Effects of Aerosol Species on Atmospheric Visibility in Kaohsiung City,Taiwan

Abstract

Visibility data collected from Kaohsiung City, Taiwan, for the past two decades indicated that the air pollutants have significantly degraded visibility in recent years. During our study period, the seasonal mean visibilities in spring, summer, fall, and winter were only 5.4, 9.1, 8.2, and 3.4 km, respectively. To ascertain how urban aerosols influence the visibility, we conducted concurrent visibility monitoring and aerosol sampling in 1999 to identify the principal causes of visibility impairments in the region. In this study, ambient aerosols were sampled and analyzed for 11 constituents, including water-soluble ions and carbon materials, to investigate the chemical composition of Kaohsiung aerosols. Stepwise regression method was used to correlate the impact of aerosol species on visibility impairments. Both seasonal and diurnal variation patterns were found from the monitoring of visibility. Our results showed that light scattering was attributed primarily to aerosols with sizes that range from 0.26 to 0.90 μm, corresponding with the wavelength region of visible light, which accounted for ～72% of the light scattering coefficient. Sulfate was a dominant component that affected both the light scattering coefficient and the visibility in the region. On average, (NH₄)₂SO₄, NH₄NO₃, total carbon, and fine particulate matter (PM_2.5)-remainder contributed 53%, 17%, 16%, and 14% to total light scattering, respectively. An empirical regression model of visibility based on sulfate, elemental carbon, and humidity was developed, and the comparison indicated that visibility in an urban area could be properly simulated by the equation derived herein.     

Optical properties of the San Joaquin Valley aerosol collected during the 1995 integrated monitoring study

Optical, filter chemistry, and cascade impactor data collected during the winter intensive of the IMS95 Study in the San Joaquin Valley (SJV) of California were analyzed to determine the light-extinction efficiency of aerosol species. Regression of light scattering by particles (b_sp) measured by a heated nephelometer without a size selective inlet against PM_2.5 front filter mass gave a scattering efficiency of 3.67±0.05 m²/g with an R² (fraction of variance explained) of 0.94. Division of the aerosol into two components and applying two different corrections to the filter data for nitrate and organic carbon on the backup filter gave scattering efficiencies of 3.7±0.3 or 4.1±0.2 m²/g for the salts composed of sulfate, nitrate, and ammonium and 2.9±0.2 or 3.1±0.2 m²/g for all other species with R² of 0.985 and 0.986. The ambient b_sp measured by an open nephelometer was a simple function of PM_2.5 mass and relative humidity (RH), giving R² of 0.90 and 0.88 for two different RH sensors. Variations in PM_2.5 size distribution and composition did not have an important effect on ambient b_sp. The RH data from each sensor were repeatable enough to show the existence of a simple dependence of aerosol water uptake on RH, but RH sensor calibration uncertainties prevented determining this dependence. Inversion of MOUDI cascade impactor data gave sulfate and nitrate mass median diameters (MMD) between 0.4 and 0.8 μm. Mie scattering calculations based on MOUDI data provided humidity-dependent extinction efficiencies for the principal aerosol chemical species. These efficiencies combined with particle filter data showed that ammonium nitrate was the dominant contributor to wintertime light extinction. Source apportionment showed that light extinction was dominated by emissions sources contributing to the formation of secondary species, especially nitrate. These wintertime data are not expected to apply to summertime in the SJV.     

Impact of Light Duty Diesels on Visibility in California

This paper investigates the impact of light duty diesels on California visibility in the early 1990s. It is found that, without increased dieselization, there will be little change in statewide visibility levels from the late 1970s to the early 1990s. Visibility impacts from diesels are calculated for various scenarios of diesel use and particulate control. The likely dieselization (20%), do nothing particulate control (0.4 g/ml) scenario will change projected statewide emissions slightly for HC ( –2 % ) , NO _x (+1%), SO _x (+5%), and TSP (+1%) but will increase statewide emissions of elemental carbon (soot) by about 80%. Simplistic haze budget calculations indicate that this increase In soot emissions should reduce visibility about 10 to 25% in California. More precise and geographically detailed visibility calculations are performed by applying a lead tracer model to data for 86 California locations. The lead tracer model indicates that the likely dieselization, do nothing control scenario will reduce visibility by about 10 to 35%, with the greatest impacts occurring in and near urban areas. Actual visibility decreases for this scenario may even be much greater, 20 to 50%, because the analysis does not address two other significant factors: (1) increased SO₄ ^–2 levels due to catalytic SO₂ oxidation by soot and to higher SO₂ emissions, and (2) increased soot emissions due to dieselization of the medium and heavy duty fleets.     

Source apportionment of visibility degradation problems in Brisbane (Australia) using the multiple linear regression techniques

Different aspects of visibility degradation problems in Brisbane were investigated through concurrent visibility monitoring and aerosol sampling programs carried out in 1995. The relationship between the light extinction coefficients and aerosol mass/composition was derived by using multiple linear regression techniques. The visibility properties at different sites in Brisbane were found to be correlated with each other on a daily basis, but not correlated with each other hour by hour. The cause of scattering of light by moisture (b_sw) was due to sulphate particles which shift to a larger size under high-humidity conditions. The scattering of light by particulate matter (b_sp) was found to be highly correlated with the mass of fine aerosols, in particular the mass of fine soot, sulphate and non-soil K. For the period studied, on average, the total light extinction coefficient (b_ext) at five sites in Brisbane was 0.65×10⁻⁴ m⁻¹, considerably smaller than those values found in other Australian and overseas cities. On average, the major component of b_ext is b_sp (49% of b_ext), followed by b_ap (the absorption of light, mainly by fine soot particles, 28%), b_sg (Rayleigh scattering, 20%) and b_sw (3%). The absorption of light by NO₂ (b_ag) is expected to contribute less than 5% of b_ext. On average, the percentage contribution of the visibility degrading species to b_ext (excluding b_ag) were: soot (53%), sulphate (21%), Rayleigh scattering (20%), non-soil K (2%) and humidity (3%). In terms of visibility degrading sources, motor vehicles (including soot and the secondary products) are expected to contribute more than half of the b_ext (excluding b_ag) in Brisbane on average, followed by secondary sulphates (17%) and biomass burning (10%).     

Visibility as related to atmospheric aerosol constituents

A laboratory and field study was performed to assess the contribution to visibility reduction of both light scattering and absorption by air pollutant particles and gases. Gaseous precursors to important visibility-reducing aerosol species were measured. Emphasis was placed on minimizing sampling artifacts for nitrate and sulfate since previous visibility studies were generally subject to substantial errors from these sources. Optical techniques for measuring the particle absorption coefficient and elemental carbon were evaluated. The aerosol species measured were fine and coarse particulate mass, sulfate, nitrate and elemental carbon, plus organic carbon and ammonium ion. The gases measured were nitric acid, NH₃, SO₂, NO₂ and O₃. Sampling was done at San Jose, Riverside and downtown Los Angeles. The light-scattering efficiency of fine particulate nitrate appeared to be higher than that of sulfate, in contrast to the findings of most prior studies. At all sites light scattering by sulfate, nitrate and elemental carbon particles contributed more than half of the light extinction. Light absorption by particles, due almost exclusively to elemental carbon, contributed 10–20% of the extinction.     

Particulate contributions to light extinction and local forcing at a rural Illinois site

The light extinction and direct forcing properties of the atmospheric aerosol were investigated for a midwestern rural site (Bondville, IL) using field measurements, a semi-empirical light extinction model, and a radiative transfer code. Model inputs were based on the site measurements of the physical and chemical characteristics of atmospheric aerosol during the spring, summer, fall and winter of 1994. The light scattering and extinction coefficients were calculated and apportioned using the elastic light scattering interactive efficiency (ELSIE) model (Sloane and Wolff, 1985, Atmospheric Environment 19(4), 669–680). The average efficiencies calculated for organic carbon (OC, carbon measured as organic multiplied by 1.2) ranged from 3.81 m²/g OC at lower relative humidities (<63%) to 6.90 m²/g OC at higher relative humidities (>75%) while sulfate (assumed as ammonium sulfate) efficiencies ranged from 1.23 m²/g (NH₄)₂SO₄ to 5.78 m²/g (NH₄)₂SO₄ for the same range of relative humidities. Radiative transfer calculations showed that the rural aerosol at Bondville is most likely to have an overall negative (cooling) forcing effect on climate. Elemental carbon (EC), however, acts to counter sulfate forcing to a degree that has a significant seasonal variation, primarily due to the seasonal variation in the sulfate concentrations. Taking the loading to be the mean summer EC+ammonium sulfate loading and assuming [EC]/[(NH₄)₂SO₄] to be zero in one case (i.e. no soot present) and 0.025 (summer mean at Bondville) in another leads to a 37% difference in calculated forcing.     

Canada’s Mobile Sources Air Pollution Control Program

The chemical and optical properties of particle emissions from onroad vehicles were investigated at the Allegheny Tunnel on the Pennsylvania Turnpike during July 1981. The optical results are in agreement with earlier data: (1) in terms of light extinction per km driven, diesel particle emissions are at least an order of magnitude more important than particle emissions from spark-ignition vehicles; (2) for diesel particle emissions, light absorption is about twice as efficient as light scattering. Chemical analyses showed that: (1) 24% of the vehicle aerosol was extractable material, (2) 75% of the total mass was carbon, (3) 55% of the total mass was unextractable (elemental) carbon, and (4) the stoichiometry of the extractable fraction of the diesel particle emissions was C_nH_t.7nN_0.05n , i.e., the extractable material was composed predominantly of alkanes. The results of the chemical analyses allow the calculation of the massspecific light absorption coefficient for the elemental carbon component of the diesel particle emissions, i.e., 10.9 ± 1.8 m²/g (500 nm).     

An examination of the physical and optical properties of aerosols collected in the IMPROVE program

The Interagency Monitoring of Protected Visual Environments (IMPROVE) protocols for reconstructing the ambient light extinction coefficient (b_ext) from measured aerosol species are the basis for evaluating compliance under the Regional Haze Rule. Aerosol mass composition and optical properties have been measured as part of the IMPROVE program since 1988, providing a long-term data set of aerosol properties at 38 sites around the US. This data set is used to evaluate assumptions made in calculating reconstructed mass and b_ext by applying statistical analysis techniques. In particular, the molecular weight to carbon weight ratio used to compute particulate organic matter is investigated. An annual average value of 1.7±0.2 for the IMPROVE sites, compared to the value of 1.4 currently assumed in the IMPROVE algorithm, is derived. Regression analysis also indicates that fine soil mass concentrations are underestimated by roughly 20% on average. Finally, aerosol mass scattering and extinction efficiencies assumed in the IMPROVE reconstructed b_ext protocol are examined. Fine mode (D_p<2.5 μm) mass scattering efficiencies have a functional dependence on mass concentrations at many sites, and use of a mass-concentration-dependent adjustment factor to refine the assumed efficiencies provides for closer agreement between measured and reconstructed b_ext.     

Variation of Mass Scattering Efficiencies in IMPROVE

Abstract

The Interagency Monitoring of Protected Visual Environments (IMPROVE) equation used to assess compliance under the U.S. Environmental Protection Agency (EPA) Haze Rule assumes that dry mass scattering efficiencies for aerosol chemical components are constant. However, examination of aerosol size distributions and chemical composition during the Big Bend Regional Aerosol and Visibility Observational Study and the Southeastern Aerosol and Visibility Study suggests that volume and mass scattering efficiencies vary directly with increasing particle light scattering and aerosol mass concentration. This is consistent with the observation that particle distributions were shifted to larger sizes under more polluted conditions and appears to be related to aging of the aerosol during transport to remote locations.     

APCA Financial Statements for FY 1979-80

The U.S. Environmental Protection Agency (EPA) has proposed a new secondary standard based on visibility in urban areas. The proposed standard will be based on light extinction, calculated from 24-hr averaged measurements. It would be desirable to base the standard on a shorter averaging time to better represent human perception of visibility. This could be accomplished by either an estimation of extinction from semicontinuous particulate matter (PM) data or direct measurement of scattering and absorption. To this end we have compared 1-hr measurements of fine plus coarse particulate scattering using a nephelometer, along with an estimate of absorption from aethalometer measurements. The study took place in Lindon, UT, during February and March 2012. The nephelometer measurements were corrected for coarse particle scattering and compared to the Filter Dynamic Measurement System (FDMS) tapered element oscillating microbalance monitor (TEOM) PM_2.5 measurements. The two measurements agreed with a mass scattering coefficient of 3.3 ± 0.3 m²/g at relative humidity below 80%. However, at higher humidity, the nephelometer gave higher scattering results due to water absorbed by ammonium nitrate and ammonium sulfate in the particles. This particle-associated water is not measured by the FDMS TEOM. The FDMS TEOM data could be corrected for this difference using appropriate IMPROVE protocols if the particle composition is known. However, a better approach may be to use a particle measurement system that allows for semicontinuous measurements but also measures particle bound water. Data are presented from a 2003 study in Rubidoux, CA, showing how this could be accomplished using a Grimm model 1100 aerosol spectrometer or comparable instrument.

Implications: Visibility is currently based on 24-hr averaged PM mass and composition. A metric that captures diurnal changes would better represent human perception. Furthermore, if the PM measurement included aerosol bound water, this would negate the need to know particulate composition and relative humidity (RH), which is currently used to estimate visibility. Methods are outlined that could accomplish both of these objectives based on use of a PM monitor that includes aerosol-bound water. It is recommended that these techniques, coupled with appropriate measurements of light scattering and absorption by aerosols, be evaluated for potential use in the visibility based secondary standard.     

Effects of aerosol species on atmospheric visibility in Kaohsiung City, Taiwan

Visibility data collected from Kaohsiung City, Taiwan, for the past two decades indicated that the air pollutants have significantly degraded visibility in recent years. During our study period, the seasonal mean visibilities in spring, summer, fall, and winter were only 5.4, 9.1, 8.2, and 3.4 km, respectively. To ascertain how urban aerosols influence the visibility, we conducted concurrent visibility monitoring and aerosol sampling in 1999 to identify the principal causes of visibility impairments in the region. In this study, ambient aerosols were sampled and analyzed for 11 constituents, including water-soluble ions and carbon materials, to investigate the chemical composition of Kaohsiung aerosols. Stepwise regression method was used to correlate the impact of aerosol species on visibility impairments. Both seasonal and diurnal variation patterns were found from the monitoring of visibility. Our results showed that light scattering was attributed primarily to aerosols with sizes that range from 0.26 to 0.90 pm, corresponding with the wavelength region of visible light, which accounted for approximately 72% of the light scattering coefficient. Sulfate was a dominant component that affected both the light scattering coefficient and the visibility in the region. On average, (NH4)2SO4, NH4NO3, total carbon, and fine particulate matter (PM2.5)-remainder contributed 53%, 17%, 16%, and 14% to total light scattering, respectively. An empirical regression model of visibility based on sulfate, elemental carbon, and humidity was developed, and the comparison indicated that visibility in an urban area could be properly simulated by the equation derived herein.     

Precipitation in Light Extinction Reconstruction

Abstract

The U.S. Environmental Protection Agency (EPA) published the Regional Haze Rule (RHR) in 1999.¹ The RHR default goal is to reduce haze linearly to natural background in 2064 from the baseline period of 2000–2004. The EPA default method^2,3 for estimating natural and baseline visibility uses the Interagency Monitoring of Protected Visual Environments (IMPROVE) formula. The IMPROVE formula predicts the light extinction coefficient from aerosol chemical concentrations measured by the IMPROVE network. The IMPROVE light scattering coefficient formula using data from 1994–2002 underestimated the measured light scattering coefficient by 700 Mm^?1, on average, on days with precipitation. Also, precipitation occurred as often on the clearest as haziest days. This led to estimating the light extinction coefficient of precipitation, averaged over all days, as the light scattering coefficient on days with precipitation (700 Mm^?1) multiplied by the percent of precipitation days in a year. This estimate added to the IMPROVE formula light extinction estimate gives a real world estimate of visibility for the 20% clearest, 20% haziest, and all days. For example, in 1993, the EPAs Report to Congress projected visibility in Class I areas would improve by 3 deciviews by 2010 across the haziest portions of the eastern United States because of the 1990 Clean Air Act Amendments. Omitted was the light extinction coefficient of precipitation. Adding in the estimated light extinction coefficient of precipitation, the estimated visibility improvement declines to <1 deci-view.     

Influence of ship emissions on air quality and input of contaminants in southern Alaska National Parks and Wilderness Areas during the 2006 tourist season

The impact of ship emissions on air quality in Alaska National Parks and Wilderness Areas was investigated using the Weather Research and Forecasting model inline coupled with chemistry (WRF/Chem). The visibility and deposition of atmospheric contaminants was analyzed for the length of the 2006 tourist season. WRF/Chem reproduced the meteorological situation well. It seems to have captured the temporal behavior of aerosol concentrations when compared with the few data available. Air quality follows certain predetermined patterns associated with local meteorological conditions and ship emissions. Ship emissions have maximum impacts in Prince William Sound where topography and decaying lows trap pollutants. Along sea-lanes and adjacent coastal areas, NO_x, SO₂, O₃, PAN, HNO₃, and PM_2.5 increase up to 650 pptv, 325 pptv, 900 pptv, 18 pptv, 10 pptv, and 100 ng m^?3. Some of these increases are significant (95% confidence). Enhanced particulate matter concentrations from ship emissions reduce visibility up to 30% in Prince William Sound and 5–25% along sea-lanes.     

Measurements of Sulfur Oxides,Nitrogen Oxides,Haze and Fine Particles at a Rural Site on the Atlantic Coast

During August, 1982 and January and February, 1983, General Motors Research Laboratories operated air monitoring sites on the Atlantic Coast near Lewes, Delaware and 1250 km to the east on the southwest coast of Bermuda. The overall purpose of this project was to study the transformations of the principal acid precipitation precursors, NO _x and SO _x species, as they transport under conditions not complicated by emissions from local sources. In this paper, the measurements of gas and particulate species from Lewes are described and the composition and sources of sulfate aerosol, which is the most important haze-producing species, are investigated.

On the average, the total suspended particulate (TSP) concentration was 27.9 μg/m³ while the PM₁₀ (mass of particles with a diameter less than or equal to 10 μm) concentration was 22.0 μg/m³ or 79 percent of the TSP. The PM₁₀ consisted of 6.1 μg/m³ of coarse particles (CPM, diameter = 2.5 ? 10μm) and 15.9 μg/m³ of fine particles (FPM, diameter < 2.5 μm).

On a mass basis the most important constituents of the fine particulate fraction were sulfate compounds, 50 percent, and organic compounds, 30 percent. The mean light extinction coefficient corresponds to a visual range of 18-20 km. Most of the extinction can be attributed to the sulfate (60 percent) and organic carbon (13 percent). Particle size measurements show that the mass median aerodynamic diameter for both species is 0.43 μm. This is a typical size for a hydrated sulfate aerosol. For carbon, however, this is a larger size than previously reported and results in a more efficient light scattering aerosol. Principal component analyses indicate that coal combustion emissions from the midwestern U.S. are the most significant source of sulfate in Lewes during the summer and winter.     

Assessing the Relative Contribution of Biogenic and Fossil Processes to Visibility-Scattering Aerosols Found in Remote Areas: Near-Term Organic Research Program

Abstract

Organic carbon has been found to be a significant component of aerosols that impair visibility in remote areas across the country. Organic aerosols are particularly important in western areas of the United States and contribute roughly equally with sulfate aerosols and dust in the total extinction budget. Potential visibility enhancement resulting from various future energy management options that reduce volatile organic carbon and particulate material emissions from fossil-energy-related processes hinges on the relative contribution of the fossil-fuel-derived organic component to the extinction budget. Thus, additional studies are needed to quantify the partitioning of organic carbon between biogenic and fossil sources. Relative humidity (RH) also plays an important role in visibility impairment. It is well known that water soluble aerosol species, such as sulfate and nitrate, can increase light-scattering efficiencies of fine particles by more than an order of magnitude as RH is increased from 20-30% to 90-95%. Organic carbon aerosol has been found to be a mixture of more soluble and less soluble components, but few studies have been performed to evaluate the RH response function of aerosols composed of these components, either separately or in combination, especially at high relative humidities. The purpose of this paper is to describe some experiments that could address the major uncertainties of biogenic and fossil carbon contributions to the fine particle extinction budget and visibility impairment.     

Aerosol size distributions and visibility estimates during the Big Bend regional aerosol and visibility observational (BRAVO) study

The Big Bend Regional Aerosol and Visibility Observational (BRAVO) study was conducted in Big Bend National Park in 1999. The park is located in a remote region of southwest Texas but has some of the poorest visibility of any Class 1 monitored area in the western US. The park is frequently influenced by air masses carrying emissions from Mexico and eastern Texas. Continuous physical, optical and chemical aerosol measurements were performed in an effort to understand the sources of and contributions to haze in the park. As part of this characterization, dry aerosol size distributions were measured over the size range of 0.05<D_p<20 μm. Three instruments with different measurement techniques were used to cover this range. Complete size distributions were obtained from all of the instruments in terms of a common measure of geometric size using a new technique. Size parameters for accumulation and coarse particle modes were computed and demonstrate periods when coarse mode volume concentrations were significant, especially during suspected Saharan dust episodes in July and August. Study average (and one standard deviation) geometric volume mean diameters for the accumulation and coarse particle modes were 0.26±0.04 and 3.4±0.8 μm, respectively. Dry light scattering coefficients (b_sp) were computed using measured size distributions and demonstrated periods when contributions to b_sp from coarse particles were significant. The study average computed b_sp was 0.026±0.016 km⁻¹. Computed dry b_sp values were highly correlated with measured values (r²=0.97). Real-time sulfate measurements were correlated with accumulation mode volume concentrations (r²=0.89) and computed dry light scattering coefficients (r²=0.86), suggesting sulfate aerosols were the dominant contributor to visibility degradation in the park.     

Study of the Relationship of Distant SO2 Emissions to Dallas-Fort Worth Winter Haze

ABSTRACT

A study was conducted to estimate the changes in wintertime visual air quality in Dallas-Fort Worth (DFW) that might occur due to proposed reductions in SO₂ emissions at two steam electric generating plants in eastern Texas, each over 100 km from the city. To provide information for designing subsequent investigations, the haze was characterized broadly during the first year of the study. Meteorological data acquired then demonstrated that, during haze episodes, emissions from only one of the two plants were likely to be transported directly to DFW. Therefore, the second year of the study was centered on just one of the power plants. Air quality was then characterized within the urban area and at rural locations that would be upwind and downwind of the plant during transport to DFW. An instrumented aircraft measured plume dispersion and the air surrounding the plume on selected days. A mathematical model was used to predict the change that would occur in airborne particulate matter concentrations in DFW if SO₂ emissions were reduced to reflect the proposed limitations. The contribution of particles in the atmosphere to light extinction was estimated, and simulated photographs were produced to illustrate the visibility changes. The study concluded that the proposed emission reductions would, at most, subtly change perceived wintertime visibility.     

Aluminium production as a source of atmospheric carbonyl sulfide (COS)

Carbonyl sulfide is found as a major sulfur compound in anodic gases of commercial aluminium electrolysis. Recent spectroscopic measurements on industrial aluminium smelters found typical CO/COS ratios between 80 and 200. This results in specific COS emissions of between 1 and 7 kg/t(Al) if all COS is released into the atmosphere. In 1993 aluminium production would have been responsible for between 0.02 and 0.14 Tg of COS emissions. Currently, aluminium production does not seem to influence the total atmospheric COS budget to an extent beyond its natural variability. If recent growth rates of global aluminium production are sustained, however, COS emissions would quadruple until 2030. Together with increasing aircraft emissions into the stratosphere, an increase of the sulfate background aerosol is to be expected that could significantly enhance ozone depletion. The use of inert anodes is recommended to reduce aluminium production emissions of COS and CF₄, C₂F₆, CO₂, and CO at the same time.     

The Effect of Fuel Composition on Atmospheric Aerosol Due to Auto Exhaust

Aerosols attributable to automobile exhaust can be classified as two types—primary aerosol (initially present in the exhaust) and secondary aerosol (generated photochemically from hydrocarbons and nitrogen oxides in the exhaust). In this study, investigation was made of possible effects of motor-fuel composition on the formation of these aerosols. Secondary aerosol, of principal interest in this work, was produced by irradiating auto exhaust in Battelle-Columbus’ 610 ft³ environmental chamber. A limited number of determinations of primary aerosol in diluted auto exhaust was made at the exit of a 36 ft dilution runnel. Determination of both primary and secondary aerosol was based on light-scattering measurements.

Exhaust was generated with seven full-boiling motor gasolines, both leaded and nonleaded, in a 1967 Chevrolet which was not equipped with exhaust-emission control devices. Changes in fuel composition produced a maximum factor of three difference in light scattering due to primary aerosol. Aerosol yields, for consecutive driving cycles on the same fuel, vary considerably; as a result, ranking the fuels on the basis of average primary aerosol yield was not very meaningful. In addition to fuel composition, the more important independent variables are initial SO₂ concentration, relative humidity and initial hydrocarbon concentration. Statistical analysis of the data indicates that the seven test fuels can be divided into two arbitrary groups with regard to secondary aerosol-forming potential. The fuels in the lower light-scattering group had aromatic contents of 15 and 21%, while those in the higher light-scattering group had aromatic contents of 25, 48, and 55%. Although the fuels can be grouped on the basis of a compositional factor, the grouping of fuels with aromatic content ranging from 25 to 55% indicates that this compositional factor cannot be equated simply with aromatic content. In an associated study of the aerosol-forming potential of individual hydrocarbons prominent in auto exhaust, it was observed that aromatics produce substantially more photochemical aerosol than olefins and paraffins. However, experiments with binar/hydrocarbon mixtures containing aromatjcs, as well as in these exhaust experiments, a strong dependence of aerosol yield on the aromatic components is is not observed. Thus, the data indicate that the dependence of secondary aerosol formation on fuel factors is a complex one and cannot be predicted solely on the basis of a sirigle hydrocarbon component reactivity scale.

The two types of automobile aerosol did not have the same dependence on fuel, composition. The variation in total light scattering attributable to primary plus secondary aerosol was less than that due to either component alone. It therefore was concluded that the light scattering due to automobile exhaust emissions in these experiments was not significantly affected by changing fuel composition.     

Light Scattering Characteristics of Aerosols at Ambient and as a Function of Relative Humidity: Part II—A Comparison of Measured Scattering and Aerosol Concentrations Using Statistical Models

ABSTRACT

The eastern United States national parks experience some of the worst visibility conditions in the nation. To study these conditions, the Southeastern Aerosol and Visibility Study (SEAVS) was undertaken to characterize the size-dependent composition, thermodynamic properties, and optical characteristics of the ambient atmospheric particles. It is a cooperative three-year study that is sponsored by the National Park Service and the Electric Power Research Institute and its member utilities. The field portion of the study was carried out from July 15 to August 25, 1995.

The study design, instrumental configuration, and estimation of aerosol types from particle measurements is presented in a companion paper. In the companion paper, we compare measurements of scattering at ambient conditions and as functions of relative humidity to theoretical predictions of scattering. In this paper, we make similar comparisons, but using statistical techniques. Statistically derived specific scattering associated with sulfates suggest that a reasonable estimate of sulfate scattering can be arrived at by assuming nominal dry specific scattering and treating the aerosols as an external mixture with ammoniation of sulfate accounted for and by the use of Tang's growth curves to predict water absorption. However, the regressions suggest that the sulfate scattering may be underestimated by about 10%. Regression coefficients on organics, to within the statistical uncertainty of the model, suggest that a reasonable estimate of organic scattering is about 4.0 m²/g.

A new analysis technique is presented, which does not rely on comparing measured to model estimates of scattering to evoke an understanding of ambient aerosol growth properties, but rather relies on measurements of scattering as a function of relative humidity to develop actual estimates of f(RH) curves. The estimates of the study average f(RH) curve for sulfates compares favorably with the theoretical f(RH) curve for ammonium bisulfate, which is in turn consistent with the study average sulfate am-moniation corresponding to a molar ratio of NH₄/SO₄ of approximately one. The f(RH) curve for organics is not significantly different from one, suggesting that organics are weakly to nonhygroscopic.