2.5. pm2.5 composition, sources, and air …2.5. pm2.5 composition, sources, and air concentrations...

30 June 2004 | State of SOS-3: 1995 - 2003 | 79

2.5. PM2.5 Composition, Sources, and Air Concentrations

Composition of PM

Industrial emissions Major sources

Composition of ultrafines Motor vehicles and elemental

carbon Secondary vs primary aerosol

Local vs distant sources Exceedances in SE

Diurnal and seasonal variability

Radiative and optical properties

Internal mixing of constituents

Urban plumes under stagnation conditions

Urban plumes under advective conditions

Ozone transport from urban to rural areas

2.5. PM2.5 COMPOSITION, SOURCES, AND AIR CONCENTRATIONS Paul Solomon and David Allen

SOS was selected by EPA’s Office of Air Quality

Planning and Standards (EPA/OAQPS) to establish the

first of two initial Supersites. The other initial Supersite

was in Fresno-Bakerfield, CA. Both sites were part of

EPA's effort under the leadership of Paul Solomon and

Richard Scheffe to increase the nation's capacity to

monitor both coarse (PM10) and fine (PM2.5) fractions in

a reliable way in various parts of the United States.

The Atlanta Supersite Experiment was established

under the leadership of SOS' Chief Scientist, Bill

Chameides. SOS brought together in Atlanta during

August 1999, the most comprehensive array of

particulate-matter measurement instruments ever

assembled in the US and about 150 of the nation’s most

competent aerosol research scientists and a few from

abroad.

David Allen of the University of Texas in Austin, TX and Matt Fraser of Rice University in

Houston, TX also were selected by EPA/OAQPS to lead the Houston Supersite, one of eight

EPA Supersites around the United States that are part of EPA’s PM Supersite Program – a

cooperative endeavor with other carefully selected public-health and regional-haze investigators.

Here very detailed, chemical and physical characterization measurements of ambient coarse and

fine aerosols were made in all months of the year (2001-2002) at multiple, carefully selected

urban and rural locations in southeast Texas.

The objectives of both the Atlanta and Houston Supersite programs were to “advance

scientific understanding of atmospheric processes regarding formation and accumulation of PM.”

The Atlanta Supersite Experiment was located on Jefferson Street in Atlanta, GA at a mixed

commercial and industrial area about 8 km west of the city center. Here detailed PM

characterization measurements had been made routinely as a part of EPRI’s and the Southern

Company’s SEARCH and ARIES programs for more than a year. An extraordinarily diverse

30 June 2004 | State of SOS-3: 1995 - 2003 | 80

variety of instruments were assembled at this site for comparison purposes and to determine the

mass, particle size distribution, chemical composition of individual particles, and chemical

composition of hourly-collected filter samples collected on all days of the week, and

simultaneous gas and particle measurements.

Scientific findings are summarized below – mainly from the Atlanta Supersite Experiment in

August 1999, year-round, rural and urban measurements at the Houston Supersite in southeast

Texas, and both January and July measurements at a rural site near Anderson, SC.

30 June 2004 | State of SOS-3: 1995 - 2003 | 81

2.5.1. Composition of PM in the Southeastern US The major components of total PM mass on average for the urban Atlanta Supersite study

[and both rural and urban locations within the Houston Supersite program] were: organic

material (~35%) [25-30%], sulfate (~34%) [30-40%]; ammonium (~12%) [7-10%], elemental

carbon (~3%) [2-5%], nitrate (~2%) [1-4%], and crustal material (~3%) [<1-3%]. Minor

differences between these two sets of numbers are mostly due to differences in the mix of

sources, chemical processing, and averaging period (summer – August vs annual average) and

possibly due in part to differences in sampling and chemical analysis methods.

• Hourly PM2.5 data in Atlanta indicated two types of events – morning peaks dominated by carbonaceous material and afternoon events dominated by sulfate (Figure 2.5.1.). Carbon and sulfate accounted for ~75% of aerosol mass during these peak events. Nitrate concentrations were generally low. However, the hourly data clearly indicated the temporal nature of nitrate, with nitrate concentrations peaking in the early morning before sunrise, when temperature was at its minimum.

Figure 2.5.1. Chemical composition of the fine aerosol compared with the total PM2.5 mass for the second half of the Atlanta Supersite Study. Plot (a) shows results from a 24-hr average and (b) 1-hr average. Events dominated by SO4

2- are indicated by 2s-4s and events dominated by organic matter by 3c-4c (after Weber et al., 2003).

• Significant semi-volatile organic material was present in PM2.5 particles collected during the Atlanta Supersite Experiment.

At a rural sampling site near Anderson, SC, organic compounds made up the largest single

component of aerosol mass – 38.5% in July vs 64.1% in January. This substantial difference

may be attributed to word burning for heating purposes in winter months. Sulfate was the

second largest contributor to PM2.5 mass – with 35% in July and 22% during January.

Data on the composition of PM2.5 in southeast Texas indicate that:

30 June 2004 | State of SOS-3: 1995 - 2003 | 82

• Sulfate, ammonium ion (which neutralizes the sulfate ion), organic carbon, and elemental carbon are the major constituents of PM2.5; the annual average concentrations of these major components were spatially homogeneous across southeast Texas.

• Nevertheless, localized events with high mass fractions of sulfate or carbon occurred frequently at many monitors in this region.

• Concentrations of sulfate were slightly higher in the spring and late fall than in the summer; carbon concentrations were highest in the late fall.

• High organic-carbon to elemental-carbon ratios suggest that much of the carbonaceous material in PM2.5 in southeast Texas is not emitted directly, but is formed in the air through reactions involving both gaseous biogenic and anthropogenic VOC emissions.

KEY CITATIONS: Husain, H. and C. Christoforou. 2003. Concentration and Chemical Composition of PM2.5 Particles at a Rural Site

in South Carolina. SOS Final Report. Clemson University. 35 pp. Modey, W.K., E.J. Eatough, Y. Pang, and N.L. Eatough. 2004. Performance and evaluation of the PC-BOSS for

fine PM2.5 sampling during the summer EPA Supersite Program in Atlanta. J. Air Waste Manage. Assoc. (in press).

Russell, M.M. and D.T. Allen. 2004. Seasonal and spatial trends in primary and secondary organic carbon concentrations in southeast Texas. Atmos Environ. 38:3225-3239.

Russell, M.M., D.T. Allen, D.R. Collins, and M.P. Fraser. 2004. Daily, seasonal and spatial trends in PM2.5 mass and composition in southeast Texas. Aerosol. Sci. Technol. 38(S1):14-26, doi:10.1080/02786820390229138.

Solomon, P.A., W. Chameides, R. Weber, A. Middlebrook, C.S. Kiang, A.G. Russell, A. Butler, B. Turpin, D. Mikel, R. Scheffe, E. Cowling, E. Edgerton, J. St. John, J. Jansen, P. McMurry, S. Hering, and T. Bahadori. 2003b. Overview of the 1999 Atlanta Supersite project. J. Geophys. Res. 108(D7), 8413, doi:10.1029/2001JD001458.

Weber, R., D. Orsini, A. Sullivan, M. Bergin, C.S. Kiang, M. Chang, Y.N. Lee, P. Dasgupta, J. Slanina, B. Turpin, E. Edgerton, S. Hering, G. Allen, P. Solomon, and W. Chameides. 2003. Transient PM2.5 aerosol events in metro Atlanta: Implications for air quality and health. J. Air Waste Manage. Assoc. 53:84-91.

30 June 2004 | State of SOS-3: 1995 - 2003 | 83

2.5.2. Sources of Fine Particles in Houston, TX Fresh hydrophobic ultrafine particles are emitted by industrial sources in the Ship Channel

area of Houston; they grow in size and become more hydrophilic as they grow.

Particle growth within the VOC-rich ship channel plume exceeded that expected solely from

SO2 oxidation. But particle growth within the plume of the Parish power plant was generally

consistent with condensation of the oxidation products of SO2 when the plume did not pass over

substantial sources of VOCs.

There are five major sources of primary PM2.5 emissions in southeast Texas. They include:

1) mobile sources, 2) cooking of foods, 3) point sources, 4) geological sources, and 5) wild fires

and open burning.

• Primary mobile-source emissions are significant; evidence suggests that these emissions account for about 25-35% of PM2.5 mass in SE Texas. In fact, diesel engines in heavy duty trucks, trains, and farm or construction equipment, and gasoline engines in cars, trucks, boats, and hand tools, as well as jet-fueled aircraft, account for most primary emissions of PM2.5 in southeast Texas.

• Primary emissions from cooking of foods are significant in all urban areas; evidence suggests that these emissions account for about 10-15% of PM2.5 mass in urban areas.

• Point sources of primary PM10 particles are significant, but point-sources of primary PM2.5 particles have not yet been quantified. Thus, additional research is needed to determine the importance, size distributions, and chemical compositions of these PM2.5 primary emissions.

• Geological sources (wind-blown dust) are a relatively minor contributor to the total mass of PM2.5.

• Fires are a sporadic, but significant source of primary PM2.5 emissions in Texas. On an annual average basis, they contribute about 1-2% of the total mass of fine particles in the Houston-Galveston area; but these emissions tend to be concentrated on specific days with fire events.

KEY CITATIONS: Allen, D.T. 2002. Particulate Matter Concentrations, Compositions, and Sources in Southeast Texas: State of the

Science and Critical Research Needs. Report to the Texas Environmental Research Consortium. 93 pp. http://www.harc.edu/harc/Projects/AirQuality/Projects/Status/Reports.aspx

Brock, C.A., M. Trainer, T.B. Ryerson, J.A. Neuman, D.D. Parrish, J.S. Holloway, D.K. Nicks, Jr., G.J. Frost, G. Hübler, F.C. Fehsenfeld, J.C. Wilson, J.M. Reeves, B.G. Lafleur, H. Hilbert, E.L. Atlas, S.G. Donnelly, S.M. Schauffler, V.R. Stroud, and C. Wiedinmyer. 2003. Particle growth in urban and industrial plumes in Texas. J. Geophys. Res. 108(D3), 4111, doi:10.1029/2002JD002746.

NOAA Aeronomy Laboratory. 2003. Texas 2000 Air Quality Study - Phase II Analysis of NOAA Data. Final Report to Texas Commission on Environmental Quality Houston/Galveston Air Quality Science Evaluation. 158 pp.

30 June 2004 | State of SOS-3: 1995 - 2003 | 84

ftp://ftp.tnrcc.state.tx.us/pub/OEPAA/TAD/Modeling/HGAQSE/Contract_Reports/Data_Analysis/TexAQS2000_NOAA_Data_Analysis.pdf 2.5.3. Emission Sources of PM2.5 and PM10 in Atlanta, GA A source identification technique called positive matrix factorization was used with daily

integrated particulate matter data on mass concentration and composition collected in Atlanta

between August 1998 and August 2000. For PM2.5 particles, eight major types of emission

sources were identified: 1) SO42--rich secondary aerosol sources (56%), 2) motor vehicle sources

(22%), 3) wood smoke sources (11%), 4) NO3--rich secondary aerosol sources (7%), 5) mixed

cement kiln and organic carbon sources (2%), 6) airborne soil sources (1%), 7) metal recycling

facilities (0.5%), and 8) a miscellaneous source that includes bus stations and metal processing

facilities (0.3%). Invariably, NH4+ [presumably mainly from agricultural sources] was

associated with both the SO42--rich and NO3

--rich secondary aerosols.

For PM10 particles, five major types of sources were identified: 1) airborne soil sources

(60%), 2) NO3--rich secondary aerosol sources (16%), 3) SO4

2--rich secondary aerosol sources

(12%), 4) cement kiln facilities (11%), and 5) metal recycling facilities (1%).

Summary finding from this work can be stated:

• Sulfate-rich secondary aerosol was the primary contributor to Atlanta PM2.5 mass; airborne soil was the largest primary source of PM10 particle mass in Atlanta.

KEY CITATION: Kim, E., P.K. Hopke, and E.S. Edgerton. 2003. Source identification of Atlanta aerosol by positive matrix

factorization. J. Air Waste Manage. Assoc. 53:731-739.

2.5.4. Composition of Ultrafine Particles in Atlanta, GA • The composition of the ultrafine (less than 100 nm) particles was dominated by carbon

compounds. The major composition classes (expressed as percentage of particle mass) were: organic carbon (~74%), potassium (~8%), iron (~3%), calcium (~2%), nitrate (~2%), elemental carbon (~1.5%), and sodium (~1%).

• The total mass of ultrafine particles (<10 nm) was higher during traffic rush hours, while larger diameter particles (10-100 nm and 100-2000 nm) had higher concentrations at night than during the day, while also reaching their highest concentrations during traffic rush hours.

KEY CITATIONS: Rhoads, K.P., D.J. Phares, A.S. Wexler, and M.V. Johnston. 2003. Size-resolved ultrafine particle composition

analysis, 1. Atlanta. J. Geophys. Res. 108(D7), 8418, doi:10.1029/2001JD001211. Solomon, P.A., W. Chameides, R. Weber, A. Middlebrook, C.S. Kiang, A.G. Russell, A. Butler, B. Turpin, D.

Mikel, R. Scheffe, E. Cowling, E. Edgerton, J. St. John, J. Jansen, P. McMurry, S. Hering, and T.

30 June 2004 | State of SOS-3: 1995 - 2003 | 85

Bahadori. 2003b. Overview of the 1999 Atlanta Supersite project. J. Geophys. Res. 108(D7), 8413, doi:10.1029/2001JD001458.

30 June 2004 | State of SOS-3: 1995 - 2003 | 86

2.5.5. Elemental Carbon in PM2.5 Time-resolved ambient particulate organic and elemental carbon were measured during the

Atlanta Supersite Study using five different instruments, in order to investigate temporal trends

of carbon-containing aerosols and to determine the contributions of primary and secondary

organic carbon to particulate organic carbon. Major findings of this work are summarized

below.

• In spite of the large uncertainties inherent in measuring carbon-containing particulate matter, which is very complex in composition, and in utilizing different operational techniques for measurement, there was generally good agreement between measurement systems.

• Organic matter and elemental carbon comprised ~40% and ~8%, respectively, of PM2.5 mass on average during the August 1999 Atlanta Supersite Study.

• Motor vehicles were indicated as a primary emission source of elemental carbon in Atlanta, using carbon monoxide as a tracer. Elemental carbon concentrations tended to peak at 0600-0900 EST, which is probably indicative of motor vehicle emissions, and had a much smaller peak in the evening.

KEY CITATIONS: Lim, H.-J. and B.J. Turpin. 2002. Origins of primary and secondary organic aerosol in Atlanta: Results of time-resolved

measurements during the Atlanta Supersite Experiment. Environ. Sci. Technol. 36:4489-4496. Lim, H.-J., B.J. Turpin, E. Edgerton, S. Hering, G. Allen, H. Maring, and P. Solomon. 2003. Semicontinuous aerosol

carbon measurements: Comparison of Atlanta Supersite measurements. J. Geophys. Res. 108(D7), 8419, doi:1029/2001JD001214.

30 June 2004 | State of SOS-3: 1995 - 2003 | 87

2.5.6. Secondary Formation of Organic Aerosols Contributions of primary emissions and secondary aerosol formation to measured organic

carbon in Atlanta in August 1999 were estimated to be roughly equal. Secondary organic carbon

estimates are estimated to have a + 10% uncertainty.

Poor correlation between elemental carbon and organic carbon in Atlanta in summer is

indicative of secondary organic aerosol formation. The ratio of organic carbon to elemental

carbon closely tracked daytime variations in ozone concentrations. Thus, the ratio increased

during the afternoon in correspondence to photochemical activity.

Ratios of organic carbon to elemental carbon suggest that much of the carbonaceous material

in southeast Texas is secondary organic aerosol, that is, it is formed in the atmosphere as the

result of the reactions of gas phase VOC emissions. Point sources and area/non-road emissions

are the dominant contributors to formation of anthropogenic secondary organic aerosol in the

Houston urban area, but on a regional basis, biogenic emissions of secondary organic aerosol

precursors overwhelm anthropogenic sources.

Radiocarbon dating (14C/13C ratios) indicates that at some suburban and rural locations in

southeast Texas, formation of secondary organic aerosol is dominated by reactions involving

biogenic emissions. These locations are primarily north and southwest of Houston’s urban core.

• Secondary formation of organic aerosols tended to be large compared to primary emissions. This was true at both the Atlanta and Houston Supersites

KEY CITATIONS: Dechapanya, W., M.M. Russell, and D.T. Allen. 2002. Estimates of anthropogenic secondary organic aerosol

formation in Houston, Texas. Aerosol Sci. Technol. 38(1):156-166, doi:10.1080/02786820390229462. Lemire, K.R., D.T. Allen, G.A. Klouda, and C.W. Lewis. 2002. Fine particulate matter source attribution for

Southeast Texas using 14C/13C ratios. J. Geophys. Res. 107(D22), 4613, doi:10.1029/2002JD002339. Lim, H.-J. and B.J. Turpin. 2002. Origins of primary and secondary organic aerosol in Atlanta: Results of time-

resolved measurements during the Atlanta Supersite Experiment. Environ. Sci. Technol. 36:4489-4496. Lim, H.-J., B.J. Turpin, E. Edgerton, S. Hering, G. Allen, H. Maring, and P. Solomon. 2003. Semicontinuous

aerosol carbon measurements: Comparison of Atlanta Supersite measurements. J. Geophys. Res. 108(D7), 8419, doi:1029/2001JD001214.

30 June 2004 | State of SOS-3: 1995 - 2003 | 88

2.5.7. Local and Regional Sources of PM2.5 in Southeast Texas

Sulfate makes up about 30-40% and organic carbon and elemental carbon make up about 25-

30% of the total, annual average, PM2.5 mass in southeast Texas. Point sources of SO2 emissions

are the dominant source of locally generated sulfate in PM2.5. There is also evidence for

significant local sources of the carbonaceous material found in PM2.5. The sulfate in PM2.5 often

is not completely neutralized; thus ammonia emissions influence the total mass of PM2.5. The

dominant source of ammonia in Texas is cattle and other livestock; in most of the state’s urban

areas, on-road sources and domestic activities (use of cleaning products, human perspiration and

respiration, human wastes and pet wastes) dominate ammonia emissions.

But it appears that not all of the sulfate or carbonaceous material observed in PM2.5 in

southeast Texas is emitted or formed locally. Examination of the spatial distributions of PM2.5

and composition and analysis of air parcel back trajectories indicate that:

• When high concentrations of fine particulate matter mass, sulfate and organic carbon are observed throughout southeast Texas, back-trajectory analyses of these air parcels often indicate high concentrations of background sulfate and organic carbon in PM2.5 that extend far upwind. This suggests that much that much sulfate and carbonaceous aerosol is transported into southeast Texas from the eastern half of North America.

• But, high concentrations of fine particulate matter mass and organic carbon are observed at isolated monitors in southeast Texas, suggesting local source contributions are important on some days.

KEY CITATION: Russell, M.M., D.T. Allen, D.R. Collins, and M.P. Fraser. 2004. Daily, seasonal and spatial trends in PM2.5 mass

and composition in southeast Texas. Aerosol. Sci. Technol. 38(S1):14-26, doi:10.1080/02786820390229138.

30 June 2004 | State of SOS-3: 1995 - 2003 | 89

2.5.8. Concentrations of PM2.5 Mass in the Southeastern US Data collected and analyzed in connection with the Atlanta Supersite Experiment lead to the

following conclusions:

• The average air concentration of PM2.5 mass during the August 1999 Atlanta Supersite Experiment was 31.3 µg m-3, with a peak value of 47.2 µg m-3. Thus, the 24-hour PM2.5 standard was not exceeded. Interestingly, the 1-hr ozone standard was exceeded on several days, for multiple hours, during the study. Sulfate and ammonium ion concentrations were well correlated with PM2.5 mass; but organic carbon and elemental carbon concentrations were not very well correlated.

• Monitoring data from January 2001 to February 2002 at the Atlanta Speciation Trends Network site showed an average total PM2.5 concentration in the winter of 12 µg m-3, and 20 µg m-3 in the summer, with the highest 24-hour average totaling µg m-3.

• Annual PM2.5 mass concentrations measured from March 1999 to February 2000 in the ASACA study in Atlanta exceeded the annual NAAQS of 15 µg m-3 at all four monitoring sites, with annual averages ranging from 19.3 to 21.2 µg m-3. One site violated the daily NAAQS of 65 µg m-3.

Analyses of PM measurements at a rural site during summer and winter indicate:

• At a rural site near Anderson, SC, the average PM2.5 mass during July 2001 was 20.9 µg m-3, with a high of 41.2 µg m-3 on July 18 and a low of 4.4 µg m-3 on July 25. The overall average in January 2002 was 9.4 µg m-3, with a high of 18.2 µg m-3 on January 18 and a low of 3.7 µg m-3 on January 25. Across all sampling events, the average annual mass concentration was 15.1 µg m-3, just above the new NAAQS annual standard of 15 µg m-3.

Data on air concentrations of PM2.5 mass in southeast Texas indicate that:

• Annual average mass concentrations of PM2.5, over wide regions of eastern and southeastern Texas, range from approximately 10 µg m-3 to 15 µg m-3, which is close to the annual average NAAQS of 15 µg m-3.

• When averaged over long time periods, PM2.5 mass concentrations are spatially homogeneous throughout southeast Texas.

• Local events with moderately high air concentrations of PM2.5 mass occur at many monitors in this region but only rarely exceed the annual average NAAQS.

KEY CITATIONS: Butler, A.J., M.S. Andrew, and A.G. Russell. 2003. Daily sampling of PM2.5 in Atlanta: Results of the first year

of the Assessment of Spatial Aerosol Composition in Atlanta study. J. Geophys. Res. 108(D7), 8415, doi:10.1029/2002JD002234.

Husain, H. and C. Christoforou. 2003. Concentration and Chemical Composition of PM2.5 Particles at a Rural Site in South Carolina. SOS Final Report. Clemson University. 35 pp.



30 June 2004 | State of SOS-3: 1995 - 2003 | 90

2.5.9. Variability in Concentrations of PM2.5 in the Southeastern US Measurements of the air concentrations of PM2.5 mass in Atlanta and surrounding rural areas

indicate that:

• Winter and summer data from a rural site near Anderson, SC showed higher mass concentrations in summer than in winter. Sulfate ion and ammonium ion concentrations increased in summer, but nitrate ion concentrations decreased in summer at this site. Comparison of these SC air concentration data with those for similar rural sites in GA and NC showed that the NC sites generally had higher and the GA sites generally had lower air concentrations of PM2.5 mass during late 2001 and early 2002.

• ASACA data in Atlanta showed that most PM2.5 constituents peaked during summer months; but nitrate, metals, and elemental carbon usually showed some enhancement during winter due mainly to lower inversion heights. Diurnally, there were discernible early morning and late night peaks that corresponded to rush-hour traffic patterns and inversion heights, respectively.

• Comparison of data from Atlanta, GA and Fresno, CA showed that seasonal differences in meteorology and amounts of emissions have significant influences on seasonal variability in the composition of PM2.5 at both locations.

Data on the air concentrations of PM2.5 mass in southeast Texas indicate that:

• Throughout the region, concentrations are slightly higher in the spring and late fall than in summer.

• A consistent and strong morning peak in PM2.5 mass concentrations is observed throughout the region and a weaker and slightly less consistent peak in mass concentration is observed in the late afternoon to early evening.

KEY CITATIONS: Butler, A.J., M.S. Andrew, and A.G. Russell. 2003. Daily sampling of PM2.5 in Atlanta: Results of the first year

of the Assessment of Spatial Aerosol Composition in Atlanta study. J. Geophys. Res. 108(D7), 8415, doi:10.1029/2002JD002234.

Chu, S.-H., J.W. Paisie, and B.W.-L. Jang. 2003. PM data analysis – A comparison of two urban areas: Fresno and Atlanta. J. Geophys. Res. (in review)

Husain, H. and C. Christoforou. 2003. Concentration and Chemical Composition of PM2.5 Particles at a Rural Site in South Carolina. SOS Final Report. Clemson University. 35 pp.


30 June 2004 | State of SOS-3: 1995 - 2003 | 91

2.5.10. Radiative Forcing by Aerosols in Atlanta During the Atlanta Supersite Experiment, studies were made of both the optical and radiative

forcing properties of the aerosols observed in Atlanta. These studies showed that:

• Light scattering was dependent on a wide range of chemical components of the aerosols. • Light absorption is most strongly linked to the elemental carbon component.

• The average direct aerosol radiative forcing properties of the Atlanta was about minus 11 + 6 watts m-2; this value is about 10 times larger than global mean estimates for aerosols.

KEY CITATION: Carrico, C.M., M.H. Bergin, J. Xu., K. Baumann, and H. Maring. 2003. Urban aerosol radiative properties:

Measurements during the 1999 Atlanta Supersite experiment. J. Geophys. Res. 108(D7), 8422, doi:10.1029/2001JD001222.

2.5.11. Composition of Aerosols in Atlanta Chemical composition of single particles was measured using the Particle Analysis by Laser

Mass Spectrometry (PALMS) instrument during the Atlanta Supersite Study. Particle

composition was generally internally mixed. The predominant particle consisted of organic

species and sulfate, and often contained other components such as nitrate, ammonium, halogens,

metals, soot/hydrocarbon, and aluminosilicates. More than 20% of the negative ion spectra of

single particles contained nitrate ion peaks, which showed a clear maximum during the morning

at high relative humidity and a smaller maximum in the afternoon. About 45% of the negative

spectra contained ions indicative of oxidized organics, with similar morning and afternoon

maxima. At high relative humidity, the nitrate peaks were often mixed internally with sulfate.

The single particle data also indicated the presence of organic sulfur containing compounds,

which might account for 10 — 15% of sulfur observed in Atlanta PM2.5.

• The composition of particles measured during the Atlanta Supersite Study was generally internally mixed, with components of organic matter, sulfate, nitrate, ammonium and other constituents.

KEY CITATIONS: Lee, S.-H., D.M. Murphy, D.S. Thomson, and A.M. Middlebrook. 2002. Chemical components of single

particles measured with particle analysis by laser mass spectrometry (PALMS) during the Atlanta Supersite Experiment: Focus on organic/sulfate, lead, soot, and mineral particles. J. Geophys. Res. 107(D1), 4003, doi:10.1029/2000JD000011.


30 June 2004 | State of SOS-3: 1995 - 2003 | 91

2.6. Particle Phase Measurement Technologies - Development and Intercomparison of Techniques

Integrated Filter Methods for Fine Particle Composition

Integrated Filter Measurements of Semi-Volatile Fine Particulates

Semi-Continuous Methods for Measuring Fine Particle Chemical Composition–Atlanta Supersite Intercomparison

Mass Spectroscopic Methods for Measuring Fine Particle Chemical Composition

Methods for Measuring Fine Particle Density

2.6. PARTICLE PHASE MEASUREMENT TECHNOLOGIES - DEVELOPMENT AND INTERCOMPARISON OF TECHNIQUES

Rodney Weber

Measurements of the physical and chemical properties

of aerosols are needed to gain insight into sources, atmospheric transformations, and impacts of ambient

aerosol particles on human health and the environment. Aerosol particle properties of interest include size

distributions, density, optical properties, total mass,

chemical composition, and the aerosol particle mixing state. Methods capable of fast sampling rates and

automated operation are especially useful since they provide large data sets that are better integrated with

meteorological and gas phase measurements and thus

have the potential to provide greater insights than highly time-integrated measurements. Although improvements

are still being made, automated instrumentation to measure aerosol physical properties, such as number

concentrations, size distributions, and optical properties

have existed for some time. It is only recently that significant progress has been achieved in developing online methods for measuring particle chemical composition, much of it with SOS

support. This section focuses on recent developments in aerosol measurement instrumentation

that has been aided by the SOS research program. Following the approach of the Atlanta Supersite Study, instrumentation is divided into three categories, integrated methods, semi-

continuous methods, and methods based on mass spectrometry.

30 June 2004 | State of SOS-3: 1995 - 2003 | 92

2.6.1.1 Assessment of Integrated Filter Methods for Measuring Fine Particle Chemical Composition

Collection of particles onto filter substrates for extended integration periods, followed by off-line

extraction and analysis, is the standard method for measuring fine particle total mass and chemically speciated mass concentrations. One of the findings from the Atlanta Supersite Study

was gained from a unique experiment in which side-by-side comparisons were made between 12 different integrated filter methods for measuring PM2.5 mass and chemical composition. The

results indicate:

• Integrated filter methods showed good agreement for PM2.5 mass (most samplers within ±20%), and sulfate and ammonium (most samplers within 10%).

• There were larger discrepancies between methods (±3 0 to 35%) for nitrate, possibly due to the low ambient concentrations. Higher variability also was found for the organic (OC) and elemental (EC) carbonaceous fractions of PM2.5. For all OC samplers the variability ranged between 35 and 45%. EC variability was high, especially between the different analytical methods.

KEY CITATION: Solomon, P., K. Baumann, E. Edgerton, R. Tanner, D. Eatough, W. Modey, H. Maring, D. Savoie, S.

Natarajan, M. B. Meyer, and G. Norris. 2003. Comparison of integrated samplers for mass and composition during the 1999 Atlanta Supersites project. J. Geophys. Res. 108(D7), 8423, doi:10.1029/2001JD001218.

30 June 2004 | State of SOS-3: 1995 - 2003 | 93

2.6.1.2 Integrated Filter Measurements of Semi-Volatile Fine Particulates by the PC-BOSS

Semi-volatile compounds associated with aerosol particles include ammonium nitrate and

semi-volatile organics that are mainly from secondary organic aerosol formation. Although these compounds can comprise a significant fraction of the PM2.5 mass, they are difficult to

measure accurately due to evaporation loss during sampling. Standard single filter methods, such as the methods currently accepted by the US EPA, including the PM2.5 Federal Reference

Method (FRM), will underestimate semi-volatile compounds. The PC-BOSS sampler has been

developed to quantify the fine particle composition, including the semi-volatile compounds. The instrument has been deployed in a number of studies, including the Atlanta Supersite Study, to

quantify the semi-volatile aerosol compounds. A key finding from the range of studies with the PC-BOSS is that:

• An estimated 10% to 50% of the fine particulate mass was not measured with the PM2.5 FRM sampler due to the loss of semi-volatile organic material and ammonium nitrate during sampling.

These types of methods are necessary to accurately characterize PM2.5 chemical components and

to aid in setting relevant policy standards. KEY CITATIONS: Eatough, D.J., R.W. Long, W.K. Modey, and N.L. Eatough. 2003. Semi-volatile secondary organic aerosol in urban

atmospheres: Meeting a measurement challenge. Atmos. Environ. 37:1277-1292. Modey, W.K., Y. Pang, N.L. Eatough, and D.J. Eatough. 2001. Fine particulate (PM2.5) composition in Atlanta, USA:

Assessment of the particle concentrator-Brigham Young University organic sampling system, PC-BOSS, during the EPA supersite study. Atmos. Environ. 35(36):6493-6502.

Modey, W.K., E.J. Eatough, Y. Pang, and N.L. Eatough. 2004. Performance and evaluation of the PC-BOSS for fine PM2.5 sampling during the summer EPA Supersite Program in Atlanta. J. Air Waste Manage. Assoc. (in press).

30 June 2004 | State of SOS-3: 1995 - 2003 | 94

2.6.2 Advances in Semi-Continuous Methods for Measuring Fine Particle Chemical Composition–Atlanta Supersite Intercomparison

A significant outcome of SOS efforts in instrument development was the assistance given to

a range of methods for automated measurements of particle chemical composition. Five relatively new methods for measurement of bulk fine particle ionic composition, and five for

measurement of carbonaceous compounds were intercompared as part of the Atlanta Supersite Study. These methods included:

Inorganic composition: • Aerosol Dynamics, Inc Integrated Collection and Vaporization Cell (ICVC) for nitrate

and sulfate, now commercially available through R&P (e.g., Rupprecht and Patashnick Sulfate and Nitrate Monitors, Albany NY). (Hering and Stolzenburg, 1997; Stolzenburg and Hering, 2000)

• Netherlands Energy Research Foundation (ECN): nitrate and sulfate. (Slanina et al., 2001) • Georgia Tech/Brookhaven National Lab Particle Into Liquid Sampler (PILS). (Orsini et al.,

2003; Weber et al., 2001). • Atmospheric Research and Analysis (ARA) nitrate. • Texas Tech University (TT): nitrate and sulfate. (Boring et al., 2002)

Carbonaceous Composition • Rutgers University/Oregon Graduate Institute Thermal optical carbon analyzer (OC, EC,

TC). A similar instrument is commercially available (Sunset Labs) • Rupprecht and Patashnick (R&P) 5400 ambient carbon particulate monitor (OC, EC, TC)

(Turpin et al., 1990). • Radiance Research particle soot absorption photometer (PSAP, Seattle WA) (EC). • Aerosol Dynamics Inc integrated collection and vaporization cell (ICVC) for carbon

(OC). • Magee Scientific AE-16 aethalometer (EC).

Two separate publications detailed the performance of the semi-continuous methods based on an unprecedented side-by-side ambient study.

1) Ionic Compounds: Five semi-continuous PM2.5 instruments for measurements of fine particle (PM2.5) nitrate and sulfate were compared. It was found that most instruments were in close agreement with r-squared values between instrument pairs typically ranging from 0.7 to 0.94. Based on comparison between individual semi-continuous devices and 24-hour integrated filter measurements, most instruments were within 20 to 30% for nitrate (~0.1-0.2 µg m-3) and 10 to 15% for sulfate (1-2 µg m-3). Within 95% confidence intervals, linear regression fits suggest no biases existed between the semi-continuous techniques and the 24-hour integrated filter measurements of nitrate and sulfate, however, for nitrate, the semi-continuous intercomparisons showed significantly less variability than intercomparisons amongst the 24-hour integrated filters. (Weber et al., 2003b, JGR)

2) Carbonaceous Compounds: As observed in the integrated filter inter-comparisons, the measurement of the aerosol carbonaceous component is subject to larger uncertainties than the inorganic compounds. Semicontinuous aerosol carbon measurements made during

30 June 2004 | State of SOS-3: 1995 - 2003 | 95

Atlanta using five different samplers showed moderate correlations between pairs of organic carbon measurements and high correlations between pairs of elemental carbon measurements. The two semi-continuous samplers capable of a total carbon (TC) measurement were in good agreement, within 5% based on the slope of a Deming least squares fit, and r-squared of 0.83. Regression slopes between pairs of the OC measurements were within 22%, but in some cases intercepts as large as 2.3 ugC/m3 suggest there were significant systematic differences. Large differences were found between pairs of EC measurements. Regression slopes indicate differences ranged from 4 to 45%. The study demonstrates the successful operation of automated semi-continuous carbon analyzers and illustrates the need for standards to decrease uncertainties in current OC-EC measurements. (Lim et al., 2003, JGR) Many of the semi-continuous methods developed with SOS assistance have had a significant

impact on other federally funded projects, and are likely to impact regulation of PM2.5. Key findings are summarized below.

Data from new semi-continuous instrumentation deployed at the Atlanta Supersite have

enabled the following new insights into aerosol sources and atmospheric processing: • Secondary organic aerosol, which comprised about 46% of the measured organic

carbon, was from a combination of in situ photochemical production and transport of more aged secondary organic aerosol. This is based on diurnal patterns and correlations with ozone and carbon monoxide and estimates of the fraction of OC from secondary organic aerosol formation processes from mean 1-hour fine particle OC and EC data.

• Transient PM2.5 episodes in which particle mass rapidly rises and falls over a period of a few hours but which go undetected with traditional time-integrated measurements are ubiquitous. Continuous highly time resolved measurements of fine particle mass and chemical composition at the EPA Atlanta Supersite Study, August 1999, revealed the transient episodes. Speciated composition data show that these events are driven by sudden increases of two specific aerosol chemical components that dominate at different times, carbonaceous events in the early morning and sulfate events in late afternoon. Apart from providing insights into sources, the unique chemical nature of these transient events may have specific health effects that previous epidemiologic studies based on highly averaged aerosol data could not readily resolve.

• Partitioning between the gaseous and condensed phases was in reasonable agreement with predictions of a thermodynamic equilibrium model. Application of the thermodynamic equilibrium model ISORROPIA to near real-time measurements of fine particle sulfate, nitrate, and ammonium, and gas phase ammonia and nitric acid showed good agreement, with an indication of potential bias in estimates of acidity/alkalinity.

KEY CITATIONS: Lim, H.-J. and B.J. Turpin. 2002. Origins of primary and secondary organic aerosol in Atlanta: Results of time-resolved

measurements during the Atlanta Supersite Experiment. Environ. Sci. Technol. 36:4489-4496.

30 June 2004 | State of SOS-3: 1995 - 2003 | 96

Lim, H.-J., B.J. Turpin, E. Edgerton, S. Hering, G. Allen, H. Maring, and P. Solomon. 2003. Semicontinuous aerosol carbon measurements: Comparison of Atlanta Supersite measurements. J. Geophys. Res. 108(D7), 8419, doi:1029/2001JD001214.

Weber, R., D. Orsini, Y. Duan, K. Baumann, C.S. Kiang, W. Chameides, Y.-N. Lee, F. Brechtel, P. Klotz, P. Jongejan, H. ten Brink, J. Slanina, C.B. Boring, Z. Genfa, P. Dasgupta, S. Hering, M. Stolzenburg, D.D. Dutcher, E. Edgerton, B. Hartsell, P. Solomon, and R. Tanner. 2003a. Intercomparison of near real time monitors of PM2.5 nitrate and sulfate at the U.S. Environmental Protection Agency Atlanta Supersite. J. Geophys. Res. 108(D7), 8421, doi:10.1029/2001JD001220.

Weber, R., D. Orsini, A. Sullivan, M. Bergin, C.S. Kiang, M. Chang, Y.N. Lee, P. Dasgupta, J. Slanina, B. Turpin, E. Edgerton, S. Hering, G. Allen, P. Solomon, and W. Chameides. 2003b. Transient PM2.5 aerosol events in metro Atlanta: Implications for air quality and health. J. Air Waste Manage. Assoc. 53:84-91.

Zhang, J., W.L. Chameides, R. Weber, G. Cass, D. Orsini, E. Edgerton, P. Jongejan, and J. Slanina. 2002. Validity of thermodynamic equilibrium assumption for fine particulate composition: Nitrate and ammonium during the 1999 Atlanta Supersite Experiment. J. Geophys. Res. 108(D7), 8414, 10.1029/2001JD001592.

30 June 2004 | State of SOS-3: 1995 - 2003 | 97

2.6.3 Advances in Mass Spectroscopic Methods for Measuring Fine Particle Chemical Composition

Four particle mass spectrometers operated side-by-side during the Atlanta Supersite Study:

1) Particle Analysis by Laser Mass Spectrometer (PALMS) of the National Oceanic and Atmospheric Administration, 2) University of California at Riverside’s Aerosol Time-of-Flight

Mass Spectrometer (ATOFMS), 3) University of Delaware’s Rapid Single-Particle Mass Spectrometer II (RSMS-II), and 4) Aerodyne’s Aerosol Mass Spectrometer (AMS). Of these,

the PALMS and ATOFMS were fairly well established techniques, whereas the RSMS-II and

AMS were in earlier stages of development and thus the Atlanta Supersite Study provided a valuable opportunity to assess these instrument’s performance. Key findings are summarized

below.

• Particle sizes were measured most accurately with ATOFMS, RSMS-II, and AMS. The RSMS-II system can obtain composition information on individual particles as small as 15 nm. The three systems that utilize laser desorption/ionization, (PALMS, ATOFMS, and RSMS-II), produce mass spectra that are qualitative and representative of individual particles. The AMS instrument, which uses a two-step volatilization on a heated surface and ionization by electron impact, can produce quantitative results representative of the ensemble of particles measured.

• Single-particle positive ion classifications from the Atlanta data by the laser-based instruments are broadly consistent and revealed similar trends as a function of size for organic, sulfate, and mineral particles. The AMS, which is the most quantitative of the mass spectrometers compared, had nitrate to sulfate molar ratios that were highly correlated with those of the semi-continuous instruments discussed above. Based on insights from the Atlanta study, subsequent studies, such as those undertaken at the New York EPA Supersite in the summer of 2002 (PEMTACS), demonstrated the quantitative measurement capabilities of the AMS. Overall, the strength and primary focus of the laser-based instruments are their ability to find associations between chemical species in individual particles with high time resolution.

KEY CITATIONS: Drewnick, F., J.J. Schwab, O. Hogrefe, S. Peters, L. Husain, D. Diamond, R. Weber, and K. Demerjian. 2003.

Intercomparison and evaluation of four semi-continuous PM2.5 sulfate instruments. Atmos. Environ. 37(2003):3335-3350.

Middlebrook, A.M., D.M. Murphy, S.-H. Lee, D.S. Thomson, K.A. Prather, R.J. Wenzel, D.-Y. Liu, D.J. Phares, K.P. Rhoads, A.S. Wexler, M.V. Johnston, J.L. Jimenez, J.T. Jayne, D.R. Worsnop, I. Yourshaw, J.H. Seinfeld, and R.C. Flagan. 2003. A comparison of particle mass spectrometers during the 1999 Atlanta Supersite Project. J. Geophys. Res. 108(D7), 8424, doi:10.1029/2001JD000660.

30 June 2004 | State of SOS-3: 1995 - 2003 | 98

2.6.4 Advances in Methods for Measuring Fine Particle Density A new technique that measures the density of aerosol particles in the diameter range 0.1 to 0.3

um was developed and used during the Atlanta Supersite Study. The technique offers an

alternative to calculating density from measured aerosol composition. Particles are selected based on known electrical mobility and then measurements of their mass are made with an

aerosol particle mass analyzer; density is determined based on geometric diameter, which is

equal to electrical mobility equivalent diameter for spherical particles, and mass. • Particles measured in August 1999 in urban Atlanta typically included a major

mass peak with a density in the ~1.5 to 1.7 g cm-3 range at 3-6% relative humidity. These data agreed with calculated densities based on measured size-resolved composition within about 5%.

KEY CITATION: McMurry, P.H., X. Wang, K. Park, and K. Ehara. 2002. The relationship between mass and mobility for atmospheric

particles: A new technique for measuring particle density. Aerosol Sci. Techol. 36:227-238.

30 June 2004 | State of SOS-3: 1995 - 2003 | 99

2.7. Gas Phase Measurement Technologies - Development and Intercomparison of Techniques

Chemical ionization & proton transfer mass spectrometry

Odd hydrogen radical measurements

NMHC measurements Formaldehyde measurements

Carbonyl measurements NO2 and NOy measurements

Organic nitrate measurements CO measurements

2.7. GAS PHASE MEASUREMENT TECHNOLOGIES - DEVELOPMENT AND INTERCOMPARISON OF TECHNIQUES

Eric Apel and David Parrish

Full understanding of the photochemistry that

produces ozone and PM2.5 requires ambient

measurements of the precursors, intermediates (radicals

as well as more stable species), and products. These

measurements must be made with high and well-defined

sensitivity, accuracy, precision, and specificity, and with

fast time response (particularly ~ 1 Hz for aircraft

measurements). The achievement of these goals is an

evolutionary process requiring careful instrument

development and operation in the field. This section

summarizes the major progress of the SOS research

program in this area.

30 June 2004 | State of SOS-3: 1995 - 2003 | 100

2.7.1 Major Advance: Chemical Ionization and Proton Transfer Mass Spectrometry The combination of mass spectrometry with chemical ionization (CIMS) potentially provides

a sensitive and specific measurement technique for many atmospheric species. SOS-sponsored

work includes the development, testing, and deployment of CIMS techniques for the

measurement of HNO3, isoprene, and ammonia as well as the deployment and testing of proton

transfer (PTR-MS) instruments. The PTR-MS instruments can detect most gas-phase organic

species with excellent time response, but their sensitivity and specificity are limited for many

compounds by the manifold of organic species and fragment ions with similar masses.

Highlights of this work are summarized below.

• Specific HNO3 Measurement Developed. A CIMS instrument to measure gas-phase HNO3 was developed and demonstrated to be sensitive with fast response (detection limit of approximately 10 pptv at 1 Hz), accurate, precise, and interference-free. It has been tested in a ground-based intercomparison and deployed on aircraft during SOS 1999 and 2000 field intensives.

• PTR-MS Deployed and Tested. PTR-MS instruments have been deployed at ground sites and on aircraft along with other instruments for measurement of several organic species. Intercomparison of the coincident measurements will help to define the capabilities of the PTR-MS instruments.

• CIMS Isoprene and Ammonia Measurements Developed. CIMS techniques for measurement of isoprene and ammonia have been developed and tested in ground-based intercomparisons. These techniques promise to provide sensitive and fast aircraft measurements of those species. The selectivity of the isoprene technique must be tested by comparison to other techniques.

KEY CITATIONS: Huey, L.G., E.J. Dunlea, E.R. Lovejoy, D.R. Hanson, R.B. Norton, F.C. Fehsenfeld, and C.J. Howard. 1998.

Fast time response measurement of HNO3 in air with a chemical ionization mass spectrometer. J. Geophys. Res. 103:3355-3360.

Fehsenfeld, F.C., L.G. Huey, D.T. Sueper, R.B. Norton, E.J. Williams, F.L. Eisele, R.L Mauldin, III, and D.J. Tanner. 1998. Ground-based intercomparison of nitric acid measurement techniques. J. Geophys. Res. 103:3343-3353.

Leibrock, E., and L.G. Huey. 2000. Ion chemistry for the detection of isoprene and other volatile organic compounds in ambient air. Geophys. Res. Letters 19:1763-1766.

30 June 2004 | State of SOS-3: 1995 - 2003 | 101

2.7.2 Major Advances: Odd Hydrogen Radical Techniques The hydroxyl radical (OH), the main oxidant in atmospheric chemistry, cycles rapidly with

the hydroperoxyl radical (HO2), initiating the production of ozone and other pollutants. A now

widely accepted measurement technique for OH and HO2, together called HOX, is laser-induced

fluorescence in detection chambers at low pressure. This technique was applied for the first time

to continuous, 24-hour measurements on 10-meter tall towers in Nashville in 1999 and at

TEXAQS 2000 in Houston. In addition, a new, unique instrument was developed to measure the

OH loss rate due to reactions with other atmospheric chemicals. This new instrument, the Total

OH Loss-rate Measurement (TOHLM), was successfully deployed for the first time in Nashville

in 1999. The combination of HOX measurements and TOHLM provides powerful new

diagnostics for understanding and testing the oxidation chemistry of any environment.

• First continuous OH and HO2 measurements in urban environments. The measurements, when compared to models, test the fundamental atmospheric chemistry that underpins chemical transport models. For Nashville SOS, OH and HO2 measurements agree to within a factor of two with model calculations near midday, but tend to be larger than models in the evening, at night, and for periods when nitrogen oxides are especially abundant. These observations indicate unidentified HOx sources and questions about HOx-NOx chemistry.

• Total OH Loss-rate Measurement tests the completeness of measured VOC inventories. The presence of unmeasured VOC is indicated if TOHLM-measured OH loss rates are greater than those calculated from the sum of VOC measurements and OH reaction rate coefficients. Preliminary Nashville observations indicate that OH loss rates are about twice those calculated, suggesting unmeasured VOC.

KEY CITATION: Kovacs, T.A. and W.H. Brune. 2000. Total OH loss rate measurement. J. Atmos. Chem. 39(2):105-122,

doi:10.1023/A:1010614113786.

30 June 2004 | State of SOS-3: 1995 - 2003 | 102

2.7.3 Non-Methane Hydrocarbons (NMHCs): Intercomparisons of Techniques NMHCs, emitted from a variety of sources, are precursors to ozone and PM2.5, and some

species are considered toxins. NMHCs are the primary photochemical fuel in urban and many

rural areas whereas carbon monoxide and methane play this role in regions remote from sources.

Biogenic NMHCs dominate the VOC chemistry in many highly forested regions and even in

urban areas situated in regions surrounded by forests, as is the case in most cities located in the

SOS domain (southeastern U.S.) [e.g., Chameides et al., 1997]. Anthropogenic NMHCs

dominate in most other urban areas.

• Accuracy of NMHC Measurements Tested. Because of the large role of NMHCs in ozone formation, it is imperative that measurements accurately reflect the true atmospheric composition. SOS has taken a leadership role in the atmospheric science community and has partnered with the NOAA Climate and Global Change Program in conducting “The Non Methane Hydrocarbon InterComparison Experiment” (NOMHICE). This experiment was designed to assess the accuracy of analytical methods used to determine mixing ratios of atmospheric non-methane hydrocarbons (NMHCs).

• NMHC Measurements Intercompared in the SOS Network. To help ensure quality measurements and to understand where there are problems, intercomparisons of NMHC measurements are conducted in the SOS network before and during every field study. Prior to the Nashville 99 Field Study nine measurements were intercompared. For each measurement, a set of canisters were simultaneously collected at the Cornelia Ft. site in Nashville, TN and distributed to 4 groups for analysis. Figure 2.6.1 shows the mean ratio of each laboratory’s analysis of a given compound to that of the reference laboratory (EPA-Lonneman). It is apparent that for some species, the accuracy of the analyses is good (agreement to within 11% for all groups), but for some species there were errors up to a factor of 5. Some of the problems could be corrected during the experiment. For example, the discrepancy of the Argonne data for i-pentane was determined to be due to an overlap with another peak and corrected. The precision of the measurements of the individual species by the different laboratories is indicated by the standard deviations of the ratios. These standard deviations ranged from very good (±3%) to quite poor (factor of 3).

• High Quality Multi-Component Standards Developed. Eight identical standards containing 100 NMHC were developed and tested for the TexasAQS 2000 study. These standards were used as a common calibration source for all participants in the study.

30 June 2004 | State of SOS-3: 1995 - 2003 | 103

0.1

2

3

4

5

6

1

2

3

4

5

Rati

o t

o E

PA

- L

onnem

an

EPA - Mantech

Ohio U.

NCAR

Argonne

eth

ane

i-buta

ne

pro

pane

n-b

uta

ne

i-penta

ne

n-p

enta

ne

acety

lene

eth

yle

ne

pro

pyle

ne

1-b

ute

ne

isopre

ne

benzene

tolu

ene

eth

yl benzene

m,p

-xyle

ne

o-x

yle

ne

1,3

,5 t

rim

eth

yl

1,2

,4 t

rim

eth

yl

1,2

,3 t

rim

eth

yl

Figure 2.7.1. Comparison of the mean ratios of selected NMHC measurements from 4 different SOS measurement groups using EPA-Lonneman results as reference.

KEY CITATIONS: Chameides, W.L., R.D. Saylor, and E.B. Cowling. 1997. Ozone pollution in the rural United States and the new

NAAQS. Policy Forum. Science 276:916. Apel, E.C., J.G. Calvert, T.M. Gilpin, F.C. Fehsenfeld, D.D. Parrish, and W.A. Lonneman. 1999. The Non-

Methane Hydrocarbon Intercomparison Experiment (NOMHICE): Task 3. J. Geophys. Res. 104:26069-26086.

30 June 2004 | State of SOS-3: 1995 - 2003 | 104

2.7.4 Formaldehyde Technique Development and Intercomparisons Formaldehyde (CH2O) is a primary emission product from internal combustion engines and

is produced in the atmosphere by the photochemical oxidation of methane and non-methane

hydrocarbons (NMHCs). It is the most abundant gas-phase carbonyl compound in both urban

areas and the remote troposphere. Formaldehyde is extensively connected with the odd

hydrogen (HOx = H+OH+HO2) and odd nitrogen (NOx =NO + NO2) chemical cycles. It is also a

major source for HO2 and for CO in air not strongly perturbed by anthropogenic sources.

Consequently, accurate measurements of formaldehyde are critical to understanding the overall

tropospheric chemistry leading to hydrocarbon oxidation, the processes controlling the odd

hydrogen cycles and nitrogen cycles, and the global budgets of OH and CO.

• Two Measurement Techniques Developed. SOS has encouraged the development of fast and sensitive techniques to measure formaldehyde. Two techniques, tunable diode laser absorption spectroscopy (TDLAS) and coil/2,4-dinitrophenylhydrazine (CDNPH), have emerged which satisfy the necessary criteria for ground-based and aircraft-based measurements. These techniques have been used extensively in SOS field studies. The TDLAS technique can provide 1-second averages, while the CDNPH technique has been limited to averages of at least 1 minute.

• Two Field Intercomparisons Completed. A blind intercomparison of six ambient formaldehyde measurement techniques was conducted at the National Center for Atmospheric Research (NCAR) in Boulder, CO, from May 29 to June 3, 1995. It was concluded that gas phase standards should be employed with any of the measurement techniques, and the cartridge measurement methods are limited by long collection periods, generally lower precision, and the incomplete understanding of potential interferences from ozone and possibly other compounds. Airborne CH2O measurements by TDLAS and CDNPH techniques indicated that, on average, both instruments measured identical ambient CH2O concentrations to better than 0.1-ppbv over the 0 to 0.8-ppbv-concentration range. However, significant differences, larger than the combined 2σ total uncertainty estimates, were observed in 29% of data set. It is clear that careful attention must be paid to the behavior of CH2O in the inlet for accurate airborne measurements.

• CH2O Gas Phase Standards Developed. Through SOS support, formaldehyde standards have been produced in high-pressure cylinders at the ppmv level. Techniques have been developed, including FTIR and GC-FID, to verify the concentration of the standard. Long-term stability tests are presently being conducted.

KEY CITATION: Gilpin, T., E. Apel, A. Fried, B. Wert, J. Calvert, Z. Genfa, P. Dasgupta, J.W. Harder, B. Heikes, B. Hopkins,

H. Westberg, T. Kleindienst, Y.-N. Lee, X. Zhou, W. Lonneman, and S. Sewell. 1997. Intercomparison of six ambient [HCHO] measurement techniques. J. Geophys. Res. 102:21,161-21,188.

30 June 2004 | State of SOS-3: 1995 - 2003 | 105

2.7.5 Carbonyl Technique Development and Intercomparisons Aldehydes and ketones and other oxygenates are produced by the oxidation of hydrocarbons,

and some are emitted directly. Until recently, this compound class has received much less

attention than its VOC counterpart, the NMHCs. This is largely due to difficulties encountered

in measuring these compounds. SOS has taken a leading role in developing carbonyl standards,

carbonyl measurement techniques, and evaluating the techniques with intercomparisons.

Through these studies, it is becoming apparent that the carbonyls and other oxygenates may be

more ubiquitous than previously thought, and hence contribute significantly to photochemical

processes in the troposphere.

• The first quantified and verified carbonyl standards. Carbonyl standards have been prepared gravimetrically with both calibrated permeation sources and in specially treated aluminum cylinders. Techniques such as atomic emission detection (AED) (Apel et al., 1998a) and FTIR have been applied to verify the accuracy of the standards. Table 2.6.1. demonstrates the verification of a prepared standard.

Table 2.6.1. Quantification of Standards Results Compound High Pressure Cylinder

gravimetric (ppm) High Pressure Cylinder

analyzed (ppm)* FTIR (ppm)

Methanol 3.00 ± 0.03 3.0 ± 0.1 2.9 ± 0.2 Ethanol 3.02 ± 0.03 n.d. 2.9 ± 0.3 Acetone 3.03 ± 0.03 3.03 ± 0.08 3.2 ± 0.2 MEK 3.04 ± 0.03 n.d 2.5 ± 0.3 Acetaldehyde 3.03 ± 0.03 n.d 3.2 ± 0.4 Propanal 3.03 ± 0.03 n.d n.d. Butanal 3.02 ± 0.03 n.d n.d.

*analysis based on calibration factors from permeation tubes n.d. - not determined • Intercomparison completed. Cartridge–based (Si-Gel and C18) and GC-MS

measurements have been intercompared through SOS. Serious discrepancies were found and more work is needed to resolve these differences.

• New techniques developed. A new relatively fast-response (15 minute cycle) GC-FID technique has been developed to measure carbonyls and other oxygenates aboard aircraft. A new GC-MS technique is currently being developed to measure carbonyls with a 5-minute time response.

KEY CITATIONS: Apel, E.C., J.G. Calvert, J.P. Greenberg, D. Riemer, R. Zika, T.E. Kleindienst, W.A. Lonneman, K. Fung,

and E. Fujita. 1998a. Generation and validation of oxygenated volatile organic carbon standards for the 1995 Southern Oxidants Study Nashville Intensive. J. Geophys. Res. 103:22,281-22,294.

Apel, E.C., J.G. Calvert, D. Riemer, W. Pos, R. Zika, T.E. Kleindienst, W.A. Lonneman, K. Fung, E. Fujita, P.B. Shepson, T.K. Starn, and P.T. Roberts. 1998b. Measurements comparison of oxygenated volatile organic compounds at a rural site during the 1995 SOS Nashville Intensive. J. Geophys. Res. 103:22,295-22,316.

30 June 2004 | State of SOS-3: 1995 - 2003 | 106

2.7.6 NO2 and NOy Technique Development and Intercomparisons. The oxides of nitrogen (NO and NO2) are precursors of ozone and PM2.5, and the total

oxidized nitrogen family (NOy) in an air parcel represents the total emissions of these precursors

that remain in the atmosphere. Improvements in the measurement of these species are

summarized below.

• Improved NO2 measurements by photolysis-chemiluminescence. A new photolytic converter utilizing the focused UV output from a high-pressure mercury (Hg) arc lamp was developed and tested. The new configuration permits simple and accurate retrieval of ambient NO2 data at very high time resolution, is more specific, provides increased sensitivity, and is less expensive to operate than previous photolytic converter designs.

• Validation of an aircraft inlet for HNO3 and NOy. Rapid and quantitative sampling of NOy species, including HNO3, has been demonstrated using a short, heated Teflon inlet. In flight, standard addition calibrations of HNO3 at the aircraft inlet demonstrate freedom from significant surface adsorption of HNO3, which has significantly compromised measurements through other aircraft inlets.

• Intercomparisons of ground-based NO2 and NOY measurements. Intercomparisons during SOS field intensives have demonstrated that laser-induced fluorescence, differential optical absorption spectroscopy, and photolysis-chemiluminescence techniques are all capable of accurately quantifying atmospheric NO2 above 1 ppbv. Further, MoO and Au converters were shown to be capable of accurately measuring NOy above 2 ppbv in the urban and suburban environments typical of the SOS region. These studies further concluded that generation of reliable NO2 or NOy data still demands skilled operators and dedicated, critical oversight during the measurement process.

KEY CITATIONS: Ryerson, T. B., E.J. Williams, and F.C. Fehsenfeld. 2000. An efficient photolysis system for fast-response NO2

measurements,. J. Geophys. Res. 105:26,447-26,461. Ryerson, T.B., L.G. Huey, K. Knapp, J.A. Neuman, D.D. Parrish, D.T. Sueper, and F.C. Fehsenfeld. 1999.

Design and initial characterization of an inlet for gas-phase NOy measurements from aircraft, J. Geophys. Res. 104:5483-5492.

Williams, E.J., K. Baumann, J.M. Roberts, S.B. Bertman, R.B. Norton, F.C. Fehsenfeld, S.R. Springston, L.J. Nunnermacker, L. Newman, K. Olszyna, J. Meagher, B. Hartsell, E. Edgerton, J.R. Pearson, and M.O. Rodgers. 1998. Intercomparison of ground-based NOy measurement techniques. J. Geophys. Res. 103:22,261-22,280.

30 June 2004 | State of SOS-3: 1995 - 2003 | 107

2.7.7 Organic Nitrate Technique Development Several advances have been made in the measurements of organic nitrates through the course

of SOS. Emphasis has been put on improving separation and quantitation of compounds that are

of biogenic origin, and on developing rapid methods for aircraft measurements.

• Measurements of peroxyacyl nitrates (PANs). Gas chromatographic methods for PANs have been refined to provide rapid and sensitive measurement. Aircraft-based instrumentation was developed to measure four of the major compounds of interest, PAN, PPN, PiBN, and MPAN, every 3.5 min. The measurement of PANs by proton-transfer reaction mass spectrometry (PTR-MS) was deployed during the Nashville 99 intensive (Hansel and Wisthaler, 2000). While still in the development stage, this method has the potential to provide rapid (10 sec) measurements of PAN aboard aircraft.

• Development of PAN calibration systems. Two different calibration methods for PAN have been developed: a diffusion source and photochemical production of PAN in acetone/air/CO/NO mixtures. The diffusion system relies on an NOy measurement for calibration, while the photochemical source relies on a known, efficient conversion of an NO standard to PAN. The two were compared during the TEXAQS 2000 study and were found to agree within 5%.

• Development of organic nitrate measurements. An automated system for the measurement of the organic nitrates produced from OH radical attack on isoprene was developed and deployed at the Dickson site during SOS 99. These compounds result when the peroxy radicals derived from OH reaction with isoprene, react with NO to produce a set of 8 isomeric RONO2 species. The maximum concentrations of the sum of these species were in the 100-200 pptv range, much higher than observed in a previous study. Comparison of these two data sets provides a good opportunity to examine the NOx-dependence of this aspect of isoprene photochemistry.

KEY CITATIONS: Williams, J., J.M. Roberts, S.B. Bertman, C.A. Stroud, F.C Fehsenfeld, K. Baumann, M.P. Buhr, K. Knapp,

P.C. Murphy, M. Nowick, and E.J. Williams. 2000. A method for the airborne measurement of PAN, PPN and MPAN. J. Geophys. Res. 105(D23):28,943–28,960.

Hansel, A. and A. Wisthaler. 2000. A method for real-time detection of PAN, PPN, and MPAN in ambient air. Geophys. Res. Lett. 27:895-898.

Grossenbacher, J.W., P.B. Shepson, T. Thornberry, M. Witmer-Rich, M.A., Carroll, I. Faloona, D. Tan, W. Brune, E. Apel, D. Riemer, and H.A. Westberg. 2000. Measurements of isoprene nitrates above a forest canopy. J. Geophys. Res. 109, D11311, doi:10.1029/2003JD003966.

30 June 2004 | State of SOS-3: 1995 - 2003 | 108

2.7.8 Carbon Monoxide Technique Development and Intercomparisons Carbon monoxide is a long-lived gas primarily emitted in automobile exhaust, which makes

it a useful tracer for urban pollution plumes. The result of an SOS effort to develop an

instrument for aircraft measurements of this species is summarized below.

• Development of an instrument based upon vacuum UV resonance fluorescence of CO. An instrument was developed that is capable of fast (~1 Hz), accurate (~5%), and precise (~1 ppbv) measurement of CO from an aircraft platform. Intercomparisons with other techniques demonstrate that it is highly specific with no identified interferences (see Figure 2.7.2).

7

6

5

4

3

2

1

0

Altitu

de

(km

)

14:00 16:00 18:00 20:00

UTC 2 September 1997

300

250

200

150

100

50

0

CO

(p

pb

v)

1

2 3

4

Figure 2.7.2. Time series of coincident 5-second-average measurements of CO. The solid lines give tunable diode laser absorption (darker) and the vacuum ultraviolet fluorescence (lighter) results and the dotted line indicates the aircraft altitude. The features labeled 1 through 4 are intercepted urban plumes. KEY CITATION: Holloway, J. S., R.O. Jacoubek, D.D. Parrish, C. Gerbig, A. Volz-Thomas, S. Schmitgen, A. Fried, B. Wert, B.

Henry, and J.R. Drummond. 2000. Airborne intercomparison of vacuum ultraviolet fluorescence and tunable diode laser absorption measurements of tropospheric carbon monoxide. J. Geophys. Res. 105:24,251-24,261.

30 June 2004 | State of SOS-3: 1995 - 2003 | 109

2.8. Atmospheric Dynamics and Mixing on Urban and Regional Scales

Daytime transport processes Nighttime transport processes

Variations in mixing height Vertical distribution of

ozone, precursors, and aerosols

Off- to on-shore flow reversal Heterogeneity of industrial

plumes

2.8. ATMOSPHERIC DYNAMICS AND MIXING ON URBAN AND REGIONAL SCALES

Allen White and Christoph Senff

While a thorough understanding of the atmospheric

chemistry associated with the formation and destruction

of atmospheric pollutants is critical in air quality

research, knowledge of atmospheric dynamics and

mixing processes is equally important to gain a full

understanding of the problem. Vertical and horizontal

transport mechanisms as well as the temporal and spatial

evolution of mixed layer height often play a critical role

in the distribution of pollutants on urban and regional

scales. In this section, we summarize key scientific

findings from the SOS’ Nashville ’95 and Nashville ‘99

studies, and TexAQS 2000 that pertain to atmospheric dynamics and mixing processes.

2.8.1. Daytime Transport Processes During SOS field studies, a number of meteorological processes contributed to horizontal

and vertical transport of pollutants during the daytime. A schematic summarizing many of these

processes is shown in Fig. 2.8.1. SOS used a variety of ground-based and airborne observing

systems to study transport mechanisms. Key findings are summarized below.

• Horizontal Advection. We found that synoptically driven winds were the dominant daytime horizontal transport mechanism. Mesoscale circulations caused by topography or land use differences also contribute to daytime transport. During TexAQS 2000, synoptic flow exported the Houston/Ship Channel and Texas City pollution plumes to rural, source-free areas, resulting in ozone concentrations well above the ozone standard far downwind of the Houston metropolitan area. Many of these ozone exceedances were missed by the surface monitoring network due to sparse network coverage in rural areas.

• Boundary-layer Venting. Under light wind conditions, we observed substantial (~40%) horizontal variations in daytime mixing height due to the urban-rural contrast in the surface energy balance (the “urban dome”). The dome allowed venting of urban emissions aloft, making them available for horizontal transport but unavailable for vertical mixing downwind of the dome during the day (refer to D in Fig. 2.8.1).

• Convection. Cumulus clouds vented pollutants from the boundary layer and reduced the sunlight available for photochemistry. Because direct measurements of cumulus venting are difficult to obtain experimentally, we were unable to quantify this process during SOS’ Nashville ‘95 and Nashville ’99 studies or during TexAQS 2000. Deep vertical

30 June 2004 | State of SOS-3: 1995 - 2003 | 110

mixing associated with convective storms may have resulted in stratosphere/troposphere exchange.

Figure 2.8.1. Transport between the surface and the atmospheric boundary layer (ABL) and between the ABL and free troposphere: A. fluxes from nearly homogeneous surfaces to ABL, B. fluxes from inhomogeneous surfaces to ABL, C. transport across capping inversion (Zi) by entrainment/detrainment and cumulus venting, and D. direct injection to troposphere caused by horizontal variations in boundary-layer depth.

• Subsidence. Synoptic scale subsidence associated with high pressure strengthened the boundary-layer capping inversion, thereby inhibiting vertical transport of momentum and pollutants and cumulus convection. This behavior, combined with the stagnant conditions resulting from relaxation of the synoptic-scale pressure gradient, allowed pollutants to accumulate locally during the day (Banta et al., 1998).

• Morning Transition. The morning transition caused photochemically aged pollutants from the residual layer to interact with pollutants emitted at night into the shallow nocturnal boundary layer. During the 1999 SOS Nashville Intensive, the breakup of the nocturnal inversion occurred at an urban site 1-2 h earlier than at three rural sites. In the humid environment in the SOS field studies, surface water vapor mixing ratio was often an excellent meteorological tracer for the timing of the morning transition (Fig. 2.8.2).

Figure 2.8.2. 1-min time series of CO and NOy concentrations, the concentration ratio of NOx to NOy, and water vapor mixing ratio (rmix) measured near the surface on 15 July 1999 at the Dickson, Tennessee site. The early morning increase in rmix is due to surface evaporation. The sharp transition in rmix near 0920 CDT occurs after the nocturnal inversion is fully eroded and as the mixed layer grows rapidly through the residual layer and entrains drier air from aloft. The CO and NOy concentrations decrease because nocturnal surface emissions of these gases were confined to the shallow nocturnal boundary layer. The NOx/NOy ratio decreases because the residual layer contains photochemically aged air from the previous day. KEY CITATIONS: Banta, R.M., C.J. Senff, A.B. White, M. Trainer, R.T. McNider, R.J. Valente, S.D. Mayor, R.J. Alvarez, R.M.

Hardesty, D. Parrish, and F.C. Fehsenfeld. 1998. Daytime buildup and nighttime transport of urban ozone in the boundary layer during a stagnation episode. J. Geophys. Res. 103:22,519–22,544.

White, A.B., B.D. Templeman, W.M. Angevine, R.J. Zamora, W.W. King, C.A. Russell, R.M. Banta, W.A. Brewer, and K.J. Olszyna. 2002. Regional contrast in morning transitions observed during the 1999 Southern Oxidants Study Nashville/Middle Tennessee Intensive. J. Geophys. Res. 107(D23), 4726, doi:10.1029/2001JD002036.

30 June 2004 | State of SOS-3: 1995 - 2003 | 111

2.8.2. Nighttime Transport Processes During the night, the effect of surface friction is confined to a shallow layer (10s of meters)

near the surface. In general, vertical motions not associated with convective storms are

considerably weaker at night than during the day. Thus, during SOS, nocturnal transport tends to

redistribute pollutants horizontally rather than vertically. Key features are summarized below.

• Low-Level Jet. At night, the winds above a shallow (10s of meters) layer at the surface accelerated as the atmosphere decouples from the surface.

• Inertial Oscillation. The nocturnal winds rotated in time in accordance with the principles of the inertial oscillation. McNider et al. (1998) demonstrated the persistent nature of this phenomenon using wind spectra obtained from wind profilers deployed during the SOS95 Nashville Intensive. Under sufficiently weak synoptic forcing, the low-level jet and inertial oscillation dominated nocturnal transport. Trajectories derived from the wind profiler network deployed during SOS95 demonstrated the combined effect of these important features (see Fig. 2.8.3).

• Vertical transport is suppressed at night. In the absence of convective storms, the atmosphere stabilized at night, which suppressed any significant vertical transport. In many cases, intermittent turbulence has been observed in the nocturnal boundary layer, which may be linked to wind shear associated with the low-level jet. The effect of intermittent turbulence on pollution levels at the surface is an important topic of current SOS research.

Figure 2.8.3. Overnight forward trajectories calculated from the network of 915-MHz boundary-layer wind profilers deployed during SOS95. Winds were averaged over the 400-m vertical intervals shown in the key. Trajectories were calculated from an origin centered on Nashville, at 36.2Ε N, 86.8Ε W (after Banta et al., 1998).

KEY CITATIONS: Banta, R.M., C.J. Senff, A.B. White, M. Trainer, R.T. McNider, R.J. Valente, S.D. Mayor, R.J. Alvarez, R.M.


McNider, R.T., W.B. Norris, R.I. Clymer, S. Gupta, R.M. Banta, R.J. Zamora, A.B. White, and M. Trainer. 1998. Meteorological conditions during the 1995 Southern Oxidants Study Nashville/Middle Tennessee Field Intensive. J. Geophys. Res. 103:22,225–22,243.

-160 -140 -120 -100 -80 -60 -40 -20 0 20 40

East-West Distance (km)

-60

-40

-20

0

20

40

60

80

100

120

140

Nort

h-S

out

h D

ista

nce

(km

)

Profiler ForwardTrajectories

7/12/95 2000 CDT -7/13/95 1100 CDT

400-800 m

800-1200 m

1200-1600 m

1600-2000 m

30 June 2004 | State of SOS-3: 1995 - 2003 | 112

2.8.3. Variations in Mixing Height During the SOS95 and SOS99 studies, it was found that the daytime mixing height can vary

considerably in the Nashville, TN area, especially when comparing the urban and adjacent rural

areas. As mixing height is an important parameter affecting air pollution concentrations, this

finding has significant implications for the regional distribution of ozone and particulates. Key

results are summarized below.

• Remote sensors provide reliable measurements of mixing height. A comparison of mixed layer depth estimates deduced from wind profiler and airborne lidar data showed very good agreement under clear or partly cloudy conditions (White et al., 1999). This result confirmed that radar wind profilers and lidars are well suited to measure the depth of the mixed layer and its variability.

• Mixing height variability is tied to land use differences. Variations in mixing height are related to differences in surface characteristics, such as soil and vegetation type as well as surface moisture (see Fig. 2.8.4). During SOS99 Nashville Intensive, these different surface characteristics were reflected in varying energy, ozone, and carbon fluxes at the surface.

Figure 2.8.4. Time height cross section of aerosol backscatter gradient indicating the top of the mixed layer measured with the NOAA/ETL airborne lidar during midday on 12 July 1995 during a northwest to southeast transect over Nashville, TN (city of Nashville is in center of panel). The urban heat island over Nashville is clearly visible at flight times near 11:30 LST. Over forested areas to the northwest of Nashville (left side of panel) the mixed layer is strongly suppressed, while over suburban and agricultural terrain to the southeast (right side of panel) the mixed layer is only slightly shallower than over the city.

• Differences in mixing height are most pronounced under light wind conditions. Under stagnant conditions, air parcels tend to dwell over regions of one surface type, which allows surface heating differences to express themselves as variations in mixing height. Stronger flow moves air parcels over many surface types, thus producing a more uniform mixing height

• Urban heat island. The strong differences in surface heating between the Nashville urban area and the surrounding agricultural and forested areas resulted in significantly deeper mixed layers over the city, especially under stagnant conditions. We frequently observed urban mixing heights of 2 km or more, which was as much as 800 m higher than the mixing heights over adjacent rural areas (see Figs. 2.8.4 and 2.8.5).

30 June 2004 | State of SOS-3: 1995 - 2003 | 113

Figure 2.8.5. Hourly measurements of mixed layer depth on 4 July 1999 from the wind profiler network deployed in and around Nashville, TN during SOS 99. The wind profiler at Cornelia Fort airport shows a maximum mixed layer depth of about 2.1 km while all other profilers deployed in rural areas around Nashville detect peak mixed layer heights of only about 1.5 km.

Time (LST) • Ozone concentrations are anti-correlated with mixing height. Peak ozone

concentrations in the Houston/Ship Channel pollution plume downwind of the sources were found to be anti-correlated with mixing height. In the Houston area, mixing depth typically increases with distance away from the coast. Thus, transport of the Houston/Ship Channel pollution plume to coastal areas tends to produce higher ozone peak values than transport to inland areas.

• Model prediction of mixing depth. We found that the Pennsylvania State National Center for Atmospheric Research Mesoscale Model 5 (PSU/NCAR MM5) had difficulty accurately predicting mixing depths under both stable and unstable atmospheric conditions. Deficiencies in the atmospheric radiation parameterizations caused excess energy in the model’s surface layer (see Fig. 2.8.6). MM5’s response to this additional input resulted in daytime mixed layers that were too deep.

Figure 2.8.6. Solar radiative flux predicted by MM5 (line) and measured with a spectral pyranometer (asterisks) at New Hendersonville, Tennessee during SOS95. The model bias leads to model errors in the surface energy budget which impact the depth and strength of vertical mixing and thermally driven circulations such as the land-sea breeze. KEY CITATIONS: Angevine, W.M., A.B. White, C.J. Senff, M. Trainer, R.M. Banta, and M.A. Ayoub. 2003. Urban-rural contrasts

in mixing height and cloudiness over Nashville in 1999. J. Geophys. Res. 108(D3), 4092, doi:1029/2001DJ0001061.

White, A.B., C.J. Senff, and R.M. Banta. 1999. A comparison of mixing depths observed by ground-based wind profilers and an airborne lidar. J. Atmos. Oceanic Technol. 16, No. 5:584-590.

Zamora, R.J., S. Solomon, E.G. Dutton, J.W. Bao, M. Trainer, R.W. Portmann, A.B. White, D.W. Nelson, and R.T. McNider. 2003. Comparing MM5 radiative fluxes with observations gathered during the 1995 and 1999 Nashville Southern Oxidants Studies. J. Geophys. Res. 108(D2), 4050, doi:10.1029/2002JD002122.

30 June 2004 | State of SOS-3: 1995 - 2003 | 114

2.8.4. Vertical Distribution of Ozone, Precursors, and Aerosols The vertical distribution of ozone, precursors, and aerosols is influenced to a large extent by

the dynamical processes described above. The key findings are:

• The vertical distribution of pollutants in the daytime mixed layer. Mixing processes due to convective or mechanical turbulence acted to smooth out vertical inhomogeneities in the daytime boundary layer. Figure 2.8.7 shows a high-ozone layer near the surface mixing vertically as convective turbulence increased over the course of the morning.

Figure 2.8.7. Series of vertical profiles of ozone concentration measured with the NOAA/ETL airborne ozone lidar on the morning of 12 July 1995. The ozone profiles are each spaced about 50 min apart, starting at 7:30 CDT and show the vertical redistribution of ozone as the mixed layer grows.

• The vertical distribution of pollutants at night. Due to a lack of vertical mixing (see

Section 2.8.2) pollutants tended to form horizontal layers or patches that persisted throughout the night until they were mixed out by the growing boundary layer the next morning. Fig. 2.8.8 depicts the cross section of a power plant plume that had been injected into the stabilizing evening atmosphere. In the absence of any significant vertical mixing the power plant plume stayed confined to a thin vertical layer.

Figure 2.8.8. Cross section of Cumberland power plant plume measured with the NOAA/ETL airborne ozone lidar on the evening of 4 July 1999 at about 10 km downwind of the power plant. The plume is easily identified by its low-ozone signature. Note that the plume is confined to the layer between 800 and 1200 m ASL.

30 June 2004 | State of SOS-3: 1995 - 2003 | 115

520 560 600 640 680

Solar Irradiance (W m-2)

0

0.2

0.4

0.6

0.8

1

Aero

so

l O

pti

ca

l D

ep

th

Durham, NH (41.0o zenith angle) August 11-16, 2002

Observed Irradiance

ETA Forecast Irradiance

Average Bias ~ 84.4 W m-2

Y = -0.019 * X + 15.7R-squared = 0.52

Y = -0.0089 * X + 6.8R-squared = 0.99

• Pollutant concentrations in the free troposphere are affected by long-range transport or stratosphere-troposphere exchange processes. Ozone sonde and aircraft measurements from the Nashville ‘95 and ‘99 campaigns showed that concentrations of ozone and other pollutants in the free troposphere were highly variable and were primarily affected by regional to continental-scale advection of clean or polluted air masses. Another significant process contributing to high tropospheric ozone concentrations is the intrusion of stratospheric air. Through entrainment processes (see Section 2.8.1) pollutant concentrations in the lower free troposphere can impact the air quality in the atmospheric boundary layer and at the surface.

• The improper treatment of aerosols in models contributed to forecast errors in solar radiation reaching the surface. Aerosol absorption and scattering reduce the amount of sunlight that reaches the Earth’s surface. During the 2002 New England Air Quality Study, the correlation between observed aerosol optical depth and incoming solar radiation measured at Durham, NH was greater than the correlation between observed aerosol optical depth and predicted incoming solar radiation from the ETA model (see Fig. 2.8.9). The model also was unable to produce the observed slope for the line of regression between these two variables. This improper treatment of aerosols in the ETA model contributed to a large radiation bias (see also Fig. 2.8.6).

Figure 2.8.9. Correlation between observed solar irradiance and aerosol optical depth measured at Thompson Farm in Durham, NH during the 2002 New England Air Quality Study along with correlation between the observed optical depth and solar irradiance predicted by the ETA model.

KEY CITATIONS: Banta, R.M., C.J. Senff, A.B. White, M. Trainer, R.T. McNider, R.J. Valente, S.D. Mayor, R.J. Alvarez, R.M.


Senff, C.J., R.M. Hardesty, R.J. Alvarez II, and S.D. Mayor. 1998. Airborne lidar characterization of power plant plumes during the 1995 Southern Oxidants Study. J. Geophys. Res. 103(D23):31,173-31,189.

Wotawa, G. and M. Trainer. 2000. The influence of Canadian forest fires on pollutant concentrations in the United States. Science 288:324-328.

Zamora, R.J., E.G. Dutton, M. Trainer, S.A. McKeen, J.M. Wilczak, and Y.-T. Hou. 2004. The accuracy of solar irradiance calculations used in medium range forecast models. Mon. Wea. Rev. (accepted for publication).

30 June 2004 | State of SOS-3: 1995 - 2003 | 116

2.8.5. Off- to On-Shore Flow Reversal In coastal areas, the land - sea breeze circulation can act to focus large concentrations of

ozone and other pollutants. Very high concentrations of pollutants can be expected when the

morning offshore flow is followed by a period of stagnant winds and the sea breeze recirculates

the aged pollutants released in the morning back over the sources areas. Key features are

summarized below.

• Off- to on-shore flow reversal was observed in Houston during TexAQS 2000 in conjunction with very large accumulations of ozone. Severe ozone exceedances on flow-reversal days were linked to a combination of two meteorological factors: (1) Light-wind conditions that allow buildup of ozone plumes over source areas during the middle of the day, and (2) Afternoon sea breeze that transports aged pollutants back over source areas, thus reinforcing the already high ozone concentrations (see Fig. 2.8.10). The distribution of pollutants and the severity of the ozone event depend on the morning offshore flow regime, the timing of the sea breeze onset, and the strength of the sea breeze.

Figure 2.8.10. Cross section of ozone measured with the NOAA/ETL airborne ozone lidar during TexAQS 2000 in the late afternoon on 30 August. Very high ozone concentrations approaching 200 ppbv extended to about 2 km above sea level in the sea breeze convergence zone along the western shore of Galveston Bay.

30 June 2004 | State of SOS-3: 1995 - 2003 | 117

• Other occurrences of high ozone are associated with a coupling of the sea breeze and the inertial oscillation. The two phenomena are nearly in resonance at the latitude of Houston, and they produce a few hours of nearly calm winds during late morning or early afternoon when large-scale winds are light from the south or southeast (see Fig. 2.8.11). Large-scale mean winds must be lighter than a threshold value of about 3 m/s for flow reversal to occur. This reversal takes place almost simultaneously throughout the metropolitan area, not only in association with the sea breeze front. When winds are sufficiently light, the likelihood of an ozone exceedance is greater than 50%. Exceedances are also relatively likely when winds are stronger from the northeast.

Figure 2.8.11. Hodograph of wind as a function of time of day, 247 m level, Liberty wind profiler, averaged over August 16-30, 2000. The wind undergoes a steady clockwise rotation throughout the 24-hour cycle, with the strongest onshore winds taking place at night.

KEY CITATIONS: Senff, C.J., R.M. Banta, L.S. Darby, R.J. Alvarez II, S.P. Sandberg, R.M. Hardesty, and W.M. Angevine.

2002. Ozone distribution and transport in the Houston area: Insights gained by airborne lidar. Fall Meeting of the American Geophysical Union, 6 – 10 December 2002.

Nielsen-Gammon, J. W. 2001. Initial Modeling of the August 2000 Houston-Galveston Ozone Episode. 71 pp. http://www.met.tamu.edu/temp/dec19.pdf

2.8.6. Heterogeneity of Industrial Plumes Measurements, made by the Baylor aircraft downwind of industrial sources in the fall of

2001, suggest that while some industrial plumes are well mixed, other plumes are spatially

heterogeneous. The spatially heterogeneous plumes can contain regions with high

concentrations of VOC, regions with high concentrations of NOx and regions with high

concentrations of both VOC and NOx. Whether a plume is well mixed or heterogeneous is likely

to depend on the distance from the source and atmospheric stability conditions.

KEY CITATION: Daum, P.H., J. Meagher, D. Allen, and C. Durrenberger. 2002. Accelerated Science Evaluation of Ozone Formation in

the Houston-Galveston Area. Summary. 6 pp. http://www.utexas.edu/research/ceer/texaqsarchive/accelerated.htm.

-2

-1

0

1

2

3

4

5

6

7

-2 -1 0 1 2 3 4 5 6 7

Alongshore Wind (m/s)

On

sh

ore

Win

d (

m/s

)

Noon Dawn

wn

Midnight

Sunset

Dawn Dawn

wn

Afternoon

Morning Dawn

wn

30 June 2004 | State of SOS-3: 1995 - 2003 | 118

2.9. Observation-Based Analysis O3 production rate and

sensitivity Effects of emission controls:

NOx vs VOC sensitivity Process-based evaluation of

emission-based models Comparison of observations

with inventories and modeled ozone

2.9. OBSERVATION-BASED ANALYSIS Larry Kleinman, Carlos Cardelino, Sandy Sillman, Gail Tonnesen

A challenge that the atmospheric sciences

community faces is to translate the measurements made

during field programs into knowledge about the

processes responsible for elevated pollutant

concentrations and the efficiency of possible control

measures. The traditional way of proceeding is to use

field observations to validate emissions-based models.

A complementary approach, which has been pioneered

by the SOS community, is to use Observation-Based

Analysis—a family of techniques that attempts to relate

questions about processes and emission controls directly to measured concentrations. The

promise of observation-based methods is that by bringing in the actual atmospheric

concentrations, there is less reliance on ill-characterized emissions, computational demands are

lessened, and an element of reality is imposed on the problem.

Measurements are intrinsically local in that they provide information about the state of an air

mass at the time and place that the observations occur. From local measurements we hope to do

two things: 1) Deduce rates and sensitivities that cannot be measured, for example the production

rate of O3 and its sensitivity to NOx and VOCs; and 2) Deduce the effects of emission controls,

which depend on the entire time history of the air mass.

Observation-based analysis establishes a link between processes and observations. This

linkage is one that an emissions-based model must reproduce to give credible predictions on

emission controls. Observation-based techniques thereby suggest criteria for evaluating model

performance that go beyond the usual matching of ozone concentration fields.

30 June 2004 | State of SOS-3: 1995 - 2003 | 119

2.9.1 O3 Production Rate and Sensitivity The comprehensive sets of

measurements obtained at surface sites and

from aircraft allow one to calculate the local

rate at which O3 is produced and the

dependence of this rate on precursors.

Analytic formulas provide a framework for

generalization and a theoretical rationale for

Indicator Ratios (see Section 2.9.2).

1. Ozone Production Rate, P(O3). Consistency between different methods of determining P(O3) is an important test of theoretical understanding and measurement procedures.

2. Ozone Production Rate – NOx and VOC Sensitivity. The SOS community has developed two complementary ways of analyzing the local rate and sensitivity of O3 production: by means of a radical budget and by means of radical propagation efficiency. These methods yield insights and useful formulas. Generalizations provide the basis for Indicator Ratio methods.

• Radical Budget. Photochemistry under NOx-sensitive conditions preferentially forms peroxides; under VOC-sensitive conditions NOz compounds are preferentially formed. The sensitivity of P(O3) to NO and VOCs is given by a simple analytic function of "LN/Q,” the fraction of free radicals removed by reacting with NOx.

Figure 2.9.1. Diurnal profile for P(O3) calculated from data collected during the ROSE campaign at a rural site in Alabama in June-July, 1990. P(O3) calculated using a steady state model (SSM), photostationary state (PSS), radical budget (RB), and from peroxy radical concentrations measured with a chemical amplifier (CA). Vertical bars indicate variability of average hourly values. (Frost et al., 1998).

Ln/Q, fraction radicals reacting with NOx

0.0 0.2 0.4 0.6 0.8 1.0

dln

P(O

3

)/dln

[X]

-1.0

-0.5

0.0

0.5

1.0

d[Peroxide]/dt> 1/2 d[NO

z]/dt

1/2 d[NOz ]/dt >

d[Peroxide]/dt

NOx sensitive HC sensitive

dlnP(O3)/dln[NO]

dlnP(O3)/dln[HC]

Figure 2.9.2. The relative sensitivity of O3 production rate to [NO] and [VOC] as a function of the fraction of radicals reacting with NOx. Solid lines are analytic formulas. Data points are from constrained steady state box model calculations driven from observations taken during the 1995 Nashville campaign. (Kleinman et al., 1997).

10

8

6

4

2

0

P(O

3),

ppbv h

r-1

181614121086

Hour of Day (CST)

PSS CA RB SSM

G. J. Frost et al., J. Geophys. Res., 103, 22491-22508, 1998, Figure 4b

30 June 2004 | State of SOS-3: 1995 - 2003 | 120

• Radical Propagation Efficiency. O3 is formed in a chain reaction. A simple version is: OH + VOC → HO2 HO2 + NO → OH + NO2

NO2 + h< + O2 → O3 + NO Numerical calculations show that the O3 yield is maximized when the chain length is long. Radical loss processes limit the chain length. At low NOx, HO2 radicals combine to form peroxides; at high NOx, OH reacts with NO2 forming HNO3. The combination of these two loss reactions causes a maximum in P(O3) at a particular NOx concentration. The ratio of the peroxide to HNO3 production rate tells us whether the atmosphere is on the low or high NOx side of the maximum.

Figure 2.9.3a and b. A trajectory model, simulating Atlanta, GA was used to calculate O3 (left panel) and OH radical chain length (right panel) as a function of initial NOx and VOC concentration. Emissions were proportional to initial concentrations. Note that the O3 ridgeline (bold diagonal line) is also a maximum in OH chain length. (Tonnesen and Dennis, 2000).

KEY CITATIONS: Frost, G.J., M. Trainer, G. Allwine, M.P. Buhr, J.G. Calvert, C.A. Cantrell, F.C. Fehsenfeld, P.D. Goldan, J.

Herwehe, G. Hübler, W.C. Kuster, R. Martin, R.T. McMillen, S.A. Montzka, R.B. Norton, D.D. Parrish, B.A. Ridley, R.E. Shetter, J.G. Walega, B.A. Watkins, H.H. Westberg, and E.J. Williams. 1998. Photochemical ozone production in the rural southeastern United States during the 1990 Rural Oxidants in the Southern Environment (ROSE) program. J. Geophys. Res. 103D:22,491-22,508.

Kleinman, L.I., P.H. Daum, J.H. Lee, Y.-N. Lee, L.J. Nunnermacker, S.R. Springston, L. Newman, J. Weinstein-Lloyd, and S. Sillman. 1997. Dependence of ozone production on NO and hydrocarbons in the troposphere. Geophys. Res. Lett. 24(18):2299-2302.

Kleinman, L.I. 2000. Ozone process insights from field experiments – part II: Observation-based analysis for ozone production. Atmos. Environ. 34:2023-2033.

Tonnesen, G.S. and R.L. Dennis. 2000. Analysis of radical propagation efficiency to assess ozone sensitivity to hydrocarbons and NOx .1. Local indicators of instantaneous odd oxygen production sensitivity. J. Geophys. Res. 105D:9213-9225.

30 June 2004 | State of SOS-3: 1995 - 2003 | 121

2.9.2 Effects of Emission Controls: NOx vs. VOC sensitivity 1. Indicator Ratios. The local analysis

described in 2.9.1 shows that under NOx-sensitive conditions, peroxides are preferentially formed as a by-product of O3 production; under VOC-sensitive conditions NOz is preferentially formed. Observed concentration ratios containing these compounds thereby contain information on whether O3 was formed under NOx- or VOC-sensitive conditions. Ratios among these species have been referred to as indicator ratios, and their values contain information about the upwind conditions under which ozone has been formed.

Specific values of indicator ratios have been identified as corresponding to either NOx-sensitive or VOC-sensitive conditions. This identification of NOx-sensitive or VOC-sensitive indicator values is based on predictions of emission-based models, and is subject to the uncertainties associated with these models. However, the indicator ratios provide a way of evaluating the accuracy of NOx-VOC predictions from emission-based models, and for comparing and contrasting measurements in different locations.

Measured indicator ratios in the Nashville urban plume (O3/NOz and H2O2/NOz) are close to the values associated with the transition between NOx- and VOC-sensitive chemistry. In contrast, indicator ratios suggest that Atlanta is NOx-sensitive and Los Angeles is VOC-sensitive. In Nashville, the optimum control strategy will involve reductions in both categories of precursors.

Figures 2.9.4a. and b. The correlation between O3 and summed NOx reaction products (NOy-NOx, also referred to as NOz) from emission-based models and from measurements. Model results (a) are coded based on the response of the model to reductions in NOx and VOC, as either NOx-sensitive (green circles), VOC-sensitive (X’s), mixed (pink dashes), or with a pattern dominated by NOx titration (blue asterisks). The equivalent measurements (b) are from sites in Atlanta (green circles), Nashville (pink dashes) and Los Angeles (X’s). Source: Sillman and He (2002) and measurements reported in Sillman et al. (1997, 1998).

30 June 2004 | State of SOS-3: 1995 - 2003 | 122

2. The Observation-based Model (OBM). The Observation-based Model provides an independent check on the precursor relationships predicted by emission-based air quality models. The OBM utilizes atmospheric observations (instead of emission inventories) to drive a photochemical model and infer sensitivities of ozone in an urban atmosphere to changes in the emissions of VOC, NOx, or CO. The OBM was used to analyze data gathered at three Photochemical Assessment Monitoring sites (PAMS): Washington, DC, Bronx, NY, and Houston, TX. Results of the analysis are summarized below. • Natural hydrocarbons, primarily

isoprene, represented a significant fraction of the total hydrocarbon reactivity and significantly degraded the efficacy of VOC emission reductions as an ozone mitigation strategy.

• Afternoon NO concentrations typically fell to levels at or below the limit of detection of the PAMS instrumentation and, as a result, it is not possible to determine whether ozone was more sensitive to reductions in anthropogenic hydrocarbons or nitrogen oxides.

Figure 2.9.5. Variation in OBM-calculated relative incremental reactivities (RIRs) for ntiric oxide (NO), anthropogenic hydrocarbons (AHC), natural hydrocarbons (NHC), and CO at the Bronx, New York site as a function of the assumed minimum afternoon NO concentration (from Cardelino and Chameides, 2000).

KEY CITATIONS: Sillman, S. and D. He. 2002. Some theoretical results concerning O3- NOx -VOC chemistry and NOx -VOC

indicators. J. Geophys. Res. 107(D22), 4659, doi:10.1029/2001JD001123. Sillman, S., D. He, M.R. Pippin, P.H. Daum, D.G. Imre, L.I. Kleinman, J.H. Lee, and J. Weinstein-Lloyd.

1998. Model correlations for ozone, reactive nitrogen, and peroxides for Nashville in comparison with measurements: Implications for O3-NOx-hydrocarbon chemistry. J. Geophys. Res, 103D:22,629-22,644.

Cardelino, C.A. and W.L. Chameides. 2000. The application of data from photochemical assessment monitoring stations to the observation-based model. Atmos. Environ. 34: 2325-2332.

30 June 2004 | State of SOS-3: 1995 - 2003 | 123

2.9.3 Process-Based Evaluation of Emission Based Models An evaluation of models used to establish ozone control policy must include a comparison

between observed and predicted O3 concentration fields. But models can have O3 predictions

that are substantially identical yet predict different sensitivities to NOx and VOCs. Sillman et al.

(1997) have shown that alternate model predictions on O3 sensitivity can be distinguished one

from the other according to their predictions of indicator ratios. The premise developed within

the SOS community is that there is a robust relation between indicator ratios and O3 sensitivity

so that predicting these ratios correctly is strong evidence that a model can accurately predict the

effects of emissions controls. Because models can give the right O3 for wrong reason, process-

oriented tests of model performance provide a more robust evaluation of model performance.

In a process-based evaluation:

• Models should be evaluated by how well they perform for specific ratios (O3/NOz, H2O2/HNO3) that are closely linked with NOx or VOC sensitive chemistry

• The evaluation should be based on measurements that are concurrent with peak O3 in the event.

• An evaluation should be done for a series of plausible model scenarios that give different results for O3 – NOx – VOC sensitivity.

Figure 2.9.6. Peak O3 and concurrent NOy in the Atlanta urban plume on August 10, 1992. The X's represent helicopter measurements at 600m elevation, 4:00-5:00 pm, located within 4 km of the location of measured peak O3. Bold letters represent domain-wide peak O3 and concurrent NOy at 600m elevation, 4:00-5:00 pm for different model scenarios. The line represents the transition between NOx and VOC sensitive chemistry based on O3/NOy. Note that only 2 model scenarios are consistent with both O3 and NOy observations. Both scenarios are NOx-sensitive (from Sillman et al., 1997).

KEY CITATION: Sillman, S., D. He, C. Cardelino and R.E. Imhoff. 1997. The use of photochemical indicators to evaluate ozone

NOx-hydrocarbon sensitivity: Case studies from Atlanta, New York and Los Angeles. J. Air Waste Manage. Assoc. 47(10):1030-1040.

30 June 2004 | State of SOS-3: 1995 - 2003 | 124

2.9.4 Comparison of Observations with Inventories and Modeled Ozone Comparison of ambient measurements of emissions made during routine monitoring or in

research studies provide a means to verify bottom-up emissions inventories and, along with

measurements of ozone, to assess photochemical model performance. Work in TexAQS 2000

showed that photochemical model simulations are very sensitive to the choice of VOC emission

rates used for the petrochemical facilities in the Houston area. Key finding is summarized

below.

• Use of emissions estimated based on ambient observations compared to inventory emissions resulted in model results for ozone concentrations that agreed better with measurements from the NCAR Electra and from the surface regulatory network.

KEY CITATION: NOAA Aeronomy Laboratory. 2003. Texas 2000 Air Quality Study - Phase II Analysis of NOAA Data. Final

Report to Texas Commission on Environmental Quality Houston/Galveston Air Quality Science Evaluation. ftp://ftp.tnrcc.state.tx.us/pub/OEPAA/TAD/Modeling/HGAQSE/Contract_Reports/Data_Analysis/TexAQS2000_NOAA_Data_Analysis.pdf. 158 pp.

30 June 2004 | State of SOS-3: 1995 - 2003 | 125

2.10. Emissions Based Modeling Strengths and limitations

Photochemical mechanisms Role of temperature in ozone

production Regional surface ozone

predictability Meteorological predictability

Control strategy sensitivity Performance evaluation using

chemical signals Role of mixing processes

.Implementation of DDM-3D in URM and CMAQ

Development of Reactivity Metrics for VOC

Assimilation of Satellite Data into MM5

Chemical Indicators

Analysis of Winds Observed by Radar Profilers

Web-Based Lagrangian Particle Model

UAM in Forecast Mode

2.10. EMISSIONS-BASED MODELING Richard McNider and Ted Russell

Emissions-based modeling has always been at the

core of the SOS research and assessment program.

While the emphasis within SOS clearly has been on

observations, it was always intended that emissions-

based models would serve as a synthesis point for

observations and as a means to test model performance

and fidelity. Because of the delay in actually getting

observational and supporting auxiliary data sets in

place, many of the modeling efforts have involved

sensitivity studies and in improving model components

or processes.

2.10.1 Strengths and Limitations 2.10.1.1. SOS’ Emissions-Based Modeling

Activities During the last 7-10 years, the SOS program has

focused so thoroughly on field observation activities,

that the financial support for emissions-based modeling

from the core budget of SOS was relatively small. In

2001-2003, however, modest funding for model

development and model-based analysis of SOS’

extensive field measurements was made available from

SOS’ core funds. At the same time, the SOS Task

Group on Models and Model Evaluation (SOS-MME) facilitated and nurtured a substantial

number of emissions-based modeling activities that were funded outside the core EPA budget by

other agencies, EPA grant programs, or through in-house activities by SOS cooperators within

EPA and NOAA. SOS-MME fostered these activities through workshops, observational data

support, and intellectual contributions. Some of these separately funded and collaborative

activities are discussed below in the larger context of SOS contributions.

30 June 2004 | State of SOS-3: 1995 - 2003 | 126

1. The Southeastern States Air Resources Managers (SESARM) Seasonal Modeling of Regional Air Quality (SMRAQ). The Microelectronics Center of North Carolina (MCNC), Duke University, and Georgia Tech developed a very innovative program to test the fidelity of emissions based models in a seasonal context rather than an episodic mode. This program was fostered by SOS and funded by SESARM with some contribution from utilities. SOS investigators played a major role in the design and review of this modeling effort. SOS’ observations during the 1995 Nashville/Middle Tennessee Ozone Study were a key conceptual opportunity for model evaluation. Some of the model frameworks developed by this study were used in real time forecasting of ozone concentrations for TexAQS 2000. 2. Southern Appalachian Mountain Initiative (SAMI). Georgia Tech, TVA, UAH, Alpine Geophysics, and other groups are carrying out a comprehensive study of air quality in the Southern Appalachians. This program was funded by SESARM, the U.S. EPA, and direct Congressional appropriations. The core understanding of Southeast modeling issues and supporting data were key to the conceptual design and evaluation of this activity. Several SOS investigators contributed to this effort. 3. EPA Competitive Grants Programs. Several SOS investigators have been successful in winning emissions-based modeling grants from EPA’s STAR Grants Program. The SOS connection almost certainly helped during the policy-relevance review portions of these competitive processes. Specifically, UAH won grants for modeling activities related to satellite assimilation, plume-in-grid modeling, and coupled chemical/large eddy simulations. University of Michigan won grants related to developing indicator species. Georgia Tech won grants dealing with inverse modeling of emissions, the Southeast Center for Integrated Study of Secondary Air Pollutants, and, more recently, on organic carbon source apportionment, and the impact of global change on air quality. UAH also won STAR grant support related to satellite assimilation and coupled chemical large eddy simulation modeling. Observations from TexAQS 2000 were used as tests for both of these UAH STAR grant activities. UAH and Georgia Tech continue to work together on modeling-related studies. 4. NOAA and EPA In-House Modeling Programs. Both the NOAA Aeronomy Laboratory and EPA’s National Exposure Research Laboratory have conducted large-domain regional simulations and also small domain plume studies. These studies have been spurred in part by the SOS observational program, and SOS scientists have been involved in the planning and evaluation of these modeling activities.

5. Fall Line Air Quality Study. The state of Georgia is funding a combined monitoring and modeling study of ozone and PM formation, transport, and control in Georgia. This work builds on the modeling studies conducted for SAMI, SOS, and SCISSAP, with a focus on identifying strategies that would be most effective at improving air quality and meeting air quality goals in Georgia.

6. Partnership for Environmental Research and Community Health (PERCH). Georgia Tech is participating in an environmental study investigating environmental and public health related issues in the Florida panhandle. PERCH is using CMAQ results to understand the sources of various air pollutants in this most southerly part of the SOS region.

7. TexAQS 2000 Satellite Assimilation Activities. These modeling activities were supported by funding from the Texas Commission on Environmental Quality (TCEQ) and NOAA Coastal

30 June 2004 | State of SOS-3: 1995 - 2003 | 127

Impact funds. They used the CMAQ/MM5 modeling infrastructure developed under SOS. Preliminary results from this activity show that model winds and planetary boundary layer height estimates were improved by assimilating satellite data. It was also recommended that this MM5 run be used as the State of Texas’ State Implementation Plan MM5 model atmosphere.

8. TexAQS 2000 Chemical Large Eddy Simulation. Short-term spikes in ozone concentration have been a major concern in understanding rapid ozone formation events in Houston. Studies of these ozone spikes used the large eddy model developed in part by SOS core funding to the NOAA Turbulence and Diffusion Laboratory. The LES–Chem model also was used to evaluate fluctuations in point-source emissions and the role of correlation among various VOC chemical species in accelerating plume chemistry.

2.10.1.2. SOS Community Modeling Activities While the SOS program early on saw the long-range value of coupling SOS’ robust chemical

and meteorological observations with a strong emissions-based modeling effort, it was clear that

because of program priorities and the distributed nature of SOS that a new paradigm was needed

for the SOS modeling and model evaluation (SOS-MME) program to be viable.

In conjunction with EPA’s NERL Modeling Branch, UAH coordinated a workshop in April

1999 to explore and define the structure for an enhanced SOS Modeling Program. At the

workshop, overviews of SOS modeling activities were given by a variety of SOS investigators.

A broad discussion was held concerning the future and structure of a successful SOS Modeling

Program. From this discussion, the concept of distributed SOS Community Modeling built

around the use of Models3 was formulated. This structure was based on the premise that no one

organization had all the skills, talents, and the human-resource infrastructure necessary to carry

out all activities in a state-of-the-art emissions-based modeling program striving to make

significant forward progress in air-quality modeling. If each university (or government-agency

partner such as TVA) could contribute in the areas in which they had special competence,

however, then the combined skills and talents were there. For example, UAH could support the

meteorological modeling component, while Georgia Tech could contribute in the area of

emissions modeling, the University of Texas in the area of gas and condensed phase chemistry,

and the University of Michigan in the area of chemical mechanisms.

Thus the SOS Community Modeling Framework was begun. UAH is taking the lead in

coordination and putting the Models3 framework in place. Meetings were held with Georgia

Tech and TVA modelers to plan the emissions development process. While the heavy field

program load in past years limited SOS core funding, the SOS Community Modeling Program

30 June 2004 | State of SOS-3: 1995 - 2003 | 128

provided the proper framework for SOS investigators to make substantial contributions in the

Nashville, Atlanta, and Houston/Galveston field program settings. It has also led to significant

advances in the more general CMAQ concept, and its application by various SOS and SOS-

related investigations both in the SOS region and in many other parts of the US.

UAH and Georgia Tech teamed up to conduct detailed modeling of the Altanta Supersite

period. Georgia Tech and the University of Texas conducted detailed modeling of the Houston

Supersite observations.

As the SOS Community Modeling Program has evolved, the system as changed somewhat in

organization and focus. For example, during TexAQS 2000, UAH used the common

CMAQ/MM5 system as a foundation for joint activities with Texas A&M University, Lamar

University, the University of Houston, and the air-quality modelers in TCEQ. These activities

included data assimilation studies, development of point-source plume-trajectory models, and

other model evaluation activities.

30 June 2004 | State of SOS-3: 1995 - 2003 | 129

2.10.2 SOS Photochemical Mechanism Studies The Carbon Bond IV Mechanism (CB4) has been the primary photochemical mechanism

used in regulatory settings for testing and setting strategies by local, state, and federal regulators.

Recent research suggested a possible change in the CB4 mechanism related to organic peroxide

radicals that affect organic nitrate formation. This proposed change decreases the model-

estimated rate of formation of organic nitrates in the atmosphere, and thus increase the amount of

NOy that would remain active in ozone formation. Previous tests of this proposed CB4

modification in the northeastern region of the US indicated that the proposed modification

change would lead to only modest changes in estimated ground-level ozone concentrations.

Thus, SESARM sponsored further modeling studies to determine if the proposed peroxide-

radical modification of the CB4 mechanism would have a similarly small effect on estimated

ground-level ozone concentrations in the southeastern region. In contrast to the relatively small

model-estimated change in ground-level ozone observed in the Northeast, this SESARM-

sponsored research demonstrated that simulated ground-level ozone concentrations in regions of

relatively smaller emissions of NOx, such as the southeastern US, would be extremely sensitive

to this organic-peroxide modification of the CB4 mechanisms and that estimated ground-level

ozone concentrations could be altered by as much as 10-25 ppb. The SESARM-sponsored

researchers also stressed that the sensitivity of ozone to this proposed mechanism change

increased the importance of understanding the air-borne nitrate budget in the Southeast

(Kasibhatla et al., 1997).

In a similar vein, Biazar (1996) showed another example of the extreme sensitivity of how

organic nitrates are handled in air quality modeling. He studied the effect of including

atmospheric deposition losses of MPAN in the models. Biazar discovered that once organic

nitrates are formed, the estimated rate of removal of reactive nitrogen in the system through

surface deposition losses was very similar to the estimated rate of thermal decomposition of

reactive nitrogen for subsequent ozone formation. With the inclusion of deposition losses of

MPAN in model runs for a rural environment, the model indicates that reactive nitrogen is

rapidly removed from the system, leading to much smaller ground-level concentrations of ozone.

Somewhat similar uncertainties regarding the adequacy of the CH4 mechanism and the

possible importance of chlorine chemistry arose during research seeking to understand the rapid

ozone formation events observed in Houston, TX and especially in the Ship Channel area of this

30 June 2004 | State of SOS-3: 1995 - 2003 | 130

city. Thus, both modeling studies and combinations of chlorine-focused chamber studies,

modeling studies, and real-world field observations during rapid ozone formation events were

undertaken in connection with the TexAQS 2000 field study.

Key findings and policy implications from these studies included the following:

• Photochemical models using the common Carbon Bond 4 chemical mechanism gave reasonably good estimates of ozone concentrations during rapid ozone formation episodes in Houston.

• These same models and mechanisms tended to overpredict NOz concentrations (and especially HNO3 concentrations) during rapid ozone formation episodes in Houston {and especially in the Ship Channel area of Houston?}.

• These reasonably good estimates of ozone concentrations and underpredictions of NOz concentrations (and especially HNO3 concentrations) also were observed in recent applications of the Comprehensive Air Quality Model with Extensions (CAMx) photochemical model

These three findings are very important in developing confidence in the air quality models

used in assessing the relative efficacy and cost-effectiveness of NOx and VOC control strategies

because they suggest different Ozone Production Efficiencies than were observed in specific

urban and regional ozone non-attainment areas, and different ozone-indicator species

relationships.

• Chemical reactions with chlorine can increase ozone in Houston. Chemical reactions involving chlorine were incorporated into the Comprehensive Air Quality Model with Extensions (CAMx) photochemical model. An inventory of chlorine sources in the Houston urban area also was developed. Results from the chorine-included CAMx model indicated that estimated ozone concentrations were increased by 5-15 ppb compared to model results without chlorine chemistry. This was true in both the Ship Channel area and in other areas of Houston (Allen, 2003).

Anthropogenic emissions of chlorine should be considered in developing air quality plans for

the Houston-Galveston region of Texas.

• Both temporal and spatial fluctuations in air turbulence and distances between major NO and VOC sources can change the effective rates of ozone formation reactions. Comparison of modeling results using both coarse-grid and fine-grid versions of the LES-Chem model show that: a) if NO and reactive VOC plumes are mixed by turbulent meteorological conditions soon after emission, rapid ozone formation will occur immediately and in close proximity to the NO and VOC emission sources; and conversely b) if NO and reactive VOC plumes occur under low-turbulent conditions, convergence of the two plumes can be delayed, and rapid ozone formation will occur at longer distances from emissions sources.

30 June 2004 | State of SOS-3: 1995 - 2003 | 131

KEY CITATIONS: Allen, D. 2003. Southern Oxidants Study Research Program at the University of Texas Final Report. (EPA Contract

No. R-82902801-0 and EPA Contract No. R-82902801-1 continuation). 19 pp. Kasibhatla, P., W. Chameides, B. Duncan, M. Houyoux, C. Jang, R. Mathur, T. Odman, and A. Xiu. 1997.

Impact of inert organic nitrate formation on ground level ozone in a regional air quality simulation using the Carbon Bond IV Mechanism. Geophys. Res. Letters, Vol. 24, No. 24 pp 3205-3208

Biazar, A.P. 1995: The role of natural nitrogen oxides in ozone production in the Southeast environment. Ph.D Dissertation, University of Alabama in Huntsville.

Herwehe, J., R. McNider, and A. Biazar. 2002. LES modeling of industrial sources. AMS Turbulence and Diffusion Conference, Wageningen, Netherlands.

30 June 2004 | State of SOS-3: 1995 - 2003 | 132

2.10.3 Regional Surface Ozone Predictability In order to determine model performance outside of the traditional episode mode, the results

and performance of a seasonal photochemical model for the eastern U.S. was carried out for the

summer of 1995. The model was the MAQSIP Model driven by the MM5 meteorological model

in a data assimilation mode with the MM5 reinitialized every five days. Primary emphasis was

placed on model performance in rural areas.

In an analysis of the seasonal model results, a comparison of daytime average ozone

produced by the model with observed values, showed virtually no bias and a relatively strong

correlation at both low and high concentrations (Figure

2.10.1). On the other hand, when model performance on a

day-to-day basis was examined, the results showed that on

many days the model showed very little skill, especially

on some of the higher average ozone days within the

modeling domain. Figure 2.10.1. Scatter plots of modeled versus observed 10am-5pm average surface ozone concentrations at various percentiles (blue: 10th percentile; magenta:25th percentile; green 50th percentile; yellow: 75th percentile; red 90th percentile. The dashed line is the one-to-one line (from Kasibhatla and Chameides, 2000).

In addition to this study, SMRAQ investigators have teamed with investigators in the

northeast to co-evaluate modeling for the eastern U.S. by the NARSTO NE modeling team for

the month of July 1995, which overlapped the SMRAQ study period. This study concluded that

predictions of ozone are insignificant on the intra-day time scale, high for the diurnal component,

low for the amplitude of the diurnal component and highest yet for the synoptic scale. The

modeling results show that:

• Higher-frequency (several-day or less) mesoscale chacteristics are not well replicated in the seasonal photochemical models. Thus it was suggested that for regulatory purposes, longer-term simulations may be preferable to episode simulations.

These tools and approaches are continuing to be used in analysis and model verification for

Nashville 1999 and TexAQS 2000.

KEY CITATIONS: Kasibhatla, P. and W. Chameides. 2000. Seasonal modeling of regional ozone pollution in the eastern United

States. Geophys. Res. Letters Vol. 27, No. 9: 1415-1418. Hogrefe, C., S.T. Rao, P. Kasibhatla, W.Hao, G. Sistla, R. Mathur, and J. McHenry. 2001a. Evaluating the

performance of regional-scale photochemical modeling systems: Part II ozone predictions. Atmos. Environ. 35(24):4175-4188.

30 June 2004 | State of SOS-3: 1995 - 2003 | 133

2.10.4 Meteorological Predictability Meteorological models are used to describe the physical atmosphere, to carry out

photochemical simulations, and to consider alternative control strategies. The characteristics of

the meteorological model used in large part controls model performance – especially for paired-

in-space-and-time statistics. Joint studies of meteorological model performance for the SMRAQ

1995 seasonal modeling and Northeast NARSTO July 1995 modeling were carried out looking at

meteorological model performance. SMRAQ used the MM5 model, and the NARSTO NE study

used the RAMS model. These studies showed that, in both models:

• Higher frequency intra-day fluctuations were not well correlated with observations. Correlations were highest for the longer time scale (synoptic) components.

In a separate study of the temporal spectral characteristics of profiler data taken during

Nashville ’95, it was shown that models underestimate the magnitude of the diurnal/oscillations

compared to observations. This study also argued that there is an inherent spectral gap in models

where the spectral energy at spatial scales less than several model grid scales and above the PBL

scale is not included in the simulations. This conclusion is evidently in keeping with the MM5

and RAMS results.

KEY CITATIONS: Hogrefe, C., S.T. Rao, P. Kasibhatla, G. Kallos, C. Tremback, W.Hao, G. D. Olerud, A.Xiu, J. McHenry, and

K. Alapaty. 2001b. Evaluating the performance of regional-scale photochemical modeling systems: Part I - meteorological predictions. Atmos. Environ. 35(24): 4159-4174.

Gupta, S., R. McNider, M. Trainer, R. Zamora, K. Knupp, and M. Singh. 1997. Nocturnal wind structure and plume growth rates due to inertial oscillations. J. Appl. Met. 36(8):1050-1063.

30 June 2004 | State of SOS-3: 1995 - 2003 | 134

2.10.5 Control Strategy Sensitivity In this study, the Regional Oxidant Model (ROM) was used to carry out a study regarding the

relative sensitivity of changes in ambient ozone concentrations to decreases in emissions of VOC

and NOx. A simulation period July 2-10, 1988 was used. A very deliberate and methodical set

of decreases in NOx and VOC emissions in 25% increments were carried out – both separately

and concurrently. The initial state was based on 1988 emission estimates. A comparison of

modeled versus observed ozone patterns showed relatively good skill, but the point-to-point

comparisons were less favorable, showing large scatter. ROM underpredicted daily maximum

ozone concentrations throughout most of the modeling domain. There was some tendency for

bias toward underprediction of ozone in the Southeast. This bias was absent in the Northeast, but

scatter was greater.

This sensitivity study showed that decreases in NOx emissions led to significant decreases in

maximum ozone concentrations over large areas. Decreases in VOC emissions did not lead to

decreases in maximum ozone concentrations near major urban areas. However, decreasing NOx

emissions by 50% in general had a larger impact than a 50% decrease in VOC emissions. These

results indicate that much of the eastern U.S. is NOx-limited. While some of the northern

subregions (Northeast Corridor and Lake Michigan) were VOC-limited at least some of the time,

the Southeast was predicted to be NOx-limited throughout the modeling domain. Because the

anthropogenic VOC emissions generally are believed to be underestimated, the sensitivity to

NOx emissions could be even greater than the model indicated. Because of the large-scale grid

resolution (18.5 km), Roselle and Schere caution against use of these findings in making

management decisions for areas near large NOx and VOC sources.

• This study was one of the first to explore ozone sensitivity to NOx and VOC reductions on regional scales. While the ROM model has perhaps been supplanted by higher-resolution, newer-generation models, the basic results shown in this study probably still prevail and are consistent with regional observations carried out under SOS for rural NOx sensitivity.

KEY CITATION: Roselle, S. and K. Schere. 1995. Modeled response of photochemical oxidants to systematic reductions in

anthropogenic volatile organic compounds and NOx emissions. J. Geophys. Res. Vol.100, No. D11: 22929-22941.

30 June 2004 | State of SOS-3: 1995 - 2003 | 135

2.10.6 Model Performance Evaluation Using Chemical Signals Comparisons of UAM and UMICH model results with observations were carried out during

the 1992 Atlanta Intensive. Two days were examined in detail – 10-11 August 1992. Vertical

variation of primary organic species was used to assess the extent of vertical mixing of the

model. The model underestimated isoprene, but overestimated reaction products of isoprene

such as HCHO. Meteorology could be manipulated within the range of uncertainty to alter the

modeled ratio of O3/NOy from too low to too high. Most of the model domain showed peak

ozone sensitivity to NOx and little sensitivity to ROG. However the locations downwind from

Atlanta, characterized by the highest peak ozone concentrations, also include ROG-sensitive

chemistry. Peak ozone was more sensitive to NOx in all cases. The O3/NOy ratio was used as a

test for O3-NOx-ROG sensitivity. Scenarios created to force the system into ROG sensitivity

tended to estimate O3 correctly but underestimated the O3/NOy ratio.

Net ozone production efficiencies (OPE) based on observed O3/NOz slopes likely

overestimate the true net OPE. This is due to the rapid deposition of NOz species, especially

HNO3. Model results were consistent with a dry deposition velocity of HNO3 and H2O2 of 5 cm

s-1, which is higher than values commonly used (2.5 cm s-1) but are consistent with

measurements of differential loss rates between SO2 and NOy observed by the TVA helicopter in

power plant plumes. The modeled OPE for various sources were compared, taking into account

the NOz loss. The OPE for Nashville and a small power plant, Gallatin, were similar at about 3.2

to 5, while the OPE for the much larger Cumberland power plant ranged from 2 to 3. The

indicator ratios in the Nashville urban plume indicate that the chemistry is intermediate between

VOC- and NOx-limited. The authors could not distinguish firmly between the two alternatives.

• These studies demonstrate the power of vertical concentration profiles, indicator ratios, and ozone production efficiencies in diagnosing model outputs. Simply achieving an acceptable statistical performance for ozone is not sufficient to ensure that control actions taken based on model results will have the desired effectiveness.

KEY CITATIONS: Sillman, S., K. Al-Wali, F. Marsik, P. Nowacki, P. Samson, M. Rodgers, L. Garland, J. Martinez, R. Imhoff,

J. Lee, L. Newman, J. Weinstein-Lloyd, and V. Aneja. 1995. Photochemistry of ozone formation in Atlanta, GA: Models and measurements. Atmos. Env. 29:3055-3066.

Sillman, S., D. He, M.R. Pippin, P.H. Daum, D.G. Imre, L.I. Kleinman, J.H. Lee, and J. Weinstein-Lloyd. 1998. Model correlations for ozone, reactive nitrogen, and peroxides for Nashville in comparison with measurements: Implications for O3-NOx-hydrocarbon chemistry. J. Geophys. Res. 103:22,629-22,644.

30 June 2004 | State of SOS-3: 1995 - 2003 | 136

2.10.7 Role of Mixing Processes in Emissions-Based Models In a study using UAM-IV in Atlanta, it was found that air pollution concentrations depend

strongly on vertical diffusivity used in UAM-IV modeling. Vertical mixing is believed to be

overestimated during the day by the normal procedures used in UAM-IV. The calculated

vertical diffusion coefficient, Kv, exceeds measured and comparable literature values. Although

generated using the UAM-IV model and the UAM-IV Kv parameterization, the results show that

a large error potential is possible in air quality grid models due to Kv parameterization. Varying

the Kv made little difference to the surface O3 concentrations, but did change significantly the

vertical profile of O3 concentrations. Over urban areas, the surface NOy concentration is

increased using lower Kv ‘s due to the high surface source strength. For reactive hydrocarbons

such as isoprene, lower Kv ‘s can more than double the surface concentrations in the model,

improving the comparison with measurements.

There has been considerable discussion about vertical profiles during several of the SOS

observational campaigns and their consistency with modeling results such as the study above.

When comparing model profiles to observed profiles it is very important that the observed

profiles are averaged for long periods since the first order closure models are based on ensemble

statistics in which the effect of individual eddies have been averaged out. Otherwise significant

structure in the vertical profiles even in a convective boundary can exist. Large-scale

fluctuations about the long-term mean can be manifested especially in the deep boundary layers

in the Southeast. In an effort to understand the appropriate averaging time for reactive pollutants

in convective boundary layers a coupled large eddy (LES) /photochemical model was built and

tested for the Southeast. It allows one to look at the variation in the vertical profiles of reactive

compounds such as isoprene and NOx for different averaging times. It also allows one to address

the question of whether the fluctuations that do exist in the real atmosphere but don’t exist in a

first order closure model have any impact on the ultimate chemical solution. In probably the first

LES study of this kind using a complete chemical mechanism, it was shown that isoprene

profiles have to be averaged for several hours to produce a statistically stationary profile. It also

showed that in large part first order closure models can be trusted within the deep boundary

layers of the Southeast, i.e., there was agreement between the LES Chem model and first order

closure models. It did suggest, however, that the chemistry can evolve quite differently in the

LES Chem and first order closure models just outside the planetary boundary layer.

30 June 2004 | State of SOS-3: 1995 - 2003 | 137

• Recent applications of CMAQ for PM2.5 show that results can be sensitive to the parameterization of vertical mixing, and minimum diffusivities assumed. At present, there does not appear to be a universally optimal approach. This can be critical for current efforts by Regional Planning Organizations dealing with visibility assessments and related policy decisions.

• The coupled LES chemical models results show that the turbulence paradigm that most air quality models are based upon, i.e. first order closure, can be trusted even within the deep boundary layers of the Southeast. However, details on how the first order closures (K profiles) are formulated can make a difference in model results. Thus, additional research is needed to ensure that the K-profiles imposed in air quality models reflect the appropriate turbulent intensities and scales in real-world boundary layers.

KEY CITATIONS: Nowacki, P., P. Samson, and S. Sillman. 1996. Sensitivity of urban airshed model (UAM-IV) calculated air

pollutant concentrations to the vertical diffusion parameterization during convective meteorological situations. J. Appl. Met. 35:1790-1803.

Herwehe, J.H. 2000. A numerical study of of the effects of large eddies on the trace gas measurements and chemistry in the convective boundary layer. Ph.D Dissertation, Atmospheric Science Department, University of Alabama in Huntsville.

30 June 2004 | State of SOS-3: 1995 - 2003 | 138

2.10.8 Implementation of DDM-3D in URM and CMAQ As part of SOS and related studies, the Decoupled Direct Method in Three Dimensions

(DDM-3D) was implemented in CMAQ and used for source apportionment within the Fall Line

Air Quality Study and for assessing the impact of Ohio-based power plants on ozone and

particulate matter formation in the Fall Line region of GA. Previously, DDM-3D was

implemented in the Urban-to-Regional Multiscale Model (URM), and, as part of SAMI, was at

the core of assessing the transport and chemistry of pollutants from one source region to another

and how emissions controls might improve air quality. As part of that study, the importance of

long-range transport was quantified. Further, the use of annual averaging approaches led to the

discovery that while over any one episode, it often appears as though transport from distant states

can be very large, when averaged over a large number of days, more local sources have a

proportionally larger impact. Thus, air quality management should continue to consider local

sources, as well as upwind sources, as part of an overall strategy, and should not focus on just a

few days and meteorological patterns.

Continued research in the Fall Line Air Quality Study being conducted by SOS

researchers shows that DDM-3D in CMAQ can be a valuable tool for assessing point source

impacts and can distinguish between linear and non-linear responses. Key findings include:

• Non-linearities can be important for major sources (e.g., power plants) when controls of greater than about 50% are simulated. This work also finds that in each city investigated, local sources, on top of a “regional background” have the major, if not dominating, role in the source-receptor relationships found, but the make-up of other sources that impact the cities was significantly different. On a ppt-of-ozone per ton-of-NOx basis, local, ground-level sources tend to have the greatest effectiveness.

DDM-3D, as implemented in CMAQ and other advanced air quality models, provides a

method for conducting detailed source-impact assessment. One of the weaknesses of

photochemical models has been their inability to quantitatively assess the impact of individual

sources. DDM-3D implemented in an air quality model, provides an effective and well tested

method, and can be used to study the transport, chemistry and fate of pollutants from many

sources at the same time. Example applications include quantifying the ozone formation

potential of VOCs and the development of source-receptor matrices.

• Consideration of air quality impacts across multiple episodes increases the importance of local sources. Regional modeling of source impacts on air quality show that while long range transport of pollutants and their precursors can appear very significant during any one episode, when viewed across multiple episodes and

30 June 2004 | State of SOS-3: 1995 - 2003 | 139

meteorological conditions, the impact of local emissions becomes of increased importance. Different long-distance source regions affect an area under different conditions, while local sources tend to have a role on a day-in/day-out basis. Thus, local controls will still have benefits, even in areas that are impacted from more distant sources.

KEY CITATIONS: Boylan, J.W., M.T. Odman, J.G. Wilkinson, A.G. Russell, K.G. Doty, W. B. Norris, and R T. McNider. 2002.

Development of a comprehensive muiltiscale ‘one-atmosphere’ modeling system: Application to the Southern Appalachian Mountains. Atmos. Environ. 36 (23):3721-3734.

Odman, M.T., J.W. Boylan, J.G. Wilkinson, A.G. Russell, S.F. Mueller, R.E. Imhoff, K.G. Doty, W.B. Norris, and R.T. McNider. 2002a. Integrated modeling for air quality assessment: The Southern Appalachian Mountains Initiative project. Journal de Physique IV, Vol. 12(PR10): 211-234.

Southern Appalachians Mountains Initiative. 2002. Final Report of the Southern Appalachians Mountains Initiative, Asheville, NC.

30 June 2004 | State of SOS-3: 1995 - 2003 | 140

2.10.9 Development of Reactivity Metrics for Volatile Organic Compounds The hundreds, possibly thousands of organic compounds emitted into the atmosphere each

impact ozone (and PM2.5) differently, and are emitted with differing temporal and spatial

characteristics depending on the major sources of the compounds. EPA is considering regulatory

strategies to account for the differing impacts, or ‘reactivity’ of organic compounds. VOC

reactivities, using both 1- and 8-hour average ozone metrics, were quantified spatially for

approximately 20 individual VOC using two multi-day episodes (20 total days) to show the

variability in the reactivities for varying meteorologies, along with the spatial variability over the

eastern U.S. To further assess VOC variability, two different emissions amounts were used, one

representing 1995 conditions, the other 2010 conditions.

While the absolute reactivity of various compounds showed significant variability between

metrics used to quantify ozone formation potential, the relative reactivities showed remarkably

similar results. Modeling suggests that even across scales as large as the eastern United States,

the relative ozone forming potential of VOC does not vary tremendously. This provides a means

for regulatory agencies to base control policies and provide for a scientifically based approach to

trade VOC emissions and/or reformulate products to improve air quality.

• Ozone formation potentials of individual VOCs are consistent between metrics across wide geographical domains. As noted in other SOS-related research, VOC controls are potentially beneficial in many urban areas. However, individual VOCs can have very different impacts on the amount of ozone formed. Currently, the regulatory structure treats all VOCs as either reactive or nonreactive, not taking in to account the spectrum of impacts. Regional modeling showed that the use of relative reactivities led to consistent results for ozone reactivity, suggesting that such a scale can be used in regulatory policy setting.

KEY CITATIONS: Hakami, A., M. Bergin, and A.G. Russell. 2004. Ozone Formation Potential of Organic Compounds in the Eastern

US: A Comparison of Episodes, Inventories, and Domains. (submitted) Hakami et al. 2003. Final Report to the California Air Resources Board, Sacramento, CA.

30 June 2004 | State of SOS-3: 1995 - 2003 | 141

2.10.10 Assimilation of Satellite Data into MM5 In a joint activity between SOS and an EPA STAR grant on satellite assimilation, GOES

satellite data was processed for the SOS 1995 Nashville and Atlanta field program periods. A set

of CDs with the processed data was developed and transferred to EPA/NERL’s Atmospheric

Modeling Division. Data was processed on the model grid for the resolutions agreed upon by

EPA/NERL, TVA, and Georgia Tech. Preliminary model runs of MM5 with and without the

satellite assimilation were used. Issues related to satellite data assimilation were discovered and

corrected using TexAQS 2000 data. This data set developed to facilitate evaluation of the value

of assimilating satellite data into MM5 is now ready for further testing and evaluation.

KEY CITATION: Biazar, A. 2003. Annual Report to EPA’s STAR Grants Program. Assimilation of Satellite Data into Regional

Scale Models.

30 June 2004 | State of SOS-3: 1995 - 2003 | 142

2.10.11 Chemical Indicators The goal of this part of the SOS-MME research program is to interpret measurements made

in association with the Texas Air Quality Study (Houston) during 2000. Sandy Sillman and his

student and other faculty colleagues at the University of Michigan have responsibility for the

following activities:

1. Development of applications of air quality models for TexAQS 2000 studies, with emphasis on adequate representation of ozone chemistry.

2. Comparison of model and measured values for species that can serve as "indicators" for VOC-NOx sensitivity in Houston, or that provide a basis for evaluating photochemical processes associated with ozone formation. These include:

a) Reactive nitrogen and peroxides b) Various specific VOC species c) OH and HO2 radicals

Research during 2002-2003 focused on applications of CMAQ within the TexAQS 2000

Study. Model runs were developed based on meteorological inputs from John Nielsen-Gammon

at Texas A&M University were transferred to the University of Michigan. A version of CMAQ

is being implemented with SAPRC99 chemistry. The SAPRC 99 chemistry is much more

complete than mechanisms such as CB4 and represent more species explicitly. It is therefore

better suited to studies that involve comparisons with ambient measurements.

A preliminary version of CMAQ was applied to a single event (August 29, 2000) using

12x12 km resolution. The preliminary version was used as a basis for interpreting measurements

of photochemically reactive VOC species from the TexAQS 2000 Study in cooperation with the

Brookhaven National Laboratory

The preliminary results compared ozone, reactive nitrogen species, and H2O2 between

models and measurements for Houston, and between Houston and previous SOS results from

SOS’ Nashville ’95 and ’99 studies. These studies suggested that ozone, reactive nitrogen, and

H2O2 all show a consistent correlation pattern, despite large differences in precursor

concentrations between Houston and Nashville. The distinctive features are:

• A wide range of values was observed for H2O2 and for summed NOx reaction products (referred to as NOz, or NOy-NOx), with no correlation between high H2O2 and high NOz.

• There is a strong correlation between ozone and the sum NOz+2H2O2, which appears to be virtually identical between Nashville and Houston.

KEY CITATION: Sillman, S. 2004. Observation-based methods (OBMs) for analyzing urban/regional ozone production and ozone-NOx-VOC

sensitivity. http://www.personal.engin.umich.edu/~sillman/obm.htm.

30 June 2004 | State of SOS-3: 1995 - 2003 | 143

2.10.12 Analysis of Winds Observed by Radar Profilers A considerable amount of wind data was collected during TexAQS 2000 to assist in

understanding important features of the wind fields in the Houston area. Besides instruments for

measuring surface winds, six radar wind profilers and a Doppler lidar were used to provide high

temporal resolution of the wind fields from near the surface up to a few thousand meters above

the surface at seven locations in the study area. SOS scientists at UAH prepared a report for the

Texas Natural Resources Conservation Commission (TNRCC) [now the Texas Commission on

Environmental Quality (TCEQ)] that describes an initial analysis of the radar profiler data.

Of greatest immediate and practical importance is the guidance this report provides for those

who model the TexAQS 2000 period using data assimilation to nudge the computed wind fields

toward those that were observed. By means of structure functions, estimates were made of the

radius of applicability of wind variables such as the u- and v-components of velocity. This type

of information is required input for models employing data assimilation.

Information about these structure functions help answer such questions as: “If a wind value,

say wind speed, observed at point A is incorporated into a model, what is the size of the region

around A such that the wind speed at A is a good representation of wind speed throughout the

region?” Structure functions provide a way to answer such questions since the independent

variable is distance from a base location and the square root of the dependent variable is the wind

characteristic of interest expressed in its original units. This allows us to pose the question in a

more quantitative form. Again using wind speed as an example, we can ask “How far from the

base location must one travel in order for the average difference in wind speed, due to horizontal

separation from the base, to be 1.0 m/s (or any other selected speed)?” Such a distance can be

taken as the radius of a circle centered on the base location. The circle is the region of

applicability of the wind value.

• The larger the radius, the more horizontally homogeneous the field of a specified wind variable. The magnitude of the radius for a specified amount of change in a wind variable depends on the wind variable chosen. The radius computed for wind speed for a given location will, in general, not be the same as the radius computed for the u- or v-component. Each variable has its individual degree of degradation with distance from the base location.

KEY CITATION: Norris, W.B. 2003. Final Report to Texas Commission on Environmental Quality.

30 June 2004 | State of SOS-3: 1995 - 2003 | 144

2.10.13 Development of a Web-Based Lagrangian Particle Model The Lagrangian Particle Model (LPM) developed at UAH has proven to be a useful tool for

visualizing the transport and dispersion of atmospheric emissions. Since the model uses particles

only as massless tracers, it is equally applicable to either gaseous or fine particulate emissions.

When used for retrospective analyses of field study data, the model can help to determine

whether and when plumes pass over monitor sites or whether aircraft flight paths intersected

plumes. Based on conditioned particle concepts in which the Lagrangian turbulent velocity

fluctuation of a non-buoyant tracer particle is linearly related to the turbulent fluctuation of the

previous time step, the model deduces subgrid-scale velocities (turbulence) from the mean flow

parametrically. Unlike simpler models, which use only mean flow to compute a single

trajectory, the LPM simultaneously computes the trajectories of many particles. By accounting

for the effects of both the mean flow and turbulence on each particle, the model is able to depict

both transport and dispersion of plumes.

Time series of mean meteorological fields are provided to the LPM by mesoscale models

such as RAMS and MM5. The LPM inherits the domain and grid resolution of the

meteorological model (although the data may need to be interpolated from one type of grid

structure to another). The user determines where emission points should be located and the rate

at which particles are to be released from those points. The sources may be elevated. A plume-

rise algorithm is included to handle buoyant emissions. Output consists of a series of snapshots

of particle positions within the model domain. Time series of snapshots can be animated.

UAH modified the LPM so that it can be operated through a web-page interface

(http://texaqs.nsstc.uah.edu). This involved four main tasks: (1) converting the MM5 file produced

by Texas A&M for the TexAQS 2000 study period to suitable input for the LPM, (2) developing

a web-page interface for the model, (3) modifying the LPM to make it functional in a web-based

environment, and (4) adapting the graphics programs needed to display the LPM results on the

web. The interface gives users the opportunity to tailor a modeling run to fit the specifications of

a particular problem. The user selects a domain (108-km, 36-km, or 12 km horizontal

resolution), a period of simulation, source locations and types (buoyant or non-buoyant), and the

exit conditions (for buoyant sources). When the form containing these choices is completed, the

user submits it and waits for the LPM to run. When the run is completed, the user is given

30 June 2004 | State of SOS-3: 1995 - 2003 | 145

access to a temporal sequence of particle-position snapshots depicting transport and dispersion of

particles released from the sources.

KEY CITATION: McNider, R.T. 2004. TexAQS 2000 Web-Based Lagrangian Particle Model. http://texaqs.nsstc.uah.edu

2.10.14 Photochemical Box Model Simulations of Ozone Productivities in Houston • Results from box model simulations run under conditions based on Houston’s industrial

regions suggest that emissions of as little as 100 pounds of light alkenes (ethylene, propylene, butenes, pentenes, butadiene) and aromatics can lead to >50 ppb enhancements of ozone concentrations over a 1 km2

area. Ozone productivities of alkane emissions are generally significantly lower than for alkenes and aromatics. The box model simulations also indicate much higher ozone productivities under conditions that involve high concentrations of both VOC and NOx, as opposed to conditions that involve high concentrations of VOC alone.

KEY CITATION: Daum, P.H., J. Meagher, D. Allen, and C. Durrenberger. 2002. Accelerated Science Evaluation of Ozone

Formation in the Houston-Galveston Area. Summary. 6 pp. http://www.utexas.edu/research/ceer/texaqsarchive/accelerated.htm

2.10.15 UAM in Forecast Mode • The Urban Airshed Model, a Eulerian 3-dimensional photochemical-transport grid

model, has been used in forecast mode to predict next-day ozone concentrations. Results show that ozone is predicted with better accuracy when concentrations are high (>.095 ppmv).

KEY CITATION: Chang, M.E. and C. Cardelino. 2000. Application of the Urban Airshed Model to forecasting next-day peak ozone

concentrations in Atlanta, Georgia. J. Air Waste Manage. Assoc. 50:2010-2024.

30 June 2004 | State of SOS-3: 1995 - 2003 | 146

2.11. Integration of Observation- and Emissions-Based Modeling

Strengths and limitations Emissions inventory

assessment

2.11. Integration of Observation- and Emissions-based Modeling Ted Russell and Carlos Cardelino

As noted in sections 2.9 and 2.10, a number of key

findings have been developed from SOS’ observation-

based and emissions-based modeling research. Both types

of models have their own particular strengths and

weaknesses, and are generally complementary.

2.11.1 Strength and Limitations • Emissions-based models rely on accurately knowing the emissions inputs to models. Such

information can provide an assessment of source impacts. However, such models do not always provide reliable simulations of observed air concentrations. Observation-based models rely on ambient pollutant concentration measurements in the atmosphere, and do not provide detailed information about the sources of emissions observed. The two types of models can be formally linked to provide added confidence in air quality modeling results.

Observation-based methods will become more important and powerful as the amount of

available observational data increases, particularly as a result of the new PM monitoring network

and the availability of satellite data. At this point, however, combined emissions-based and

observation-based modeling approaches are still in their infancy.

2.11.2 Assessment of Emission Inventories Four-dimensional data assimilation coupling an inverse, error-minimizing algorithm with an Air

Quality Model was used to minimize differences between simulated and measured concentrations of

gas-phase and aerosol species to assess biases in the emissions inventory. Ambient measurements

of gas-phase inorganic and organic species and speciated PM2.5 in the eastern US for three multi-

day episodes, and the procedure used to derive emission adjustments of different sources of NOx,

VOC, CO, SO2, NH3, and fine organic aerosol emissions. Results suggest that:

• Anthropogenic SO2, elevated point-source NOx, and biogenic VOC emissions were estimated reasonably well in inventories, while area-source anthropogenic NOx, anthropogenic VOC, NH3, and organic PM2.5 emissions may be significantly biased and require revisions from the base-case inventories. The discrepancies are being investigated in follow on studies.

KEY CITATIONS: Mendoza-Dominguez, A. and A.G Russell. 2001a. Emission strength validation using four-dimensional data assimilation:

Application to primary aerosol and precursors to ozone and secondary aerosol. J. Air & Waste Manage. Assoc. 51:1538-1550.

Mendoza-Dominguez, A., and A.G. Russell. 2001b. Estimation of emission adjustments from the application of four-dimensional data assimilation to photochemical air quality modeling. Atmos. Environ. 35:2879-2894.

30 June 2004 | State of SOS-3: 1995 - 2003 | 147

2.12. SOS Contributions in Ozone Effects Research

Secondary standard recommendations

Recommendations for ecological effects research

Needs for assessment of ozone impacts on ecosystems in the Southeast

SOS interactions with the health-effects research community

2.12. SOS CONTRIBUTIONS IN OZONE EFFECTS RESEARCH Walter Heck and Cari Furiness

Ozone remains an intractable air pollution problem in

North America, Europe, and many other parts of the

world. The negative effects of ozone on human health

are well known. Ozone also has been considered the

most insidious and ubiquitous air pollutant affecting the

vegetative component of crop, forest, and natural

ecosystems. During the past five years, SOS has not

participated in experimental research on ozone effects,

either on plants or animals. However, several efforts

were undertaken to summarize and extend already

existing knowledge regarding ozone’s impacts on

ecosystems: 1) to formulate recommendations for a secondary standard for ozone different in

form from the primary standard; 2) to assess critical needs for research on the ecological effects

of ozone; and 3) to synthesize available information regarding ozone impacts on ecosystems in

the SOS region. Key findings of each of these efforts are summarized here.

2.12.1 Recommendations Regarding the Secondary Standard for Ozone – Ecological Effects

SOS sponsored a workshop in January 1996 with the primary purpose of developing

consensus statements pertinent to a Secondary Standard for ozone from a broad group of

ecologists and air quality scientists. Key statements from the workshop are listed below.

1. The public is generally unaware of the sensitivity of plants to ozone. 2. Plants are more sensitive to ozone than humans and should have a more restrictive standard. 3. The effects of ozone on plants are both cumulative and long-term. 4. Workshop participants recommended that the SUM06, summed over a running 90-day

maximum, using values from a 12-hour (0800-1959) daily window, be accepted as the form for the Standard.

5. A SUM06 value between 15 and 20 ppm-hrs was recommended as the most prudent choice for a Secondary Ozone Standard.

6. The Secondary Standard recommended was different in both form and value from the Primary and Secondary Standards that were promulgated by EPA.

KEY CITATION: Heck, W.W. and E.B. Cowling. 1997. The need for a long term cumulative secondary ozone standard: An ecological

perspective. EM January:23-33.

30 June 2004 | State of SOS-3: 1995 - 2003 | 148

2.12.2 Recommendations Regarding Ozone Ecological Effects Research Needs. SOS organized an EPA-sponsored workshop in May 1997 to assess future research needs

related to ozone effects on ecological systems. Key outcomes of the workshop are listed below.

1. A detailed listing of research needs was developed for crop, forest and natural systems (see Table 2.10.1).

2. Participants recommended that EPA take primary responsibility for developing and managing research that would address its specific needs for undertaking another review of the Secondary Standard.

3. Without the recommended research, EPA will not be in a position to address the need for a long-term, cumulative Secondary Standard for ozone and will not find such a standard acceptable.

4. Participants recommended that other agencies (e.g., USDA/ARS and FS) should accept the lead in directing and managing research that fits into their own environmental missions. This should include research needed to understand how ozone affects crop, forest, and natural ecosystems and developing models to help predict expected losses from ozone exposure.

5. EPA and other federal agencies should work cooperatively in their respective responsibilities. 6. In general, research programs should develop integrated teams that include research

scientists, statisticians, economists, and policy analysts, among others, in the planning stages for the research.

Table 2.12.1. Research tasks, identified as needed during the workshop, that should be led by EPA in collaboration with other agencies such as USDA-Agricultural Research Service, USDA-Forest Service, USDOI-Fish and Wildlife Service, and USDOI-National Park Service.* ** Crop Systems Forest Systems Natural Systems Research Across Systems METHODOLOGY - OPEN-TOP CHAMBER EFFECTS

SCALING OF SEEDLING-SAPLING RELATIONSHIPS

FOLIAR INJURY AS AN ENDPOINT - AN ASSESSMENT TOOL

EXPOSURE INDEX

INTERACTIONS - OZONE/ABIOTIC

OZONE GRADIENT STUDIES - USING NATURAL GRADIENTS

OTHER INDICATORS - BIOINDICATORS FOR NATURAL SYSTEMS

OZONE MONITORING NETWORK - RURAL

ECONOMIC LOSSES - VALIDATE CROP LOSS ESTIMATES

OZONE GRADIENT STUDIES - USING NATURAL GRADIENTS

VALUATION DETERMINATION - ECONOMIC TECHNIQUES

* The research highlighted above is essential before a reasonable review of the ecological components of the Secondary Standard can be accomplished. It is also probably essential before EPA will accept a long-term, cumulative standard. The research is mission oriented and will be accomplished better and more rapidly through the use of cooperative agreements and interagency agreements than through competitive granting mechanisms. ** See text of document for details of research tasks. KEY CITATION: Heck, W.W., C.S. Furiness, E.B. Cowling, and C.K. Sims. 1998. Effects of ozone on crop, forest, and natural

ecosystems: Assessment of research needs. EM October:11-22.

30 June 2004 | State of SOS-3: 1995 - 2003 | 149

2.12.3 Recommendations for Assessing the Impact of Ozone on Ecological Systems in the Southeast.

Following the two workshops described above, an extensive literature review was conducted

to identify literature that could be used in the development of an assessment of the ecological

effects of ozone in the SOS region. The review included a determination of what has been done

and/or is in the process of being done to assess the impact of ozone on ecological systems in the

Southeast. The study was also designed to identify what needs to be done to produce a complete

assessment of ozone impacts on ecological systems in the Southeast. Key findings of this report,

funded by NOAA, are listed below.

1. Current environmental research likely is compromised by a lack of understanding of how ozone affects the response of plants to other factors in the environment.

2. Over 5500 references, relating to ozone effects on ecological systems were surveyed, and 963 were found to be of potential value in undertaking a complete assessment of effects in the Southeast.

3. Nine programs were identified that were designed to undertake some level of effects assessment that could be used in the Southeast.

4. No program has adequately assessed the effects of ozone on ecological systems in the Southeast.

5. Summary statements from the report include:

• Ambient concentrations of ozone in the Southeast cause visible injury to sensitive species of crop, forest, and natural ecosystems.

• Ambient concentrations of ozone in the Southeast can decrease yields of sensitive species of crop, forest, and natural ecosystems.

• Agricultural crop losses across the country were estimated at $1-5 billion in 1988. • Southern commercial loblolly pine was estimated to show 2 to 5% annual growth

reduction at current levels of ozone. This could result in a 10% reduction in stemwood biomass over a ten-year period.

• Growth reduction in forest species at ambient ozone levels in the eastern USA appears to be in the range of 1-25% per year.

• Forest growth losses have not been quantified on an economic basis, either based on loss of productivity of commercial forests or based on degradation of non-commercial assets of Southeastern forests.

KEY CITATION: Heck, W.W. and C.S. Furiness. 1999. Assessment of the Effects of Ozone on Ecological Systems in the Southeast:

An Overview of Known Effects, Assessment Efforts to Date, and Recommendations for a Complete Assessment Report. Report submitted to NOAA.

30 June 2004 | State of SOS-3: 1995 - 2003 | 150

2.12.4 SOS Interactions with the Health-Effects Research Community Only a few SOS scientists have participated directly in field or laboratory research designed

to determine the direct or indirect impacts of ozone or fine particulate matter on humans or

vegetation. Most of the persons are participants in the SOS-affiliated ARIES Program sponsored

by the Southern Company and EPRI (see Section 1.4, Item 4 and Section 2.10.6). In the future,

SOS is committed to continue to work with ARIES and to develop liaison relationships with

organizations that are involved in human health effects research.

During the Nashville/Middle Tennessee Ozone Study, scientists from the Harvard School of

Public Health used personal ozone monitors to characterize individual human exposure to ozone.

During SOS work in Atlanta, an EPRI-sponsored study was conducted to study the correlation

between air quality and pediatric emergency room visits for asthma.

SOS maintains strong ties with the ozone ecological health effects community. As described

in the sections above, SOS has brought together scientists for workshops to discuss issues of

importance to ecological effects of ozone, and utilized existing knowledge about the ecological

health effects of ozone to develop pertinent summary statements, to identify gaps in knowledge,

and to identify the data needs to allow an adequate assessment of ozone impacts on ecosystems

in the SOS region. The next phase of SOS will incorporate a larger research and assessment

component dealing with ecological effects of ozone.

In the case of PM2.5, a major SOS-affiliated research effort was initiated to determine the

epidemiological effects of PM2.5 and other air pollutants in Atlanta; the results should be

applicable to the SOS study region and the nation as a whole. Results from EPRI’s and the

Soutern Company’s Aerosol Research Inhalation Epidemiology Study (ARIES), which began

data collection in July 1998 and finished the second phase of field monitoring in August 2001 are

summarized in Section 2.13.3.

Although there are no known direct effects of PM2.5 on plants, there may be indirect impacts

as a result of decreased insolation with regional haze caused by PM2.5. Certainly decreases in

visibility represent a significant welfare effect of PM2.5 and are important aspects for

consideration in determination of a secondary standard.

30 June 2004 | State of SOS-3: 1995 - 2003 | 151

2.13. Other Studies Relevant to SOS-3 Goals

Southern Center for the Integrated Study of Secondary Air Pollutants

Assessment of Spatial Aerosol Composition in Atlanta

Fall line Air Quality Study SouthEastern Aerosol

Research and Characterization Study

Aerosol Research Inhalation Epidemiology Study

Federal PM2.5 Monitoring Networks

Aerosol Characterization Experiment-Asia and Transport and Chemical Evolution Over the Pacific

2.13. OTHER STUDIES IN SUPPORT OF AND/OR RELEVANT TO SOS-3 GOALS

2.13.1 Southern Center for the Integrated Study of Secondary Air Pollutants (SCISSAP)

The Southern Center for the Integrated Study of

Secondary Air Pollutants (SCISSAP) research initiative

began under the leadership of SOS Chief Scientist, William

Chameides, in April 1998 and continued through March

2002. The project was funded through an EPA competitive

grant to Georgia Institute of Technology. Its objectives

were to answer the following scientific questions:

a) What is the composition and size distribution of fine particles in urban and rural locales in the southern United States, and to what extent do temporal and spatial variations in these parameters correlate with those of ozone and its precursor compounds?

b) What are the major precursor compounds and sources for fine particles in urban and rural locales in the South, and to what extent do these compounds and sources correspond to and/or correlate with the sources of natural and anthropogenic ozone precursors (i.e., VOC and NOx)?

c) To what extent, if any, is the chemical composition of and abundance of fine particles in urban and rural locales in the southern United States affected by the concentration of natural and anthropogenic ozone precursors and/or ozone?

d) To what extent is the concentration of ground-level ozone in urban and rural locales in the southern United States affected by the concentration and composition of fine particles and/or the concentration of the precursors of fine particles?

SCISSAP had both a measurement component, in which mobile measurement capability of

PM2.5, ozone, and their precursors was developed and applied in different field experiments, and

a modeling component, in which a modeling approach that integrated the physics and chemistry

of ozone, acid deposition, and particulate matter simultaneously was developed. The principal

sites for PM2.5 sample collection and analysis in SCISSAP were located in Atlanta, GA and in

Hendersonville and Dickson, TN (see Figure 2.13.1). The SCISSAP team also made

measurements at two sites during TexAQS2000. Measurements were made with a custom-

designed three-channel Particle Composition Monitor and associated ion chromatograph and gas-

chromatograph/mass spectrometer instruments that together permitted determinations of total

30 June 2004 | State of SOS-3: 1995 - 2003 | 152

particle mass as well as the following chemical constituents: NH3, HNO3, HONO, SO2, Na+,

Ca++, NH4+, Cl-, NO2

-, NO3-, SO4

=, total elemental carbon, total organic carbon including formic,

acetic, and oxalic acids, other speciated organic compounds, and trace metals.

Figure 2.13.1. The

SCISSAP urban rural network

Some of the major findings from the study are summarized below.

• PM2.5 composition (at the 24-hour integrated sampling time used in the study) was found to show little variability across the sites operated from Nashville, Tennessee, to Atlanta Georgia, to Houston Texas. In virtually all cases, more than 60% of the PM2.5 mass was found to comprise sulfate (and the ammonium associated with it) and organic carbon (and the other organic elements assumed to be associated with the organic carbon).

• The daily variations in the chemical components of PM2.5 exhibited little or no correlation with their gaseous precursors, and PM2.5 mass was not well correlated with local ozone concentrations.

• PM2.5 mass concentrations showed only moderate increases as one moves from rural to suburban to urban locales.

• Under very humid conditions (e.g., Atlanta in the summer) significant artifacts in the measurement of PM2.5 mass using the filter technique can arise from the presence of solid hydrates on the filter. Negative artifacts in the measurement of particulate organic carbon (OC) using EPA’s FRM filter-denuder technique can arise as a result of the liberation of semi-volatile organics from the filter during the sampling.

• Atmospheric particles of 100 nm and 300 nm in Atlanta at ~3-6% relative humidity typically had two distinct densities: 1.6 ± 0.1 g cm-3 and 0.45 ± 0.20 g cm-3.

30 June 2004 | State of SOS-3: 1995 - 2003 | 153

• PM2.5 in the Southeast is generally slightly acidic, with relatively small amounts of nitrate. • A positive correlation was found between simultaneously measured OVOC

concentrations and speciated, size-segregated particulate OC abundances in Atlanta. Calculation of the hourly new particle production potential from hourly OVOC measurements suggests that gas to particle conversion is a significant source of new organic aerosols. This calculation of new particle production predicts approximately half of the measured PM2.5 total organic carbon observed.

• While ozone and elemental carbon exhibit significant variations between urban and rural regions, most of the other components of PM2.5 have relatively uniform concentrations between urban and rural areas, though certain regions have higher sulfate than others. On the other hand, on urban scales there is a tendency for ozone and PM to be highest in or just downwind of urban areas.

• The major precursors for PM2.5 in the Southeast are SO2 (largely from coal fired power plants) and organic carbon, from a myriad of sources including biogenic (e.g., biomass burning and secondary conversion of higher organics) and anthropogenic (automobiles, cooking, etc.). Ammonia, largely from animal waste and fertilizer use, forms a fraction of the PM mass, but is important as the primary neutralizing agent. For ozone, the two primary precursors are NOx and organics.

• Sensitivity maps show that both ozone and sulfate have similar source-impact patterns. Thus, controls for precursors of both pollutants would be expected to have benefits over the same general area.

• Inverse modeling suggests that the inventory of anthropogenic VOC emissions in the eastern United States is too low by a factor of ~2.

• Over most of the domain, ozone formation is NOx-limited, though not always in urban areas where there can be a greater sensitivity to VOC emissions. Outside of primary emissions of particulate matter, SOx appears to be the most sensitive precursor for PM formation since it also captures ammonia and water.

• Elevated NOx sources were found to be less efficient at forming ozone than ground-level sources, as has been found from aircraft studies as well. Increased emissions, while increasing ozone, can decrease the ozone production efficiency (OPE). There was a much more linear response in SO2 emissions.

• Strategies to reduce NOx and SO2 simultaneously should be effective in reducing ozone and PM at the same time. Both ozone and PM share a largely uncontrollable source, biogenics, which will limit the effectiveness of controls, especially on hot, stagnant days.

• Model results show (and are supported by measurements) that reducing SO2 emissions, and hence PM sulfate, can be offset by increased nitrate aerosol as ammonium is no longer tied up neutralizing the sulfuric acid. This effect will increase in the future as SO2

emissions are reduced due to acid rain controls and ammonia emissions may increase due to increased agricultural operations.

Provisional data and other information can be found at http://www-wlc.eas.gatech.edu/SCISSAP/.

30 June 2004 | State of SOS-3: 1995 - 2003 | 154

KEY CITATION: Chameides, W.L., A.G. Russell, P. McMurry, and R. Zika. 2003. Southern Center for the Integrated Study of Secondary Ai Pollutants. Final Report to EPA for Grant R826372.

30 June 2004 | State of SOS-3: 1995 - 2003 | 155

2.13.2 Assessment of Spatial Aerosol Composition in Atlanta (ASACA) The Assessment of Spatial Aerosol Composition in Atlanta (ASACA) was initiated in 1999

as a long-term study of the sources and temporal and spatial patterns of various species related to

particulate matter. This urban-focused PM2.5 characterization study is designed to provide

further understanding of spatial variability in PM mass and chemical composition in the Atlanta

metropolitan area. It was designed to augment and strengthen the data available from both the

SEARCH regional and ARIES epidemiological studies from the standpoint of spatial variability.

The ASACA program is led by Ted Russell of Georgia Institute of Technology. Funding for

ASACA is provided by the Georgia Power Company. Continuous PM2.5 mass data were

collected at three sites, in addition to 24-hour integrated samples from a manual, filter-based

particle composition monitor; PM2.5 data were collected using a Federal Reference Method

instrument at the fourth site (see Figure 2.13.2). Speciation data were collected at all four sites.

Thus detailed records are available for all months of the year for PM2.5 mass, sulfate, ammonium,

and both elemental and organic carbon. The study is planned to continue indefinitely, allowing

the development of a large database for future research on health, welfare, and model evaluation

of particulate matter.

Figure 2.13.2. Particulate matter measurement sites in ASACA.

30 June 2004 | State of SOS-3: 1995 - 2003 | 156

Butler et al. (2003) have recently published results from analysis of the first year of data from

ASACA. Some of the key findings are summarized below.

• Annual PM2.5 mass concentrations measured from March 1999 to February 2000 in the ASACA study in Atlanta exceeded the annual NAAQS of 15 µg m-3 at all four monitoring sites, with annual averages ranging from 19.3 to 21.2 µg m-3. One site violated the daily NAAQS of 65 µg m-3.

• The composition of PM2.5 was primarily sulfate, with associated ammonium and organic carbon, and smaller amounts of elemental carbon and metal species.

• In Atlanta, most PM2.5 constituents peaked during the summer months; however, nitrate, metals, and elemental carbon showed some enhancement in the winter due to lower inversion heights.

• PM2.5 mass showed discernible early morning and late night peaks that corresponded to rush-hour traffic patterns and inversion heights, respectively.

• Secondary aerosol formation appears to be a significant source of PM2.5. • The metropolitan Atlanta area was found to be spatially homogeneous with regard to

PM2.5 mass and chemical composition.

KEY CITATION: Butler, A.J., M.S. Andrew, and A.G. Russell. 2003. Daily sampling of PM2.5 in Atlanta: Results of the first year of the Assessment of Spatial Aerosol Composition in Atlanta study. J. Geophys. Res. 108(D7), 8415, doi:10.1029/2002JD002234.

30 June 2004 | State of SOS-3: 1995 - 2003 | 157

2.13.3 Fall line Air Quality Study (FAQS) The objectives of the Fall line Air Quality Study (FAQS) are to assess urban and regional air

pollution, identify the sources of pollutants and pollutant precursors, and recommend solutions to

realized and potential poor air quality in three urban areas of Georgia – Augusta, Macon, and

Columbus. The study focuses on ground-level ozone, but observations made will provide

information about fine particulate matter as well. FAQS is funded by the state of Georgia

through the Georgia Environmental Protection Division and was built as an alliance among

stakeholders from state environmental agencies and air quality groups from the three cities. The

project began in the summer of 2000 and will continue through 2004. The study team is based at

Georgia Institute of Technology and is coordinated by Michael Chang. Specific research tasks

are:

a) Enhanced monitoring; b) Emission inventory development; c) Scenario modeling; and d) Analysis, assessment, and recommendation. During Phase I of FAQS in summer 2000, pilot studies in each urban area were conducted

utilizing the SCISSAP mobile Air Quality Research laboratory developed at Georgia Tech, long-

term continuous ozone and PM2.5 monitoring stations established in each city, and data from

existing state monitoring stations. Monitoring sites in the study are shown in Figure 2.13.3.

Figure 2.13.3. PM and ozone

measurement sites in FAQS.

30 June 2004 | State of SOS-3: 1995 - 2003 | 158

Preliminary findings from Phase I of the study completed during the summer of 2000 are

summarized below:

• Like Atlanta, the urban areas of Macon, Columbus, and Augusta appear to be NOx-sensitive with regard to ozone production. Biogenic hydrocarbons are abundant in all three cities.

• Unlike Atlanta, local anthropogenic emissions of NOx and VOC in the three cities are relatively small.

• Fine particulate matter in the FAQS area is composed primarily of organic matter and sulfate.

• Formation and accumulation of ozone in all three cities are affected by regional concentrations of ground-level ozone that extend across much of the Southeast. Augusta and Columbus appear to be particularly influenced by regional ozone concentrations, while Macon ozone concentrations seem to correlate more closely with Atlanta’s ozone concentrations.

Detailed information about the study, air quality data in each city, and reports and briefings

from the study are available at http://cure.eas.gatech.edu/faqs/index.html. Extensive efforts have

been made through the website, email listservs, and public briefings, to keep FAQS partners, the

public, and the media apprised of the status and results of the study.

KEY CITATION: Chang, M.E., K. Baumann, M. Bergin, R. Weber, D. Worsnop, and A.G. Russell. 2001. The Fall Line Air Quality Study Phase I Pilot Study: Initial Impressions of Air Quality in Augusta, Macon, and Columbus, Georgia. Georgia Institute of Technology Report. 165 pp.

30 June 2004 | State of SOS-3: 1995 - 2003 | 159

2.13.4 SouthEastern Aerosol Research and Characterization Study (SEARCH) The SouthEastern Aerosol Research and Characterization Study (SEARCH) evolved from

the Southeastern Consortium: Intermediate Oxidants Network (SCION) component of the

Southern Oxidants Study in 1998 and will be supported through 2005 by Southern Company and

EPRI. SEARCH is made up of eight paired monitoring stations (1 urban-suburban pair and 3

urban-rural pairs) in the states of AL, FL, GA and MS. At each SEARCH site, measurements of

an extensive set of gases, PM2.5 and PM10 mass and composition, meteorology, and extinction

are made. In addition to 24-hour integrated samples for PM, many measurements, including

major components of PM, are made at a temporal resolution of one hour or less. Figure 2.13.4

shows a map of the network. Several of the sites are operated in collaboration with State or local

air monitoring networks. The Jefferson Street site in downtown Atlanta is the focal point of the

epidemiological study ARIES and also was the location of the Atlanta Supersite Project in

August 1999.

Figure 2.13.4. Monitoring sites in the SEARCH network.

Oak Grove

Centreville

Pensacola

Yorkville

Jefferson St.

N.

Birmingham

Gulfport

OLF

rural urban suburban

30 June 2004 | State of SOS-3: 1995 - 2003 | 160

The objectives of the SEARCH program are:

1. To work interactively with the States and to assist them in: a) meeting their PM2.5 monitoring obligations and b) gathering a data set appropriate for evaluating and applying air quality models (as in SIP development).

2. To provide an ambient data set with minimal sample adulteration and of sufficient breadth, in terms of measured variables, geographical diversity and extent, frequency of measurements, and duration of the measurement campaign for the purposes of: a) establishing a PM and oxidant climatology for the region, b) characterizing coarse and fine PM concentrations, c) characterizing the chemical constituents of PM and their physical states and determining correlations among precursor and product materials in the atmosphere, , d) gaining insights into aerosol formation mechanisms, e) documenting differences between airborne materials in coastal vs. inland, and rural vs. urban areas, and f) illustrating biases in measurement methods, such as use of a single Teflon filter in the Federal Reference Method for PM2.5.

3. To deploy reliable continuous methods in order to observe and understand processes governing PM2.5 and co-pollutant emissions, formation, transport, and deposition.

Major policy-relevant findings from SEARCH from 1999 to 2002 are described below.

• Consistent with state findings, the PM2.5 annual standard design values have declined from 1999-2001 to 2000-2002 and no site violated the daily standard. However, in contrast to state findings, only the Atlanta (JST) and Birmingham (BHM) urban areas (and not the nearby rural areas) violated the standard.

• There were only seven (7) days across the SEARCH network that exceeded the 24-hour standard. All 7 days occurred in the first or fourth quarter of the year, when wood smoke is an important source of PM2.5. Both direct and indirect evidence shows that carbonaceous material dominates (> 50%) the composition on each of these 7 days. In contrast, on the three (3) highest sulfate days, sulfate make up less than 50% of the mass. High sulfate days have never exceeded the 24-hour standard.

• Composition: – Organic matter (organic carbon * 1.4) and sulfate are the major components of

PM2.5 in SEARCH. – For Federal Reference Method (FRM) Equivalent mass (an aggregation of PCM

sampler information to replicate what is in the FRM filter): Organic matter is the larger component at the urban sites except GFP. Sulfate is the larger component at the rural sites.

– For Best Estimate mass (an aggregation of PCM sampler information to replicate what is in the atmosphere): Organic matter is the larger component at all sites except YRK.

• Organic matter and sulfate are largely regional in nature. However, organic matter and elemental carbon account for almost all of the difference between urban and rural PM2.5 mass, thus providing evidence of important local contributions to elemental carbon and organic matter.

30 June 2004 | State of SOS-3: 1995 - 2003 | 161

• Diesel exhaust and wood smoke (from various sources) are the most important sources of primary organic carbon across the Southeast on an annual basis. Wood smoke is especially important in winter and spring. Secondary production can be important in the summer.

• Coarse particulate matter has substantial carbonaceous material. However, the contribution of organic matter is highly sensitive to the organic carbon scaling factor. Analysis of organic carbon/elemental carbon, and investigation of appropriate scaling factors for organic carbon, should be included in future monitoring.

• There are many ways to display the composition of particulate matter. The perceived message can be dramatically different depending on which method is used. Therefore, definitions and assumptions should be provided by the author and demanded by the reader.

• SEARCH uses two approaches to illustrate relative composition. The first of these seeks to explain mass reported on FRM filters. The second seeks to explain mass as it exists in the real atmosphere; that is, independent of biases imposed by the FRM sampling methodology. The latter is more useful for comparison with atmospheric models.

• FRM biases can have an important effect on attainment status as well as spatial interpretation of fine mass data. Two important biases identified to date include: 1) failure to blank correct data; and 2) statistically significant differences between vendor’s monitors based on EPA performance evaluation program (PEP) audits.

KEY CITATION: Jansen, J.J., E.S. Edgerton, B.E. Hartsell, and K. Kandasamy. 2003. SEARCH: Key findings for policy makers.

Poster presented at 2003 AAAR PM Meeting, Particulate Matter: Atmospheric Sciences, Exposure and the Fourth Colloquium on PM and Human Health, March 31-April 4, 2003, Pittsburgh, PA.

30 June 2004 | State of SOS-3: 1995 - 2003 | 162

2.13.5 Aerosol Research Inhalation Epidemiology Study (ARIES) Aerosol Research Inhalation Epidemiology Study (ARIES) is an EPRI-sponsored program

based at the Jefferson Street monitoring site in Atlanta that combines daily air quality monitoring

with epidemiology studies. The four components of ARIES are:

Air Quality Characterization: PM2.5 mass and composition, as well as related gas-phase and particle-phase pollutants, are measured with at least 24-hour time resolution. The air quality field measurements include SO2, CO, NO, NO2, NOy, O3, HNO3, NH3, and VOCs in the gas phase; major ions, including acidity, elemental/organic carbon (EC/OC), elements, water-soluble transition metals, and solvent-extractable carbon in the particle phase; pollen and mold; and particle number and size distribution from nanometers to micrometers in diameter. Air Pollution Mortality: daily mortality data are being collected and analyzed in a multi-pollutant ecological time-series study. Air Pollution Morbidity: daily data on emergency room (ER) visits are collected from practically all hospitals in the Atlanta area. The focus is on ER visits for coronary and respiratory symptoms. A parallel study is also being conducted to understand the influence of daily air quality on unscheduled physician visits at a large health maintenance organization. Finally, the health study will evaluate the physiologic responses of a group of patients with more severe cardiac conditions (those with implanted defibrillators). Exposure & Health Assessment: a personal/indoor/outdoor exposure assessment study will help the epidemiologists assess how well ambient measurements can represent personal exposures for groups of individuals with recent heart attacks and with chronic obstructive pulmonary disease. This information may also have applications in validation of personal exposure models. The health study will also examine any association between exposure and cardiac response for these The primary objectives of ARIES are to:

1. Characterize and speciate PM2.5; evaluate its relationship to covarying pollutants; and examine temporal and seasonal variability in the concentration and composition of long term trends and patterns in PM2.5 concentrations and speciation;

2. Investigate (via epidemiological and exposure studies) the associations between air quality and human health, and produce results in time for consideration of the health basis of the NAAQS;

3. Characterize sources of error and uncertainty associated with the recently promulgated Federal Reference Method (FRM) for PM2.5 mass concentration; and

4. Develop, adopt, and modify, if necessary, improved methods for PM2.5 measurement and/or speciation.

To accomplish these objectives, analysis of data from the ARIES study focuses on assessing

the plausibility and tolerances for individual observations, assessing the mutual compatibility of

different measurements of the same factor and investigating those observations that do not agree,

and identifying empirical patterns in the data. ARIES will address the impacts of and factors

30 June 2004 | State of SOS-3: 1995 - 2003 | 163

related to positive and negative sampling artifacts for organics and nitrate, identifying empirical

patterns and correlations between different factors. To facilitate cooperation and constructive

feedback between the air quality and epidemiology investigators, regular periodic reports

summarizing the measurement results and characterizing epistemic and aleatory uncertainty are

produced and distributed to principal investigators and stakeholders.

Preliminary findings include (Edgerton et al., 2003):

• Elemental and/or organic carbon (important components of PM2.5, especially in urban areas) are often associated with health endpoints, sometimes along with PM2.5 mass.

• Pollutant gases appear to be as important as PM components.

• Sulfate (the other major component of PM2.5) is not significantly associated with any health endpoint.

• There is consistency across studies and health endpoints. • There is an overall association between air pollution and health at current levels.

• Respiratory diseases appear to be associated with PM10, coarse particles, and gases.

• Heart diseases seem to be linked with PM2.5, organics, and gases.

KEY CITATION: Edgerton, E., J. Jansen, and K. Kandasamy. 2003. An ARIES, SEARCH, and Mercury Update. Presentation to

U.S. Environmental Protection Agency, Research Triangle Park, NC. April 17, 2003 Van Loy, M., T. Bahadori, R. Wyzga, B Hartsell, and E. Edgerton. 2000. The Aerosol Research Inhalation

Epidemiology Study (ARIES): PM2.5 mass and aerosol component concentrations and sampler intercomparison. J. Air Waste Manage. Assoc. 50:1446-1458.

30 June 2004 | State of SOS-3: 1995 - 2003 | 164

2.13.6 Federal PM2.5 Monitoring Networks As part of the revised National Ambient Air Quality Standards for PM2.5 promulgated in

1997, EPA is required to implement a PM2.5 monitoring program. The program is made up of

three integrated networks:

Federal Reference Methods (FRM) network, with more than 1100 monitoring sites across the US, is designed to monitor PM2.5 mass for compliance with the new National Ambient Air Quality Standards for PM2.5.

National PM2.5 Speciation Trends Network (STN) has about 250 sites across the US, of which 54 are part of EPA’s long-term National Air Monitoring Stations network for measuring air pollutants and the rest are part of the State and Local Air Monitoring Stations network. STN is designed to measure PM2.5 mass and the major chemical components of PM, including sulfate, nitrate, ammonium, organic material, elemental carbon, and crustal-related material, to provide data in support of State Implementation Plan development.

The Supersites Program is an ambient monitoring program aimed at addressing the scientific uncertainties of measuring particulate matter. The program is designed to meet the following objectives:

1) Characterize particulate matter: to obtain atmospheric measurements to characterize PM, its constituents, precursors, co-pollutants, atmospheric transport, and source categories that affect the PM in any region. This information is essential for understanding source-receptor relationships and the factors that affect PM at a given site (e.g., meteorology, sources, transport distances). This information is also essential for improving the scientific foundation for atmospheric models that investigate exposure and risk management questions.

2) Support health effects and exposure research: to obtain atmospheric measurements to address the research questions and scientific uncertainties about PM source-receptorexposure- effects relationships. Examples of these questions include, "What is the relationship between sources, ambient PM concentrations, human exposures, and health effects such as respiratory tract disease and mortality?" and "What is the biological basis for these relationships?"

3) Conduct methods testing: to obtain atmospheric measurements that will compare and evaluate different methods of characterizing PM (e.g., emerging sampling methods, routine monitoring techniques, and Federal Reference Methods). Testing new and emerging measurement methods ultimately may advance the scientific community's ability to investigate exposure and effects questions significantly.

Atlanta was chosen as the site of one of the two Phase I EPA Supersites, under the leadership

of Bill Chameides of Georgia Tech, in part because of SOS’ expertise and involvement in

ongoing research there. Seven sites have been chosen as Phase II Supersites, including Houston,

TX, under the leadership of David Allen at the University of Texas at Austin. Information about

the PM Supersites can be found at http://www.epa.gov/ttn/amtic/supersites.html.

30 June 2004 | State of SOS-3: 1995 - 2003 | 165

Based on the results from Phase II of the Supersites Program, the following policy-relevant

findings can be noted (excerpted from Solomon and Allen, 2004): • Understanding source-receptor relationships is complicated by the fact that PM observed in

urban locations, where maximum PM mass concentrations are observed, can be impacted by significant regional background with local sources superimposed. The regional influence may be from upwind urban or rural areas as well as be composed of both biogenic and anthropogenic sources with a considerable secondary aerosol component.

• Recent toxicological findings suggest that ultrafine PM may be responsible for some observed health effects. However, results presented here indicate that PM2.5 mass and ultrafine particles (mass and number) are not correlated, and therefore, PM2.5 mass cannot act as a surrogate for ultrafine particles. Thus control of the bulk of PM2.5 may have little impact on protecting public health from ultrafine-related health effects.

• Primary ultrafine particles appear to be observed in much higher concentrations near sources (e.g., motor vehicle, industrial), suggesting a higher exposure to ultrafine PM when in close proximity to these types of sources (e.g., during rush hour traffic). As well, regional “nucleation bursts” of ultrafine particles have been observed, suggesting widespread population exposure to ultrafine particles, and these tend to occur often and under clean conditions when PM2.5 mass is low.

• New continuous monitoring methods likely provide an improved estimate (less bias due to potential positive or negative artifacts) of PM2.5 mass in air than current filter-based methods.

• No significant change has occurred to the bulk composition of PM2.5 with organic carbon and sulfate as the primary species observed in the eastern half of the U.S. during either summer or annual average conditions.

• As with the unusually high rate of ozone production (Ryerson et al., 2003), the formation of certain PM components (carbonyl and organic nitrates) in Houston appear to be unique to the mix and simultaneous emissions of organic and nitrogen oxide precursor species. While some fraction of these compounds may be primary pollutants, they appear to be secondary with abundances in particles less than 1 µm and in the ultrafine PM size fraction.

KEY CITATION: Solomon, P.A. and D. Allen. 2004. Preface to Special Issue of Aerosol Science and Technology on Findings from the Fine

Particulate Matter Supersites Program. Aerosol. Sci. Technol. 38(S1):1-4, doi:10.1080/02786820390229138.

30 June 2004 | State of SOS-3: 1995 - 2003 | 166

2.13.7. Aerosol Characterization Experiment-Asia (ACE-Asia) and Transport and Chemical Evolution Over the Pacific (TRACE-P)

Development of the Particle-Into Liquid Sampler (PILS) by the Georgia Institute of

Technology and its utilization in several research studies is an excellent demonstration of the

broader impacts of instrumentation developed with SOS support. This instrument has been

deployed at the EPA Supersites at New York and St Louis. The instrument also has been used

successfully on three airborne experiments: NSF-funded ACE-Asia (2001), NASA-funded

TRACE-P (2001), and NOAA-funded ITCT-2K2 (2002). Highlights of the ACE-Asia and

TRACE-P results are summarized below.

• The chemical characteristics of fine particles advecting from Asia are summarized Lee et al., 2003, JGR.

• Intercomparisons made between airborne methods for measuring particle ionic composition between various instruments on each aircraft and between aircraft during extended periods of formation flying [Ma et al., 2003a] indicate that elative differences were typically less than ±30%, similar to what is found in ground-based studies [Drewnick et al., 2003; Weber et al., 2003a]. The PILS composition measurements were also consistent with other indirect measurements of particle composition and particle volume determined from measured size distributions. Moore et al., 2003, JGR.

• Biomass burning aerosol sources during the experiment were investigated using PILS fine particle potassium (K+) as a particle tracer for biomass smoke. Ratios of potassium to sulfate are used to estimate the percent contribution of biomass burning to fine particle mass in mixed plumes advecting from Asia. Ma et al., 2003b, JGR. High correlations between K+ and NO3

- and NH4+ indicated that biomass burning was a significant source of

these aerosol compounds in the region. We estimate that the most polluted plume recorded in both missions was composed of ~60% biomass burning emissions, possibly from the use of bio-fuels in the urban and surrounding regions of Beijing.

• Water-soluble calcium and magnesium recorded in large dust storms are used to investigate possible interactions of reactive acidic gases with mineral dust particles [Meier et al., 2003]. It is concluded that the mass accommodation coefficients for uptake of acidic gases by mineral dust are orders of magnitude lower than 0.1, the value currently used in model simulations.

• PILS composition data was also used to investigate the conditions leading to homogenous nucleation in Asia plumes [McNaughton et al., 2003; Weber et al., 2003b], for comparisons to validate model simulations of aerosol advection from Asia [Browell et al., 2003; Chin et al., 2003; Koike et al., 2003; Tang et al., 2003a; Tang et al., 2003b; Zhang et al., 2003], and for detailed investigations of the Asian aerosol optical properties [Clarke et al., 2003].

2.5. pm2.5 composition, sources, and air …2.5. pm2.5 composition, sources, and air concentrations...

Documents