3) during the experiment. Springtime 
mid-latitude wave cyclones and the associated fronts, passing the region every 4-5 
days, are the major mechanism behind the accumulation and removal of pollutants at 
the ground level. Strong northerly winds brought cold postfrontal air with little 
anthropogenic pollution but sometimes dust, while relatively warm and stagnant 
prefrontal conditions built up pollutants. 
The observed ratios between different pollutants were compared to some 
contemporary emission inventories for China. The measurement-derived CO/SO2 
ratio (~35) is at least a factor of two greater than the national (10-17, [Streets et al., 
2003, 2006]) and regional (~8, [X. Wang et al., 2005]) inventory ratios, suggesting 
substantial uncertainties in currently estimated CO (or less likely: SO2) emissions 
from China. Representing a large area with measurements at a single location might 
 
 170 
 
be problematic, but the observed CO/SO2 ratio is consistently greater than Streets et 
al. [2003, 2006], when air was from the populated and polluted region to the 
southwest of Xianghe. The inventory CO/SO2 ratio in metropolitan areas of China is 
in general very small (~10). Some other factors, for example the temporal mismatch 
between observations and inventories, pollutant release height, and chemical 
reactions may also contribute to our reported discrepancy, although the latter two are 
shown unlikely to contribute substantially in this study. To understand pollutant 
emissions from China better, especially those from the fast-developing eastern part, 
we need to make extensive observations near emission source regions and 
additionally, determine emission factors of various sources such as small boilers and 
automobiles [Streets et al., 2006]. 
Trace gas emissions from a nearby hand-fed coal-fired boiler were studied. 
The derived SO2 (15.4 g SO2/kg coal) and NOx (0.2 g NO2/kg coal) emission factors 
fall in the range of the factors applied in some emission inventories [Streets et al., 
2003]. CO emissions, however, depend on combustion conditions; the CO emission 
factor from this study is at least a factor of two lower than the inventory values 
[Streets et al., 2003, 2006]. Thus if this boiler is representative, small-scale coal 
boilers cannot explain the possible low bias in inventory CO; emissions from mobile 
sources need to be further investigated, but are beyond the scope of this research. In 
our ballpark estimate, similar small boilers altogether sum to less than 1% of China?s 
annual gaseous pollutant emissions, but can cause severe local air quality issues, as 
these boilers with their low stacks are often located in densely populated communities, 
run under cold and stagnant conditions, and lack emission control measures. 
 
 171 
 
Consistent with previous studies in eastern China, aerosols near the surface at 
the Xianghe site are found to be light-absorbing. The single scattering albedo 
calculated from the aerosol scattering and absorption data is 0.81 and 0.86 in the 
morning and in the afternoon, respectively. Assuming that black carbon is the 
dominant absorbing compound in the visible (specific absorption coefficient: 10 
m2/g), from aerosol absorption/CO correlation, I estimated the black carbon emissions 
from China at 1300-2600 Gg/yr, comparable to that from inventory [Streets et al., 
2003] but with large and ill-constrained uncertainties. 
A considerable fraction of aerosol particles in Xianghe during the experiment 
were of soil origin, as indicated by the loadings of some crustal elements (e.g., Al: 9.6 
?g/m3; Fe: 6.5 ?g/m3, Ca: 9.8 ?g/m3). However high levels of Ca, Fe, and K under 
stagnant conditions hint at strong sources related to anthropogenic activities, for 
example construction and agricultural burning. Peak concentrations of Al on the other 
hand were mostly seen in the windy postfrontal environment, likely associated with 
transported or locally emitted dust. Compared to the Earth?s crust, some trace 
elements (e.g., Zn, Cu, Ag, Pb, and As) in aerosol samples were enriched. Synoptic 
variations of these pollutant tracer elements were in phase with CO, implying their 
anthropogenic origins. Despite prohibition of leaded gasoline for an extended period, 
substantial Pb sources still exist in the area as evidenced by CO-correlated and highly 
enriched Pb (enrichment factor: 133). Chloride ion was enriched compared to sea 
water and in general more so when SO2 was high, probably from heavy coal burning. 
Our result that nitrate was more abundant than sulfate at the Xianghe site is in 
contradiction with a number of earlier studies conducted in the region. Size range of 
 
 172 
 
aerosol sampling, interference of trace gases, analysis technique, sampling time and 
location are among the factors that complicate comparison of different aerosol 
composition data sets. Recent expansion in NOx emission sources (large point sources 
and vehicle fleet) and progress in SO2 emission abatement (e.g., replace coal with 
low-S fuels in downtown Beijing) might have increased the nitrate-sulfate ratio in the 
region. The ratio shows apparent synoptic variations: high in the warm and polluted 
prefrontal atmosphere, while low under cold and clean postfrontal conditions. NOx-
to-nitrate conversion appears temperature-dependent and probably is mainly through 
gas-phase reactions. The conversion from SO2 to sulfate, in contrast, is more 
complete in air from the dry and cold north, where gas-phase reactions are likely 
suppressed. An analysis combining backward trajectories, assimilated meteorological 
data, and satellite-retrieved cloud properties suggests that during the experiment 
about 70% of the aerosol samples with relatively high SO2 conversion ratio may have 
a history of potential cloud interaction. The in-cloud processes provide a fast pathway 
for SO2 oxidization. The interaction between dust and sulfur compounds, widely 
studied in earlier research, may also contribute to the reaction of SO2, but only in 
~40% of the samples. Ground measurements in coordination with the EAST-AIRE 
Xianghe experiment carried out in Hedo, Japan, ~1000 km to the east, recorded 
surges in sulfate aerosols as wave cyclones passing the island brought continental 
flows from N China [Takami et al., 2007] behind cold fronts. SO2 in these postfrontal 
low-level transport events was almost completely reacted away, a fact that can only 
be explained with in-cloud processing [Takami et al., 2007]. Together, the two studies 
demonstrate that, in contrast to some previous conclusions [e.g., Yao et al., 2003], 
 
 173 
 
under certain circumstances in-cloud processes can be important in the springtime 
sulfur chemistry over East Asia, both upwind and downwind of the strong emission 
sources in China. Chemical transport models need to properly treat clouds to 
accurately simulate the sulfur cycle in this part of the world. 
The EAST-AIRE aircraft campaign conducted in April 2005 over the area 
near Shenyang (population: six million), the capital of China?s industrialized 
northeastern province of Liaoning and ~600 km NE of Beijing, further investigated 
the interaction between mid-latitude wave cyclones and air pollution. The change in 
air quality in response to weather conditions during the aircraft experiment was 
similar to that in Xianghe. Within the PBL, prefrontal flights generally found great 
levels (~20 ppb SO2) of anthropogenic pollutants associated with surface winds from 
the populated and polluted areas to the south; postfrontal flights on the other hand 
measured little SO2 but sometimes dust (~60 Mm-1 aerosol scattering) in northerly 
winds. I contrast the results from this experiment to a chemical climatology amassed 
through hundreds of flights over the NE U.S. mainly on smoggy summer days 
[Taubman, 2004]. Pollutant levels over China are greater than over the U.S. at almost 
every altitude, indicating the potential large-scale impacts of pollutants from China. 
The overall particle size of aerosols sampled over China is larger than over the U.S., 
probably due to dust emissions, although line-losses of coarse particles in sampling 
could be large in both studies (in particular over China). 
The prefrontal flight on April 5 recorded pollutants well above background 
levels (1-3 ppb SO2, 300 ppb CO, 50 Mm-1 aerosol scattering) in the free troposphere, 
at 2000-3000 m altitude. Backward trajectory analysis shows little evidence of lofting 
 
 174 
 
by a warm conveyor belt over the continent on previous days. However, satellites 
spotted some non-precipitating convective clouds near the industrial cities of Hohhot 
and Baotou roughly 1000 km to the west, along the backward trajectory and ~24 hr 
prior to the flight. This upwind convective process on April 4 might have lifted 
pollutants to the free troposphere, where strong westerly winds could carry them 
downwind in the next ~1 day to our area of flight. Pollutants lifted above the PBL 
could potentially travel over long distances, turning a local air pollution issue into a 
large-scale problem. Spring (MAM) cloudiness over N China is comparable to over 
the NE U.S., but it rains much less in N China. Dry (non-precipitating) convective 
processes over the emission source regions of China in spring can loft and export 
pollutants with great efficiency, but need further investigation and can be a great 
challenge to numerical models. 
OMI and MODIS satellite sensors make it possible to look into the April 5 air 
pollution episode at a regional scale, and to track the transport and evolution of the 
pollution plume from China to as far as the northwestern Pacific on the following few 
days. The total SO2 loading of the plume decayed with time. The overall SO2 lifetime 
(e-folding time) based on the decay and assuming first-order loss, is 1-4 day, in the 
range of previous modeling studies. SO2 to sulfate conversion would add up to ~0.1-
0.2 and ~0.2-0.4 in AOD near the core part of the plume on April 6 and 7, 
respectively. The signal is however not easily detected in this dust-laden plume, as 
removal of dust and introduction of sulfate may cancel out. The GOCART model 
confirms that reduction of dust and formation of sulfate are the two major processes 
changing the aerosol loadings of this particular plume. This analysis showcases the 
 
 175 
 
potential in combining different satellite sensors to trace major transport events and 
pollutant evolution, despite large uncertainties in quantitative application of satellite 
data. 
 
6.2 Recommendations 
In this dissertation I try to illustrate a few aspects of China?s air pollution 
problem, by analyzing a data set collected during the first intensive field campaign of 
EAST-AIRE. This joint China-U.S. project focuses on measurements over inland 
China, close to some of the most populated and polluted regions in the world. 
Compared to experiments conducted downwind of China over the Pacific [e.g., 
Carmichael et al., 2003], measurements near the emission source areas benefit from 
stronger signal and less chemical/physical losses, and can be directly used to evaluate 
emission inventories. However, as has been repeatedly stressed, this pilot study 
covers only a few locations over a short period of time, while emission inventories 
are in most cases compiled for a large region over much longer time (e.g., one year). 
Studying pollutant emissions from a country as massive and as diverse (e.g., 
modernized mega cities Beijing and Shanghai vs. rural areas) as China justifies long-
term observations at much more representative locations. Combined with in-situ 
measurements, chemical transport models in forward, inverse, or assimilation modes 
can help estimate emissions from China in a more quantitative way, accounting for 
factors such as chemical reactions and pollutant release heights. One challenge in 
comparing modeling results to observations near emission sources would be to 
improve the spatial and temporal resolution of both models and emission inventories. 
 
 176 
 
It was found in this study that some current emission inventories probably 
underestimate CO emissions from some of the largest metropolitan areas of China, 
for example, Beijing, Tianjin, Shanghai, and Guangzhou. A small coal-fired boiler 
studied appears to be more efficient than previously thought, but we probably need to 
test the results by sampling more similar boilers. A soaring vehicle fleet and 
deteriorating traffic conditions in these mega cities can lead to vast growth in CO 
emissions. To resolve this mystery of the ?missing CO source?, experiments directly 
determining emission factors of mobile sources in Chinese cities are needed. 
The research flights conducted around the northeastern city of Shenyang are 
among some of the first efforts to measure the vertical distribution of pollutants over 
inland China. Some interesting comparisons have been made between results from the 
China flights and the chemical climatology of the northeastern U.S., but more flights 
in multiple seasons over China are necessary to build up a chemical climatology for 
more meaningful comparisons. The April 5 flight discloses the potential role that 
upwind dry convection may play in lofting and transporting pollutants in northern 
China?s arid spring. Flights in proximity to similar non-precipitating convective 
clouds may provide direct evidence and more details for this lofting mechanism, but 
require careful planning and probably forecast modeling support. Additional profiling 
flights and chemical transport modeling are essential to determine the relative 
importance of convection and the warm conveyor belt lofting in long-range transport 
out of East Asia. Simulating dry convection processes may be particularly 
challenging, since the numerical model output cannot be graded against observed 
 
 177 
 
precipitation. Instead, satellite-retrieved cloud properties may be useful for evaluating 
model-simulated convection. 
In two case studies, I make use of some recently available satellite data to 
study the evolution of pollutants transported into and out of northern China by mid-
latitude wave cyclones. These satellite products provide information concerning 
weather and pollution at a regional scale and on a daily basis. Substantial 
uncertainties restrict their quantitative applications, but there is room for 
improvement. Multi-layer cloud retrieving algorithms [e.g., Chang and Li, 2005] and 
later-launched CloudSat can resolve the vertical distribution of clouds, which may 
allow for a more accurate space-based account of cloud-pollutant interactions. As 
shown in the study tracking the April 5 pollution plume, the OMI SO2 and AI 
products can be powerful tools tracing pollutant transport, and can be employed to 
identify pollution-influenced areas over the oceans in future studies on aerosol 
radiative forcing. The OMI SO2 retrieval algorithm needs to be improved to estimate 
quantitatively anthropogenic emissions or study pollution episodes. The correction for 
clouds in this study assumes a homogeneous cloud deck; future studies will apply a 
cell-by-cell cloud correction scheme utilizing OMI-retrieved cloud information. 
Model simulated SO2 and aerosol vertical profiles may help improve the OMI SO2 
algorithm as well. The ratio between satellite-retrieved relatively long-lived (e.g., CO) 
and short-lived (e.g., SO2) gases may change as plumes move away from their 
sources. Using this differential decay of long- and short-lived gases, we may be able 
statistically evaluate the lifetime of short-lived gas. But enhancing sensitivity to 
 
 178 
 
pollutants in the lower atmosphere especially in the boundary layer will be key to 
quantitatively applying satellite data in air pollution research. 
In summary, research in this dissertation constitutes a first step towards better 
understanding China?s air pollution issue and its large-scale impacts. Future efforts 
integrating in-situ observations, satellite retrievals, and numerical simulations will be 
crucial to appreciate and ultimately mitigate air pollution in China, a problem with 
roots in China but global effects.               
 
 179 
 
Appendix I Surface Weather Maps during the Aircraft 
Campaign 
 
Appendix I.1. Surface weather analysis of East Asia at 00 UTC on April 1, 2005. 
(Source: Korea Meteorological Administration). 
 
 
Appendix I.2. Surface weather analysis of East Asia at 00 UTC on April 2, 2005. 
 
 180 
 
 
Appendix I.3. Surface weather analysis of East Asia at 00 UTC on April 3, 2005. 
 
 
Appendix I.4. Surface weather analysis of East Asia at 00 UTC on April 4, 2005. 
 
 181 
 
 
Appendix I.5. Surface weather analysis of East Asia at 00 UTC on April 5, 2005. 
 
 
Appendix I.6. Surface weather analysis of East Asia at 00 UTC on April 6, 2005. 
 
 182 
 
 
Appendix I.7. Surface weather analysis of East Asia at 00 UTC on April 7, 2005. 
 
 
Appendix I.8. Surface weather analysis of East Asia at 00 UTC on April 8, 2005. 
 
 183 
 
 
Appendix I.9. Surface weather analysis of East Asia at 00 UTC on April 9, 2005. 
 
 
Appendix I.10. Surface weather analysis of East Asia at 00 UTC on April 10, 2005. 
 
 184 
 
 
Appendix I.11. Surface weather analysis of East Asia at 00 UTC on April 11, 2005. 
 
 
Appendix I.12. Surface weather analysis of East Asia at 00 UTC on April 12, 2005. 
 
 185 
 
Appendix II: Vertical Profiles of Pollutants during the Aircraft 
Campaign 
 
Appendix II.1. Vertical profiles of RH, aerosol scattering, and wet-bulb potential 
temperature north of Shenyang on April 5, 2005. 
 
 
Appendix II.2. Vertical profiles of RH, SO2, and wet-bulb potential temperature 
north of Shenyang on April 5, 2005. 
 
 186 
 
 
Appendix II.3. Vertical profiles of SO2, temperature, aerosol scattering coefficients 
(550 nm) and relative humidity on April 10, 2005.  Panels a) and b) present data 
obtained from a spiral ascend north of Shenyang (42.48?N 123.61?E).  Panels c) and 
d) present data obtained from a spiral ascend south of Shenyang (41.37 ?N 122.61 ?E).  
Panels e) and f) present data obtained as when descending to the airport (41.64 ?N 
123.49 ?E). 
 
 
 187 
 
 
Appendix II.4. Vertical profiles of SO2, temperature, aerosol scattering coefficients 
(550 nm) and relative humidity on April 11, 2005.  Panels a) and b) present data 
obtained from a spiral ascend south of Shenyang (41.35?N 122.65?E).  Panels c) and 
d) present data obtained from a spiral ascend south of Shenyang (42.45 ?N 123.61 ?E).  
Panels e) and f) present data obtained when descending to the airport (41.64 ?N 
123.49 ?E). 
 
 188 
 
 Appendix II.5. Vertical profiles of SO
2, temperature, aerosol scattering coefficients 
(550 nm) and relative humidity on April 12, 2005. Data were obtained as the aircraft 
was descending to the airport (41.64 ?N 123.49 ?E). 
 
 189 
 
Appendix III Polygon Boxes Used to Calculate the Total SO2 
Loading of the Plume 
 
Appendix III.1. Polygon boxes used to sample pollution plume on April 6. Color 
represents trajectory projected SO2 loading (DU). 
 Appendix III.2. Polygon boxes used to sample pollution plume on April 7. Color 
represents trajectory projected SO2 loading (DU). 
 
 190 
 
Bibliography 
 
Akimoto, H., T. Ohara, J. Kurokawa, and N. Horii (2006), Verification of energy 
consumption in China during 1996?2003 by using satellite observational data, 
Atmos. Environ., 40, 7663-7667. 
Allen, D., K. Pickering, and M. Fox-Rabinovitz (2004), Evaluation of pollutant 
outflow and CO sources during TRACE-P using model-calculated, aircraft-
based, and Measurements of Pollution in the Troposphere (MOPITT)-derived 
CO concentrations, J. Geophys. Res., 109, D15S03, doi:10.1029/2003JD004250. 
Anderson, T. L., et al. (1996), Performance characteristics of a high-sensitivity, three-
wavelength, total scatter/backscatter nephelometer, J. Atmos. Oceanic 
Technol., 13967-13986. 
Anderson, T. L., and J. A. Ogren (1998), Determining aerosol radiative properties 
using the TSI 3563 Integrating Nephelometer, Aerosol Sci. Technol., 29, 57-
69. 
Andreae, M. O., and P. Merlet (2001), Emission of trace gases and aerosols from 
biomass burning, Global Biogeochemical Cycles, 15, 955?966. 
Andreae, M. O., O. Schmid, H. Yang, D. Chand, J. Yu, Y. Zhang, and L. Zeng 
(2005), Aerosol optical and chemical properties in urban Guangzhou, China, 
Eos Trans. AGU, 86(18), Jt. Assem. Suppl., Abstract A21A-23. 
Arellano, A. F., P. S. Kasibhatla, L. Giglio, G. R. van der Werf, and J. T. Randerson 
(2004), Top-down estimates of global CO sources using MOPITT 
measurements, Geophys. Res. Lett., 31, L01104, doi:10.1029/2003GL018609. 
Arakawa, H. (1969), Climates of Northern and Eastern Asia, Elsevier, New York, 
 
 191 
 
USA. 
Arimoto, R., et al. (2006), Characterization of Asian dust during ACE-ASIA, Global 
and Planetary Change, 52, 23-56. 
Aunan, K., T. K. Berntsen, and H. M. Seip (2000), Surface ozone in China and its 
possible impact on agricultural crop yields, Ambio, 29(6), 294-301. 
Barnard, J. C., E. I. Kassianov, T. P. Ackerman, S. Frey, K. Johnson, B. Zuberi, L. T. 
Molina, M. J. Molina, J. S. Gaffney, and N. A. Marley (2005), Measurements 
of black carbon specific absorption in the Mexico City metropolitan area 
during the MCMA 2003 field campaign, Atmos. Chem. Phys. Discuss., 5, 
4083-4113. 
Beijing EPA (2002), Integrated emission standard of boiler pollutants, DB11/139-
2002, Beijing, China (in Chinese). 
Bergin, M., et al. (2001), Aerosol radiative, physical, and chemical properties in 
Beijing during June 1999, J. Geophys. Res., 106(D16), 17969-17980. 
Berglen, T. F., T. K. Berntsen, I. S. A. Isaksen, J. K. Sundet (2004), A global model 
of the coupled sulfur/oxidant chemistry in the troposphere: The sulfur cycle, J. 
Geophys. Res., 109, D19310, doi:10.1029/2003JD003948. 
Berkowitz, C. M., P. H. Daum, C. W. Spicer, and K. M. Busness (1996), Synoptic 
patterns associated with the flux of excess ozone to the western North Atlantic, 
J. Geophys. Res., 101(D22), 28923-28934, 10.1029/95JD03274. 
Berntsen, T. K., S. Karlsd?ttir, and D. A. Jaffe (1999), Influence of Asian emissions 
on the composition of air reaching the North Western United States, Geophys. 
Res. Lett., 26(14), 2171?2174. 
 
 192 
 
Bertram, T. H., et al. (2007), Direct measurements of the convective recycling of the 
upper troposphere, Science, 315, 816-820. 
Bethan, S., G. Vaughan, C. Gerbig, A. Volz-Thomas, H. Richer, and D. A. Tiddeman 
(1998), Chemical air mass differences near fronts, J. Geophys. Res., 103(D11), 
13413-13434, 10.1029/98JD00535.  
Bey, I., D. J. Jacob, J. A. Logan, and R. M. Yantosca (2001), Asian chemical outflow 
to the Pacific in spring: Origins, pathways, and budgets, J. Geophys. Res., 
106(D19), 23097-23114, 10.1029/2001JD000806. 
Bond, T. C., T. L. Anderson, and D. Campbell (1999), Calibration and 
intercomparison of filter-based measurements of visible light absorption by 
aerosols, Aerosol Sci. Technol., 30, 582-600. 
Boylan, J. W., and A. G. Russell (2006), PM and light extinction metric, goals, and 
criteria for three-dimensional air quality models, Atmos. Environ., 40, 4946-4959. 
Brajer, V., and R. W. Mead (2003), Blue skies in Beijing? Looking at the Olympic 
effect, J. Environ. Development, 12(2), 239-263. 
Bryson, R. A., and F. K. Hare (1969), The climates of North America, Elsevier, New 
York, USA. 
Cao, J. J., S. C. Lee, K. F. Ho, X. Y. Zhang, S. C. Zou, K. Fung, J. C. Chow, and J. G. 
Watson (2003), Characteristics of carbonaceous aerosol in Pearl River Delta 
Region, China during 2001 winter period, Atmos. Environ., 37, 1451-1460. 
Carmichael, G. R., et al. (2003), Evaluating regional emission estimates using the 
TRACE-P observations, J. Geophys. Res., 108(D21), 8810, 
doi:10.1029/2002JD003116. 
 
 193 
 
Chameides, W. L, P. S. Kasibhatla, J. Yienger, H. Levy II (1994) Growth of 
continental-scale metro-agro-plexes, regional ozone pollution, and world food 
production, Science, 264, 74?77. 
Chameides, W. L., et al. (1999a), Is ozone pollution affecting crop yields in China?, 
Geophys. Res. Lett., 26(7), 867-870, 10.1029/1999GL900068. 
Chameides, W. L., et al. (1999b), Case study of the effects of atmospheric aerosols 
and regional haze on agriculture: An opportunity to enhance crop yields in China 
through emission controls?, Proceedings of the National Academy of Sciences, 
96, 13626-13633. 
Chan, C. Y., X. D. Xu, Y. S. Li, K. H. Wong, G. A. Ding, L. Y. Chan, and X. H. 
Cheng (2005), Characteristics of vertical profiles and sources of PM2.5, PM10, 
and carbonaceous species in Beijing, Atmos. Environ., 39, 5113-5124. 
Chang, F.L., and Z. Li (2005), A new method for detection of cirrus overlapping 
water clouds and determination of their optical properties, J. Atmos. Sci., 62, 
3993-4009. 
Charlson, R. J., S. E. Schwartz, J. M. Hales, R. D. Cess, J. A. Coakley, J. E. Hansen, 
and D. J. Hofmann (1992), Climate forcing by anthropogenic aerosols, 
Science, 255, 423-430. 
Chatfield, R. B., and P. J. Crutzen (1984), Sulfur dioxide in remote oceanic air: Cloud 
transport of reactive precursors, J. Geophys. Res., 89(D5), 7111?7132. 
Chen, J., F. Wei, C. Zheng, Y. Wu, and D. C. Adriano (1991), Background 
concentrations of elements in soils of China, Water Air and Soil Pollution, 57/58, 
699-712. 
 
 194 
 
Chen, L., Y. Duan, Y. Zhuo, L. Yang, L. Zhang, X. Yang, Q. Yao, Y. Jiang, and X. 
Xu (2007), Mercury transformation across particulate control services in six 
power plants of China: The co-effect of chlorine and ash composition, Fuel, 86, 
603-610. 
 Chen, L. W. A. (2002), Urban fine particulate matter: Chemical composition and 
possible origins, Ph.D. Dissertation, Chemical Physics Program, University of 
Maryland. 
Chen, L-W. A., B. G. Doddridge, R. R. Dickerson, J. C. Chow, P. K. Mueller, J. 
Quinn, and W. A. Butler (2001), Seasonal variations in elemental carbon 
aerosol, carbon monoxide and sulfur dioxide: Implications for sources, 
Geophys. Res. Lett., 28(9), 1711-1714. 
Chen, L. W. A., B. G. Doddridge, R. R. Dickerson, J. C. Chow, and R. C. Henry 
(2002), Origins of fine aerosol mass in the Baltimore-Washington corridor: 
Implications from observation, factor analysis, and ensemble air parcel back 
trajectories, Atmos. Environ., 36, 4541-4554. 
Chen, T-B. et al. (2005), Assessment of heavy mental pollution in surface soils of 
urban parks in Beijing, China, Chemosphere, 60, 542-551. 
Chen, Y., G. Zhi, Y. Feng, J. Fu, J. Feng, G. Sheng, and B. R. T. Simoneit (2006), 
Measurements of emission factors for primary carbonaceous particles from 
residential raw-coal combustion in China, Geophys. Res. Lett., 33, L20815, 
doi:10.1029/2006GL026966. 
 
 195 
 
Chin, M., D. Jacob, G. Gardner, M. Foreman-Fowler, P. Spiro, and D. Savoie (1996), 
A global three-dimensional model of tropospheric sulfate, J. Geophys. Res., 
101(D13), 18667-18690. 
Chin, M., D. Savoie, B. Huebert, A. Bandy, D. Thornton, T. Bates, P. Quinn, E. 
Saltzman, and W. De Bruyn (2000), Atmospheric sulfur cycle simulated in the 
global model GOCART: Comparison with field observations and regional 
budgets, J. Geophys. Res., 105(D20), 24689-24712. 
Chinese CBS (2005), China statistical yearbook 2005, China Statistics Press, Beijing, 
China. 
Chinese EPA (1999), Online document in Chinese available at their website: 
http://www.zhb.gov.cn/law/fg/gwyw/199809/t19980902_81940.htm 
Chinese EPA and Chinese CBS (2006), China green national accounting study report 
2004 (public version), Beijing, China. 
Cooper, O. R., J. L. Moody, D. D. Parrish, M. Trainer, T. B. Ryerson, J. S. Holloway, 
G. H?bler, F. C. Fehsenfeld, S. J. Oltmans, and M. J. Evans (2001), Trace gas 
signatures of the airstreams within North Atlantic cyclones: Case studies from 
the North Atlantic Regional Experiment (NARE?97) aircraft intensive, J. 
Geophys. Res., 106(D6), 5437-5456, 10.1029/2000JD900574. 
Cooper O. R., et al. (2004), A case study of transpacific warm conveyor belt 
transport: Influence of merging airstreams on trace gas import to North America, 
J. Geophys. Res., 109, D23S08, doi:10.1029/2003JD003624. 
Delene, D. J., and J. A. Ogren (2002), Variability of aerosol optical properties at four 
North American surface monitoring sites, J. Atmos. Sci., 59, 1135-1150. 
 
 196 
 
Dickerson, R. R., A. C. Delany, and A. F. Wartburg (1984), Further modification of a 
commercial NOx detector for high sensitivity, Rev. Sci. Instrum., 55(12), 
1995-1998. 
Dickerson, R. R., et al. (1987), Thunderstorms: An important mechanism in the 
transport of air pollutants, Science, 235, 460-465. 
Dickerson, R. R., and A. C. Delany (1988), Modification of a commercial gas filter 
correlation CO detector for enhanced sensitivity, J. Atmos. Oceanic Technol., 5, 
424-431. 
Dickerson, R. R., S. Kondragunta, G. Stenchikov, K. L. Civerolo, B. G. Doddridge, 
and B. N. Holben (1997), The impact of aerosols on solar ultraviolet radiation 
and photochemical smog, Science, 278, 827-830. 
Dickerson, R. R., M. O. Andreae, T. Campos, O. L. Mayol-Bracero, C. Neusuess, and 
D. G. Streets (2002), Analysis of black carbon and carbon monoxide observed 
over the Indian Ocean: Implications for emissions and photochemistry, J. 
Geophys. Res., 107, 8017, doi:10.1029/2001JD000501. 
Dickerson, R. R., et al. (2007), Aircraft observations of dust and pollutants over NE 
China: Insight into the meteorological mechanisms of long-range transport, J. 
Geophys. Res., 112, D24S90, doi:10.1029/2007JD008999. 
Donnell, E. A., D. J. Fish, E. M. Dicks, and A. J. Thorpe (2001), Mechanisms for 
pollutant transport between the boundary layer and the free troposphere, J. 
Geophys. Res., 106(D8), 7847-7856, 10.1029/2000JD900730. 
Draxler, R.R., and G.D. Rolph, (2003), HYSPLIT (HYbrid Single-Particle 
Lagrangian Integrated Trajectory) Model access via NOAA ARL READY 
 
 197 
 
Website htttp://www.arl.noaa.gov/ready/hysplit4.html). NOAA Air Resources 
Laboratory, Silver Spring, MD. 
Dzubay, T. G., R. K. Stevens, C. W. Lewis, D. H. Hern, W. J. Courtney, J. W. Tesch, 
and M. A. Mason (1982), Visibility and aerosol composition in Houston, Texas, 
Environ. Sci. Technol., 16, 514-525. 
Ehrman, S. H.; Pratsinis, and S. E.; Young (1992), J. R. Receptor modeling of the fine 
aerosol at a residential Los Angeles site, Atmos. Environ., 26B, 473-481. 
 Esler, J. G., P. H. Haynes, K. S. Law, H. Barjat, K. Dewey, J. Kent, S. Schmitgen, 
and N. Brough (2003), Transport and mixing between airmasses in cold frontal 
regions during Dynamics and Chemistry of Frontal Zones (DCFZ), J. Geophys. 
Res., 108 (D4), 4142, doi:10.1029/2001JD001494. 
Fairlie, T. D., D. J. Jacob, and R. J. Park (2007), The impact of transpacific transport 
of mineral dust in the United States, Atmos. Environ., in press. 
Fehsenfeld, F. C., et al. (1987), A ground-based intercomparison of NO, NOx, NOy 
measurement techniques, J. Geophys. Res., 92(12), 14710-14722. 
Fiore, A. M., D. J. Jacob, I. Bey, R. M. Yantosca, B. D. Field, A. C. Fusco, and J. G. 
Wilkinson (2002), Background ozone over the United States in summer: origin, 
trend, and contribution to pollution episodes, J. Geophys. Res., 107(D15), 
10.1029/2001JD000982. 
Fuelberg, H. E., C. M. Kiley, J. R. Hannan, D. J. Westberg, M. A. Avery, and R. E. 
Newell (2003), Meteorological conditions and transport pathways during the 
Transport and Chemical Evolution over the Pacific (TRACE-P) experiment, J. 
Geophys. Res., 108(D20), 8782, doi:10.1029/2002JD003092. 
 
 198 
 
Gao, Y., and J. Anderson (2001), Characteristics of Chinese aerosols determined by 
individual-particle analysis, J. Geophys. Res., 106(D16), 18037-18045. 
Gao, Y., J. R. Anderson, and X. Hua (2007), Dust characteristics over the North 
Pacific observed through shipboard measurements during the ACE-Asia 
experiment, Atmos. Environ., 41, 7907-7922. 
Ge, S., Z. Bai, W. Liu, T. Zhu, T. Wang, S. Qing, and J. Zhang (2001), Boiler 
briquette coal versus raw coal: Part 1. Stack gas emissions, J. Air Waste 
Manage. Assoc., 51, 524? 533. 
Ge, S., X. Xu, C.J. Chow, J. Watson, Q. Sheng, W. Liu, Z. Bai, T. Zhu, and J. Zhang 
(2004), Emissions of air pollutants from household stoves: Honeycomb coal vs. 
coal cake, Environ. Sci. Technol., 38, 4612-4618. 
Gregory, G. L., A. S. Bachmeier, D. R. Blake, B. G. Heikes, D. C. Thornton, A. R. 
Bandy, J. D. Bradshaw, and Y. Kondo (1996), Chemical signatures of aged 
Pacific marine air: Mixed layer and free troposphere as measured during PEM-
West A, J. Geophys. Res., 101(D1), 1727-1742, 10.1029/95JD00410. 
Gregory, G. L., J. T. Merrill, M. C. Shipham, D. R. Blake, G. W. Sachse, and H. B. 
Singh (1997), Chemical characteristics of tropospheric air over the Pacific 
Ocean as measured during PEM-West B: Relationship to Asian outflow and 
trajectory history, J. Geophys. Res., 102(D23), 28275-28286, 
10.1029/97JD02292. 
Guttikunda, S. K., Y. Tang, G. R. Carmichael, G. Kurata, L. Pan, D. G. Streets, J.-H. 
Woo, N. Thongboonchoo, and A. Fried (2005), Impacts of Asian megacity 
 
 199 
 
emissions on regional air quality during spring 2001, J. Geophys. Res., 110, 
D20301, doi:10.1029/2004JD004921. 
Hains, J. C., B. F. Taubman, A. M. Thompson, L. T. Marufu, J. W. Stehr, and R. R. 
Dickerson (2008), Origins of chemical pollution derived from Mid-Atlantic 
aircraft profiles using a clustering technique, Atmos. Environ., 42, 1727-1741. 
Hannan, J. R., H. E. Fuelberg, J. H. Crawford, G. W. Sachse, and D. R. Blake (2003), 
Role of wave cyclones in transporting boundary layer air to the free troposphere 
during the spring 2001 NASA/TRACE-P experiment, J. Geophys. Res., 
108(D20), 8785, doi:10.1029/2002JD003105. 
Hao, J., and L. Wang (2005), Improving urban air quality in China: Beijing case 
study, J. Air & Waste Manage. Assoc., 55, 1298-1305. 
Harris, A. R., and C. I. Davidson (2005), The role of resuspended soil in lead flows in 
the California South Coast Air Basin, Environ. Sci. Technol., 39, 7410-7415. 
Hatakeyama, S., A. Takami, W. Wang, and D. Tang (2005), Aerial observation of air 
pollutants and aerosols over Bo Hai, China, Atmos. Environ., 39, 5893?5898. 
Hartmann, D.L. (1994), Global Physical Climatology, Academic Press, San Diego, 
CA, USA, 411 pp. 
He, B. S., and C. H. Chen (2002), Energy ecological efficiency of coal fired plant in 
China, Energy Conver. Mgmt., 43, 2553-2567. 
He, K., F. Yang, Y. Ma, Q. Zhang, X. Yao, C. K. Chan, S. Cadle, T. Chan, and P. 
Mulawa (2001), The characteristics of PM2.5 in Beijing, China, Atmos. Environ., 
35, 4959-4970. 
 
 200 
 
He, K., H. Huo, and Q. Zhang (2002), Urban air pollution in China: Current status, 
characteristics, and progress, Annu. Rev. Energy Environ., 27, 397?431. 
He, P., P. Liu, X. Chen, and X. Chen (1999), Investigation on atmospheric pollution 
of Zhengzhou in 1998, J. Henan Medical Univ., 34, 64-67 (in Chinese). 
Heald, C. L., et al. (2003), Asian outflow and trans-Pacific transport of carbon 
monoxide and ozone pollution: An integrated satellite, aircraft, and model 
perspective, J. Geophys. Res., 108(D24), 4804, doi:10.1029/2003JD003507. 
Heald, C. L., D. J. Jacob, D. B. A. Jones, P. I. Palmer, J. A. Logan, D. G. Streets, G. 
W. Sachse, J. C. Gille, R. N. Hoffman, and T. Nehrkorn (2004), Comparative 
inverse analysis of satellite (MOPITT) and aircraft (TRACE-P) observations to 
estimate Asian sources of carbon monoxide, J. Geophys. Res., 109, D23306, 
doi:10.1029/2004JD005185. 
Holzer, M., I. G. McKendry, and D. A. Jaffe (2003), Springtime trans-Pacific 
atmospheric transport from east Asia: A transit-time-pdf approach, J. Geophys. 
Res., 108(D22), 4708, doi:10.1029/2003JD003558. 
Holzer, M., T. M. Hall, and R. B. Stull (2005), Seasonality and weather-driven 
variability of transpacific transport, J. Geophys. Res., 110, D23103, 
doi:10.1029/2005JD006261. 
Hu, M., F. Zhou, K. Shao, Y. Zhang, X. Tang, and J. Slanina (2002), Diurnal 
variations of aerosol chemical compounds and related gaseous pollutants in 
Beijing and Guangzhou, J. Environ. Sci. Health, A37, 479-488. 
 
 201 
 
Hudman, R.C., et al. (2004), Ozone production in transpacific Asian pollution plumes 
and implications for ozone air quality in California, J. Geophys. Res., 109, 
D23S10,10.1029/2004JD004974. 
Huebert, B. J., M.-X. Wang, and W.-X. Lu (1988), Atmospheric nitrate, sulfate, 
ammonium, and calcium concentrations in China, Tellus, 40B, 260-269. 
Huebert, B. J., T. Bates, P. B. Russell, G. Shi, Y. J. Kim, K. Kawamura, G. 
Carmichael, and T. Nakajima (2003), An overview of ACE-Asia: Strategies for 
quantifying the relationships between Asian aerosols and their climatic impacts, 
J. Geophys. Res., 108 (D23), 8633, doi:10.1029/2003JD003550. 
Husar, R. B., et al. (2001), Asian dust events of April 1998, J. Geophys. Res., 
106(D16), 18317-18330, 10.1029/2000JD900788. 
Igarashi, Y., Y. Sawa, K. Yoshioka, H. Takahashi, H. Matsueda, and Y. Dokiya 
(2006), Seasonal variations in SO2 plume transport over Japan: Observations at 
the summit of Mt. Fuji from winter to summer, Atmos. Environ., 40 (36), 7018-
7033. 
IPCC (2001), Climate change 2001: The scientific basis, Cambridge University Press, 
Cambridge, United Kingdom and New York, NY, USA, 881 pp. 
IPCC (2007), Climate change 2007: The scientific basis, Cambridge University Press, 
Cambridge, United Kingdom and New York, NY, USA, 996 pp. 
Jacob, D. J., J. H. Crawford, M. M. Kleb, V. S. Connors, R. J. Bendura, J. L. Raper, 
G. W. Sachse, J. C. Gille, L. Emmons, and C. L. Heald (2003), Transport and 
Chemical Evolution over the Pacific (TRACE-P) aircraft mission: Design, 
 
 202 
 
execution, and first results, J. Geophys. Res., 108(D20), 9000, 
doi:10.1029/2002JD003276. 
Jaffe, D. A., et al. (1999), Transport of Asian air pollution to North America, 
Geophys. Res. Lett., 26, 711-714. 
Jaffe, D. A., I. McKendry, T. Anderson, and H. Price (2003a), Six 'new' episodes of 
trans-Pacific transport of air pollutants, Atmos. Environ., 37, 391-404. 
Jaffe, D. A., D. Parrish, A. Goldstein, H. Price, and J. Harris (2003b), Increasing 
background ozone during spring on the west coast of North America, Geophys. 
Res. Lett., 30(12), 1613, doi: 10.1029/2003GL017024. 
Jaffe, D. A., E. Prestbo, P. Swartzendruber, P. Weiss-Penzias, S. Kato, A. Takami, S. 
Hatakeyama, and Y. Kajii (2005), Export of atmospheric mercury from Asia, 
Atmos. Environ., 39, 3029-3038. 
Jeong, M.-J., Z. Li, E. Andrews, and S.-C. Tsay (2007), Effect of aerosol 
humidification on the column aerosol optical thickness over the Atmospheric 
Radiation Measurement Southern Great Plains site, J. Geophys. Res., 112, 
D10202, doi:10.1029/2006JD007176. 
Kaneyasu, N., K. Takeuchi, M. Hayashi, S. -I. Fujita, I. Uno, and H. Sasaki (2000), 
Outflow patterns of pollutants from East Asia to the North Pacific in the winter 
monsoon, J. Geophys. Res., 105(D13), 17361-17378, 10.1029/2000JD900138. 
Karlsd?ttir, S., and I. S. A. Isaksen (2000), Changing Methane Lifetime: Possible 
Cause for Reduced Growth, Geophys. Res. Lett., 27(1), 93?96.  
 
 203 
 
Kiley, C. M., and H. E. Fuelberg (2006), An examination of summertime cyclone 
transport processes during Intercontinental Chemical Transport Experiment 
(INTEX-A), J. Geophys. Res., 111, D24S06, doi:10.1029/2006JD007115. 
Kim, B., J. Han, and S. Park (2001), Transport of SO2 and aerosol over the Yellow 
Sea, Atmos. Environ., 35, 727-737. 
Kim, J. Y., C. H. Song, Y. S. Ghim, J. G. Won, S. C. Yoon, G. R. Carmichael, and J.-
H. Woo (2006), An investigation on NH3 emissions and particulate NH4+?NO3? 
formation in East Asia, Atmos. Environ., 40, 2139-2150. 
Koch, D., D. Jacob, I. Tegen, D. Rind, and M. Chin (1999), Tropospheric sulfur 
simulation and sulfate direct radiative forcing in the Goddard Institute for Space 
Studies general circulation model, J. Geophys. Res., 104(D19), 23799-23822. 
Krotkov, N. A., et al. (2006), Band residual difference algorithm for retrieval of SO2 
from the Aura Ozone Monitoring Instrument (OMI), IEEE Transactions on 
Geoscience and Remote Sensing, 44, 1259-1266. 
Krotkov, N. A., et al. (2008), Validation of SO2 retrievals from the Ozone Monitoring 
Instrument over NE China, J. Geophys. Res., 113, D16S40, 
doi:10.1029/2007JD008818. 
Kurata, G., G. R. Carmichael, D. G. Streets, T. Kitada, Y. Tang, J.-H. Woo, and N. 
Thongboonchoo (2004), Relationships between emission sources and air mass 
characteristics in East Asia during the TRACE-P period, Atmos. Environ., 38, 
6977?6987. 
Larssen, T., et al. (2006), Acid rain in China, Environ. Sci. Technol, 40, 418-425. 
Lary, D., A. Lee, R. Toumi, M. Newchurch, M. Pirre, and J. Renard (1997), Carbon 
 
 204 
 
aerosols and atmospheric photochemistry, J. Geophys. Res., 102(D3), 3671-
3682. 
Levy, R. C., L. A. Remer, S. Mattoo, E. Vermote, and Y. J. Kaufman (2007a), 
Second-generation operational algorithm: Retrieval of aerosol properties over 
land from inversion of Moderate Resolution Imaging Spectroradiometer spectral 
reflectance, J. Geophys. Res., 112, D13211, doi:10.1029/2006JD007811. 
Levy, R. C., L. A. Remer, and O. Dubovic (2007b), Global aerosol optical models 
and application to MODIS aerosol retrieval over land, J. Geophys. Res., 112, 
D13210, doi:10.1029/2006JD007815. 
Li, C., L.T. Marufu, R. R. Dickerson, Z. Li, T. Wen, Y. Wang, P. Wang, H. Chen, 
and J. W. Stehr (2007), In-situ measurements of trace gases and aerosol optical 
properties at a rural site in northern China during EAST-AIRE IOP 2005. J. 
Geophys. Res., D22S04, doi:10.1029/2006JD007592. 
Li, Z., et al., (2007a), Aerosol optical properties and its radiative effects in northern 
China, J. Geophys. Res., 112, D22S01, doi:10.1029/2006JD007382. 
Li, Z., et al., (2007b), Preface to special section: Overview of the East Asian Study of 
Tropospheric Aerosols: an International Regional Experiment (EAST-AIRE), J. 
Geophys. Res. D22S00, doi:10.1029/2007JD008853.  
Liang, Q., L. Jaegle?, D. A. Jaffe, P. Weiss-Penzias, A. Heckman, and J. A. Snow 
(2004), Long-range transport of Asian pollution to the northeast Pacific: 
Seasonal variations and transport pathways of carbon monoxide, J. Geophys. 
Res., 109, D23S07,doi:10.1029/2003JD004402. 
 
 205 
 
Liang, Q., L. Jaegle?, and J. M. Wallace (2005), Meteorological indices for Asian 
outflow and transpacific transport on daily to interannual timescales, J. Geophys. 
Res., 110, D18308, doi:10.1029/2005JD005788. 
Lide, D. R. (1998), CRC handbook of chemistry and physics, 79th Edition, CRC Press 
LLC., Boca Raton, FL, USA, 2496 pp.  
Lin, C.-Y. C., D. J. Jacob, J. W. Munger, and A. M. Fiore (2000), Increasing 
background ozone in surface air over the United States, Geophys. Res. Lett., 
27(21), 3465-3468, doi:10.1029/2000GL011762. 
Liu, H., D. J. Jacob, I. Bey, R. M. Yantosca, B. N. Duncan, and G. W. Sachse (2003), 
Transport pathways for Asian pollution outflow over the Pacific: Interannual and 
seasonal variations, J. Geophys. Res., 108, 8786, doi 10.1029/2002JD003102. 
Liu, J., J. R. Drummond, D. B. A. Jones, Z. Cao, H. Bremer, J. Kar, J. Zou, F. 
Nichitiu, and J. C. Gille (2006), Large horizontal gradients in atmospheric CO at 
the synoptic scale as seen by spaceborne Measurements of Pollution in the 
Troposphere, J. Geophys. Res., 111, D02306, doi:10.1029/2005JD006076. 
Liu, J., and D. L. Mauzerall (2005), Estimating the average time for intercontinental 
transport of air pollutants, Geophys. Res. Lett., 32, L11814, 
doi:10.1029/2005GL022619. 
Liu, J., D. L. Mauzerall, and L. W. Horowitz (2005), Analysis of seasonal and 
interannual variability in transpacific transport, J. Geophys. Res., 110, D04302, 
doi:10.1029/2004JD005207. 
 
 206 
 
Liu, X., and M. Shao (2004), The analysis of sources of ambient light extinction 
coefficient in summer time of Beijing city, Acta Scientiae Circumstantiae, 24(2), 
185-190 (in Chinese). 
Liu, L., and W. Song (2001), Atmosphere pollution feature analysis and 
countermeasure, J. Hebei Academy of Sciences, 18, 61-64 (in Chinese). 
Logan, J. A. (1999), An analysis of ozonesonde data for the troposphere: 
Recommendations for testing 3-D models and development of a gridded 
climatology for tropospheric ozone, J. Geophys. Res., 104, 16115?16149. 
Luke, W. T. (1997), Evaluation of a commercial pulsed fluorescence detector for the 
measurement of low-level SO2 concentrations during the Gas-Phase Sulfur 
Intercomparison Experiment, J. Geophys. Res., 102, 16255?16265. 
Lun, X., X. Zhang, Y. Mu, A. Nang, and G. Jiang (2003), Size fractionated speciation 
of sulfate and nitrate in airborne particles in Beijing, China, Atmos. Environ., 37, 
2581-2588. 
Ma, J., and J. A. van Aardenne (2004), Impact of different emission inventories on 
simulated tropospheric ozone over China: A regional chemical transport model 
evaluation, Atmos. Chem. Phys., 4, 877?887. 
Ma, Y., H. Yang, Z. Fang, and Y. Zhang (2003), Analysis of status and trends on the 
air pollution in the center multi-cities of Liaoning, Urban Environ. Urban 
Ecology, 16, 89-91 (in Chinese). 
Mari, C., M. J. Evans, P. I. Palmer, D. J. Jacob, and G. W. Sachse (2004), Export of 
Asian pollution during two cold front episodes of the TRACE-P experiment, J. 
Geophys. Res., 109, D15S17, doi:10.1029/2003JD004307. 
 
 207 
 
Martin, B. D., H. E. Fuelberg, N. J. Blake, J. H. Crawford, J. A. Logan, D. R. Blake, 
and G. W. Sachse (2002), Long-range transport of Asian outflow to the 
equatorial Pacific, J. Geophys. Res., 107, 8322, doi:10.1029/2001JD001418. 
Menon, S., J. Hansen, L. Nazarenko, and Y. Luo (2002), Climate effects of black 
carbon aerosols in China and India, Science, 297, 2250-2253. 
Methven, J., et al. (2006), Establishing Lagrangian connections between observations 
within air masses crossing the Atlantic during the International Consortium for 
Atmospheric Research on Transport and Transformation experiment, J. 
Geophys. Res., 111, D23S62, doi:10.1029/2006JD007540. 
Mi, W., Z. Li, X. Xia, B. Holben, R. Levy, F. Zhao, H. Chen, and M. Cribb (2007), 
Evaluation of MODIS aerosol products at two AERONET stations in China, J. 
Geophys. Res., 112, D22S08, doi:10.1029/2007JD008474. 
Mickley, L. J., D. J. Jacob, B. D. Field, and D. Rind (2004), Effects of future climate 
change on regional air pollution episodes in the United States, Geophys. Res. 
Lett., 31, L24103, doi:10.1029/2004GL021216.  
Miyazaki, Y., et al. (2003), Synoptic-scale transport of reactive nitrogen over the 
western Pacific in spring, J. Geophys. Res., 108(D20), 8788, 
doi:10.1029/2002JD003248. 
Nadim, F., P. Zack, G. E. Hoag, S. Liu (2001), United States experience with gasoline 
additives, Energy Policy, 29, 1-5. 
Newell, R. E., and M. J. Evans (2000), Seasonal changes in pollutant transport to the 
North Pacific: The relative importance of Asian and European sources, Geophys. 
Res. Lett., 27(16), 2509-2512, 10.1029/2000GL011501. 
 
 208 
 
Ning, D., L. Zhong, and Y. Chung (1996), Aerosol size distribution and elemental 
composition in urban areas of Northern China, Atmos. Environ., 30, 2355-2362. 
Novelli, P., K. Masarie, and P. Lang (1998), Distributions and recent changes of 
carbon monoxide in the lower troposphere, J. Geophys. Res., 103(D15), 19015-
19033. 
Pacyna, J.M., (1998), Source inventories for atmospheric trace metals. In: Harrison, 
R.M. and van Grieken, R.E., Editors, 1998. Atmospheric Particles, IUPAC 
Series on Analytical and Physical Chemistry of Environmental Systems, Vol. 5, 
Wiley, Chichester, UK, pp. 385?423.  
Palmer, P. I., D. J. Jacob, D. B. A. Jones, C. L. Heald, R. M. Yantosca, J. A. Logan, 
G. W. Sachse, and D. G. Streets (2003), Inverting for emissions of carbon 
monoxide from Asia using aircraft observations over the western Pacific, J. 
Geophys. Res., 108(D21), 8828, doi:10.1029/2003JD003397. 
Park, R. J., D. J. Jacob, B. D. Field, R. M. Yantosca, and M. Chin (2004), Natural and 
transboundary pollution influences on sulfate-nitrate-ammonium aerosols in the 
United States: implications for policy, J. Geophys. Res., 109, D15204, 
10.1029/2003JD004473. 
Parrish, D. D., Y. Kondo, O. R. Cooper, C. A. Brock, D. A. Jaffe, M. Trainer, T. 
Ogawa, G. H?bler, and F. C. Fehsenfeld (2004), Intercontinental Transport and 
Chemical Transformation 2002 (ITCT 2K2) and Pacific exploration of Asian 
Continental Emission (PEACE) experiments: An overview of the 2002 winter 
and spring intensives, J. Geophys. Res., 109, D23S01, 
doi:10.1029/2004JD004980. 
 
 209 
 
Pekney, N. J., C. I. Davidson, K. J. Bein, A. S. Wexler, and M. V. Johnston (2006), 
Identification of sources of atmospheric PM at the Pittsburgh supersite, Part I: 
single particle analysis and filter-based positive matrix factorization, Atmos. 
Environ., 40, S411-S423.  
Pickering, K. E., R. R. Dickerson, G. J. Huffman, J. F. Boatman, and A. Schanot 
(1988), Trace gas transport in the vicinity of frontal convective clouds, , J. 
Geophys. Res., 93(1), 759-773. 
Pickering, K. E., A. M. Thompson, R. R. Dickerson, B. S. Gockel, W. T. Luke, D. P. 
McNamara, J. P. Greenberg, and P. R. Zimmerman (1990), Model calculations 
of tropospheric ozone production potential following observed convective 
events, J. Geophys. Res., 95(9), 14,049-14062. 
Pike, S. M., and S. B. Moran (2001), Trace elements in aerosol and precipitation at 
New Castle, NH, USA, Atmos. Environ., 35, 3361-3366. 
Pope III, C. Arden, David V. Bates, and Mark E. Raizenne (1995), Health effects of 
particulate air pollution: time for reassessment?, Environ. Health Perspect., 103, 
472-480. 
Prados, A. I., R. R. Dickerson, B. G. Doddridge, P. A. Milne, J. L. Moody, and J. T. 
Merrill (1999), Transport of ozone and pollutants from North America to the 
North Atlantic Ocean during the 1996 AEROCE Intensive Experiment, J. 
Geophys. Res., 104(21), 26219-26234. 
Prospero, J. M., D. L. Savoie, and R. Arimoto (2003), Long-term record of nss-sulfate 
and nitrate in aerosols on Midway Island, 1981?2000: Evidence of increased 
 
 210 
 
(now decreasing?) anthropogenic emissions from Asia, J. Geophys. Res., 
108(D1), 4019, doi:10.1029/2001JD001524. 
Qian, Y., D. P. Kaiser, L. R. Leung, and M. Xu (2006), More frequent cloud-free sky 
and less surface solar radiation in China from 1955 to 2000, Geophys. Res. Lett., 
33, L01812, doi:10.1029/2005GL024586. 
Qiu, J., and L. Yang (2000), Variation characteristics of atmospheric aerosol optical 
depths and visibility in North China during 1980?1994, Atmos. Environ., 34, 
603-609. 
Rahn, K. A., (1981), The Mn/V ratio as a tracer of large-scale sources of pollution 
aerosol for the Arctic, Atmos. Environ., 15, 1457-1464. 
Reay, D. S., F. Dentener, P. Smith, J. Grace, and R. A. Feely (2008), Global nitrogen 
deposition and carbon sinks, Nature Geoscience, 1, 430-437. 
Remer, L. A., et al. (2005), The MODIS aerosol algorithm, products and validation, J. 
Atmos. Sci., 62, 947?973. 
Richter, A., J. P. Burrows, H. N? , C. Granier, and U. Niemeier (2005), Increase in 
tropospheric nitrogen dioxide over China observed from space, Nature, 437, 
129-132. 
Rosenfeld, D., U. Lohmann, G. B. Raga, C. D. O?Dowd, M. Kulmala, S. Fuzzi, A. 
Reissell, and M. O. Andreae (2008), Floor or drought: How do aerosols affect 
precipitation, Science, 321, 1309-1313. 
Seinfeld, J. H., and S. N. Pandis (1998), Atmospheric chemistry and physics: From 
air pollution to climate change, John Wiley & Sons, New York, USA, 1326 pp. 
 
 211 
 
Seiler, W., and C. Junge (1970), Carbon monoxide in the atmosphere, J. Geophys. 
Res., 75(12), 2217?2226. 
Shao, M., X. Tang, Y. Zhang, and W. Li (2006), City clusters in China: air and 
surface water pollution, Frontiers in Ecology and the Environment, 7(4), 353-
361. 
Shimizu, A., N. Sugimoto, I. Matsui, K. Arao, I. Uno, T. Murayama, N. Kagawa, K. 
Aoki, A. Uchiyama, and A. Yamazaki (2004), Continuous observations of Asian 
dust and other aerosols by polarization lidars in China and Japan during ACE-
Asia, J. Geophys. Res., 109, D19S17, doi:10.1029/2002JD003253. 
Sinton, J. E., (2001), Accuracy and reliability of China?s energy statistics, China 
Econ. Rev., 12, 373? 383. 
Stehr, J. W., R. R. Dickerson, K. A. Hallock-Waters, B. G. Doddridge, and D. Kirk 
(2000), Observations of NOy, CO, and SO2 and the origin of reactive nitrogen in 
the eastern United States. J. Geophys. Res., 105(D3), 3553-3563. 
Steiner, A. L., S. Tonse, R. C. Cohen, A. H. Goldstein, and R. A. Harley (2006), 
Influence of future climate and emissions on regional air quality in California, J. 
Geophys. Res., 111, D18303, doi:10.1029/2005JD006935. 
Stohl, A., (1998), Computation, accuracy and applications of trajectories?A review 
and bibliography, Atmos. Environ., 32, 947-966. 
Stohl, A., S. Eckhardt, C. Forster, P. James, and N. Spichtinger (2002), On the 
pathways and timescales of intercontinental air pollution transport, J. Geophys. 
Res., 107(D23), 4684, doi:10.1029/2001JD001396. 
 
 212 
 
Stohl, A., C. Forster, S. Eckhardt, N. Spichtinger, H. Huntrieser, J. Heland, H. 
Schlager, S. Wilhelm, F. Arnold, and O. Cooper (2003), A backward modeling 
study of intercontinental pollution transport using aircraft measurements, J. 
Geophys. Res., 108(D12), 4370, doi:10.1029/2002JD002862. 
Streets, D. G.., and S. T. Waldhoff (2000), Present and future emissions of air 
pollutants in China: SO2, NOx, and CO, Atmos. Environ., 34, 363-374. 
Streets, D. G., et al. (2003), An inventory of gaseous and primary aerosol emissions 
in Asia in the year 2000, J. Geophys. Res., 108(D21), 8809, 
doi:10.1029/2002JD003093. 
Streets, D. G., Q. Zhang, L. Wang, K. He, J. Hao, Y. Wu, Y. Tang, and G. R. 
Carmichael (2006), Revisiting China's CO emissions after the Transport and 
Chemical Evolution over the Pacific (TRACE-P) mission: Synthesis of 
inventories, atmospheric modeling, and observations, J. Geophys. Res., 111, 
D14306, doi:10.1029/2006JD007118. 
Sullivan, R. C., S. A. Guazzotti, D. A. Sodeman, and K.A. Prather (2007), Direct 
observations of the atmospheric processing of Asian mineral dust, Atmos. Chem. 
Phys., 7, 1213-1236. 
Sun, Y., G. Zhuang, Y. Wang, L. Han, J. Guo, M. Dan, W. Zhang, Z. Wang, and Z. 
Hao (2004), The air-borne particulate pollution in Beijing ? concentration, 
composition, distribution and sources, Atmos. Environ., 38, 5991-6004. 
Sun, Y., G. Zhuang, Y. Wang, X. Zhao, J. Li, Z. Wang, and Z. An (2005), Chemical 
composition of dust storms in Beijing and implications for the mixing of mineral 
 
 213 
 
aerosol with pollution aerosol on the pathway, J. Geophys. Res., 110, D24209, 
doi:10.1029/2005JD006054. 
Sun, Y., G. Zhuang, W. Zhang, Y. Wang, and Y. Zhuang (2006), Characteristics and 
sources of lead pollution after phasing out leaded gasoline in Beijing, Atmos. 
Environ., 40, 2973-2985. 
Takami, A., T. Miyoshi, A. Shimono, N. Kaneyasu, S. Kato, Y. Kajii, and S. 
Tatakeyama (2007), Transport of anthropogenic aerosols from Asia and 
subsequent chemical transformation, J. Geophys. Res., 112, D22S31, 
doi:10.1029/2006JD008120. 
Takemura, T., T. Nakajima, A. Higurashi, S. Ohta, and N. Sugimoto (2003), Aerosol 
distributions and radiative forcing over the Asian Pacific region simulated by 
Spectral Radiation-Transport Model for Aerosol Species (SPRINTARS), J. 
Geophys. Res., 108, 8659, doi:10.1029/2002JD003210. 
Tan, Q., W. L. Chameides, D. Streets, T. Wang, J. Xu, M. Bergin, and J. Woo (2004), 
An evaluation of TRACE-P emission inventories from China using a regional 
model and chemical measurements, J. Geophys. Res., 109, D22305, 
doi:10.1029/2004JD005071. 
Tang, X., J. Li, X. Su, and D. Chen (1990), Atmospheric Environmental Chemistry, 
Higher Education Press, Beijing, China, 343 pp (in Chinese).    
Taubman, B. F. (2004), Airborne characterization of regional aerosol origins and 
optical properties, Ph.D. Dissertation, Department of Chemistry, University of 
Maryland. 
 
 214 
 
Taubman, B. F., L. Marufu, B. Vant-Hull, C. Piety, B. Doddridge, R. Dickerson, and 
Z. Li (2004), Smoke over haze: Aircraft observations of chemical and optical 
properties and the effects on heating rates and stability, J. Geophys. Res., 109, 
D02206, doi: 10.1029/2003JD003898. 
Taubman, B. F., J. C. Hains, L. T. Marufu, B. G. Doddridge, A. M. Thompson, J. W. 
Stehr, C. A. Piety, and R. R. Dickerson (2006), Aircraft vertical profiles of trace 
gas and aerosol Pollution over the Mid-Atlantic U.S.: Statistics and 
meteorological cluster Analysis, J. Geophys. Res., 111(D10), D10S07, 
10.1029/2005JD006196. 
Taubman, B. F., et al. (2007), Aircraft observations over the Mid-Atlantic U.S. and 
North Atlantic Ocean during the 1996 Atmosphere/Ocean Chemistry 
Experiment: Evidence for linked transport of natural and anthropogenic ozone 
over Bermuda, J. Geophys. Res., in preparation. 
Thomas, V. M. (1995), The elimination of lead in gasoline, Annu. Rev. Energy 
Environ., 20, 301-324. 
Twomey, S. A., (1959), The nuclei of natural cloud formation. Part II: The 
supersaturation in natural clouds and the variation of cloud droplet 
concentrations, Pure Appl. Geophys., 43, 243?249. 
UNEP (2000), The Montreal protocol on substances that deplete the ozone layer, 
Ozone Secretariat, United Nations Environment Programme, Nairobi, Kenya 
(ISBN: 92-807-1888-6). 
 
 215 
 
U.S. DOE (1986), Global distribution of total cloud cover and cloud type amounts 
over land, Tech. Rep. DOE/ER/60085-H1, Natl. Cent. for Atmos. Res., Boulder, 
CO, USA. 
U.S. EPA (2006), Air Quality Criteria for Lead, EPA/600/R-5/144aF. 
Venkataraman, C., and G. U. M. Rao (2001), Emission factors of carbon monoxide 
and size-resolved aerosols from biofuel combustion, Environ. Sci. Technol., 35, 
2100? 2107. 
Virkkula, A., N. C. Ahlquist, D. S. Covert, W. P. Arnott, P. J. Sheridan, P. K. Quinn, 
and D. J. Coffman (2005), Modification, calibration and a field test of an 
instrument for measuring light absorption by particles, Aerosol Sci. Technol., 
39, 68-83. 
Wang, T., T. F. Cheung, Y. S. Li, X. M. Yu, and D. R. Blake (2002), Emission 
characteristics of CO, NOx, SO2 and indications of biomass burning observed at 
a rural site in eastern China, J. Geophys. Res., 107(D12), 
doi:10.1029/2001JD000724. 
Wang, T., et al. (2004), Relationships of trace gases and aerosols and the emission 
characteristics at Lin?an, a rural site in eastern China during spring 2001, J. 
Geophys. Res., 109, doi:10:1029/2003JD004119. 
Wang, T., A. Ding, J. Gao, and W. S. Wu (2006), Strong ozone production in urban 
plumes from Beijing, China, Geophys. Res. Lett., 33, L21806, 
doi:10.1029/2006GL027689. 
Wang, W., and T. Wang (1995), On the origin and the trend of acid precipitation in 
China, Water, Air, & Soil Pollution, 85, 2295-2300. 
 
 216 
 
Wang, W., and T. Wang (1996), On acid rain formation in China, Atmos. Environ., 
30, 4091-4093. 
Wang, X., Z. Han, and X. Lei (2003), Study on ozone concentration change of 
Guangzhou district, Acta Scientiarum Naturalium Universitatis Sunyatseni, 42, 
106-109 (in Chinese). 
Wang, X., D. L. Mauzerall, Y. Hu, A. G. Russell, E. D. Larson, J.-H. Woo, D. G. 
Streets, and A. Guenther (2005), A high-resolution emission inventory for 
eastern China in 2000 and three scenarios for 2020, Atmos. Environ., 39, 5917-
5933. 
Wang, X., and D.L. Mauzerall (2006), Evaluating impacts of air pollution in China on 
public health: Implications for future air pollution and energy policies, Atmos. 
Environ., 40, 1706-1721. 
Wang, Y., M. B. McElroy, T. Wang, and P. I. Palmer (2004), Asian emissions of CO 
and NOx: Constraints from aircraft and Chinese station data, J. Geophys. Res., 
109, D24304, doi:10.1029/2004JD005250. 
Wang Y., M. B. McElroy, K. F. Boersma, H. J. Eskes, and J. P. Veefkind (2007), 
Traffic restrictions associated with the Sino-African summit: Reductions of NOx 
detected from space, Geophys. Res. Lett., 34, L08814, 
doi:10.1029/2007GL029326. 
Wang, Y., G. Zhuang, A. Tang, H. Yuan, Y. Sun, S. Chen, and A. Zheng (2005), The 
ion chemistry and source of PM2.5 aerosol in Beijing, Atmos. Environ., 39, 3771-
3784. 
 
 217 
 
Wang, Y., G. Zhuang, Y. Sun, and Z. An (2006), The variation of characteristics and 
formation mechanisms of aerosols in dust, haze, and clear days in Beijing, 
Atmos. Environ., 40, 6579-6591. 
Wang, Y., G. Zhuang, A. Tang, W. Zhang. Y. Sun, Z. Wang, and Z. An (2007), The 
evolution of chemical components of aerosols at five monitoring sites of China 
during dust storms, Atmos. Environ., 41, 1091-1106. 
Wang, Z., H. Akimoto, and I. Uno (2002), Neutralization of soil aerosol and its 
impact on the distribution of acid rain over East Asia: Observations and model 
results, J. Geophys. Res., 107(D19), 4389, doi:10.1029/2001JD001040. 
Wen, T., R. R. Dickerson, Y. Wang, Y. Sun, M. Xue, D. Streets, and Z. Li (2006), 
Trace gases and aerosol observations in Beijing: 1, evaluations of emission 
inventories, in preparation. 
Westerling, A.L., H.G. Hidalgo, D.R. Cayan, and T.W. Swetnam (2006), Warming 
and earlier spring increases western U.S. forest wildfire activity, Science, 313: 
940-943 
WHO and UNDP (2001), Environment and people?s health in China, World Health 
Organization, Geneva, Switzerland, 76 pp. 
Williamson, S. J. (1973), Fundamentals of air pollution, Addison-Wesley Publication 
Company, Inc. Philippines, 472 pp. 
World Bank (1997), Clear water, blue skies: China's environment in the new century, 
World Bank, Washington DC, USA, 122 pp. 
World Bank (2006), 2006 World Development Indicators database, World Bank, 
Washington DC, USA, 242 pp. 
 
 218 
 
Wu, X., J. Sun, B. Shi, and W. Liu (2002), The source of nitrogen oxide and its 
characters in the air of Urumqi, Arid Environ. Monitoring, 16, 137-139 (in 
Chinese). 
Xia, X., H. Chen, Z. Li, P. Wang, and J. Wang (2007), Significant reduction of 
surface solar irradiance induced by aerosols in a suburban region in northeastern 
China, J. Geophys. Res., 112, D22S02, doi:10.1029/2006JD007562. 
Xin, J., et al. (2007), Aerosol optical depth (AOD) and ?ngstr?m exponent of 
aerosols observed by the Chinese Sun Hazemeter Network from August 2004 to 
September 2005, J. Geophys. Res., 112, D05203, doi:10.1029/2006JD007075. 
Xu, H., Y. Wang, T. Wen, and X. He (2007a), Size distributions and vertical 
distributions of water soluble ions of atmospheric aerosol in Beijing, Environ. 
Sci., 28, 14-19 (in Chinese). 
Xu, H., Y. Wang, T. Wen, and X. He (2007b), Size distributions and vertical 
distributions of metal elements of atmospheric aerosol in Beijing, Environ. 
Chem., 26, 675-679 (in Chinese). 
Xu, J., M. H. Bergin, X. Yu, G. Liu, J. Zhao, C. M. Carrico, and K. Baumann (2002), 
Measurement of aerosol chemical, physical, and radiative properties in the 
Yangtze delta region of China, Atmos. Environ., 36, 161-173. 
Xu, J., M. H. Bergin, R. Greenwald, J. J. Schauer, M. M. Shafer, J.-L. Jaffrezo, and 
G. Aymoz (2004), Aerosol chemical, physical and radiative characteristics near 
a desert source region of Northwest China during ACE-Asia, J. Geophys. Res., 
109, D19S03, doi:10.1029/2003JD004239 
 
 219 
 
Xu, L., K. Okada, Y. Iwasaka, K. Hara, Y. Okuhara, Y. Tsutsumi, and G. Shi (2001), 
The composition of individual aerosol particle in the troposphere and 
stratosphere over Xianghe (39.45?N, 117.0?E), China, Atmos. Environ., 35, 
3145-3153. 
Xu, M., C.-P. Chang, C. Fu, Y. Qi, A. Robock, D. Robinson, and H. Zhang (2006), 
Steady decline of East Asian monsoon winds, 1969?2000: Evidence from direct 
ground measurements of wind speed, J. Geophys. Res., 111, D24111, 
doi:10.1029/2006JD007337. 
Xu, Q. (2001), Abrupt change of the mid-summer climate in central east China by the 
influence of atmospheric pollution, Atmos. Environ., 35, 5029-5040. 
Yan, P., J. Tang, J. Huang, J. T. Mao, X. J. Zhou, Q. Liu, Z. F. Wang, and H. G. Zhou 
(2008), The measurements of aerosol optical properties at a rural site in 
Northern China, Atmos. Chem. Phys., 8, 2229-2242. 
Yang, F., K. He, B. Ye, X. Chen, L. Cha, S. H. Cadle, T. Chan, and P. A. Mulawa 
(2005), One-year record of organic and elemental carbon in fine particles in 
downtown Beijing and Shanghai, Atmos. Chem. Phys., 5, 1449-1457. 
Yang, L., B. Qin, and R. Wu (2001), Preliminary study for potential impacts on the 
aquatic environment of Lake Taihu by acid rain, J. Lake Sci., 13(2), 135-142. 
Yao, X., A. P. S. Lau, M. Fang, C. K. Chan, and M. Hu (2003), Size distributions and 
formation of ionic species in atmospheric particulate pollutants in Beijing, 
China: 1?inorganic ions, Atmos. Environ., 37, 2991-3000. 
Yienger, J. J., M. Galanter, T. A. Holloway, M. J. Phadnis, S. K. Guttikunda, G. R. 
Carmichael, W. J. Moxim, and H. Levy II (2000), The episodic nature of air 
 
 220 
 
pollution transport from Asia to North America, J. Geophys. Res., 105(D22), 
26931-26946, 10.1029/2000JD900309. 
Yu, H., S. C. Liu, and R. E. Dickinson (2002), Radiative effects of aerosols on the 
evolution of the atmospheric boundary layer, J. Geophys. Res., 107(D12), 
doi:10.1029/2001JD000754. 
Yuan, H., G. Zhuang, J. Li, Z. Wang, and J. Li (2008), Mixing of mineral with 
pollution aerosols in dust season in Beijing: Revealed by source apportionment 
study, Atmos. Environ., 42, 2141-2157. 
Yumimoto, K., and I. Uno (2006), Adjoint inverse modeling of CO emission over the 
East Asian region using four dimensional variational data assimilation, Atmos. 
Environ., 40, 6836-6845. 
Zhang, D., and Y. Iwasaka (1999), Nitrate and sulfate in individual Asian dust-storm 
particles in Beijing, China in spring of 1995 and 1996, Atmos. Environ., 33, 
3213-3223. 
Zhang, J., K. R. Smith, R. Uma, Y. Ma, V. V. N. Kishore, K. Lata, M. A. K. Khalil, 
R. A. Rasmussen, and S. T. Thorneloe (1999), Carbon monoxide from 
cookstoves in developing countries: 1. Emission factors, Chemosphere, 1, 353? 
366. 
Zhang, Q., et al. (2007), NOx emission trends for China, 1995?2004: The view from 
the ground and the view from space, J. Geophys. Res., 112, D22306, 
doi:10.1029/2007JD008684. 
 
 221 
 
Zhang, R., R. Arimoto, J. An, S. Yabuki, and J. Sun (2002), Ground observations of a 
strong dust storm in Beijing in March 2002, J. Geophys. Res., 110, D18S06, 
doi:10.1029/2004JD004589. 
Zhang, R., G. Li, J. Fan, D. L. Wu, and M. J. Molina (2007), Intensification of Pacific 
storm track linked to Asian pollution, Proceedings of the National Academy of 
Sciences, 104, 5295-5299. 
Zhang, X. Y., S. L. Gong, Z. X. Shen, F. M. Mei, X. X. Xi, L. C. Liu, Z. J. Zhou, D. 
Wang, Y. Q. Wang, and Y. Cheng (2003), Characterization of soil dust aerosol 
in China and its transport and distribution during 2001 ACE-Asia: 1. Network 
observations, J. Geophys. Res., 108(D9), 4261, doi:10.1029/2002JD002632. 
Zhang, Y., X. Zhu, S. Slanina, M. Shao, L. Zeng, M. Hu, M. Bergin, and L. Salmon 
(2004), Aerosol pollution in some Chinese cities, Pure Appl. Chem., 76. 1227?
1239. 
Zhao, C., X. Tie, and Y. Lin (2006), A possible positive feedback of reduction of 
precipitation and increase in aerosols over eastern central China, Geophys. Res. 
Lett., 33, L11814, doi:10.1029/2006GL025959. 
Zheng, M., L. G. Salmon, J. J. Schauer, L. Zeng, C. S. Kiang, Y. Zhang, and G. R. 
Cass (2005), Seasonal trends in PM2.5 source contributions in Beijing, China, 
Atmos. Environ., 39, 3967-3976. 
Zou, X., P. Zhai, and Q. Zhang (2005), Variations in droughts over China: 1951?2003, 
Geophys. Res. Lett., 32, L04707, doi:10.1029/2004GL021853.