Atmospheric Environment (v.34, #1)

This paper provides a critical examination of empirical emission factor models by considering one model that typifies the class: the model used to estimate PM10 emissions from paved roads. We show that the model can yield highly uncertain emission estimates because (1) it lacks a mechanistic basis, (2) its formulation is highly dependent on the data set used to derive it, and (3) the accuracy of the model is completely determined by the methods used to measure emissions. The paper also describes a method to relate model performance statistics to statements about the uncertainty in emission estimates.
Keywords: Emission factors; AP-42; PM10 emissions; Model uncertainty; Fugitive emissions; Emission measurement;

UV absorption cross sections of nitrous acid by A.S. Brust; K.H. Becker; J. Kleffmann; P. Wiesen (13-19).
The UV absorption cross sections of nitrous acid (HONO) and its deuterated analogue (DONO) have been measured in a slow flow system. Nitrous acid was produced by a high purity source reaching concentrations up to 22 ppm with a purity of >95% and a long-term variation of ⩽2%. The concentrations were measured by ion chromatography. Differential absorption cross sections for a spectral resolution of 0.5 nm of (2.98±0.26)×10−19  cm2 and (3.04±0.26)×10−19  cm2 have been obtained for the HONO absorption maxima at 354.2 and 368.1 nm, respectively. In addition, differential absorption cross sections of DONO were measured for the first time. The values of (2.67±0.24)×10−19  cm2 at 353.3 nm and (3.20±0.26)×10−19  cm2 at 367.6 nm are comparable to those of HONO. Absolute absorption cross sections were also derived for these species. The values of both the differential and absolute absorption cross sections determined in this work are lower than those reported previously. Therefore, the concentrations in the atmosphere may be 30–50% higher than that reported in several publications in which DOAS spectrometers were used for the determination of the HONO concentration.
Keywords: Nitrous acid; Differential absorption cross sections; Absolute absorption cross sections; Atmospheric chemistry;

Using UK data as a case study, this paper demonstrates that statistical models of hourly surface ozone concentrations require interactions and non-linear relationships between predictor variables in order to accurately capture the ozone behaviour. Comparisons between linear regression, regression tree and multilayer perceptron neural network models of hourly surface ozone concentrations quantify these effects. Although multilayer perceptron models are shown to more accurately capture the underlying relationship between both the meteorological and temporal predictor variables and hourly ozone concentrations, the regression tree models are seen to be more readily physically interpretable.
Keywords: Artificial neural networks; CART; Decision trees; Air quality modelling;

The formation yields of acrolein, 1,2-epoxy-3-butene and OH radicals have been measured from reaction of ozone with 1,3-butadiene at room temperature and atmosphere pressure. 1,3,5-Trimethyl benzene was added to scavenge OH radicals in measurements of product yields. In separate experiments, small quantities of 1,3,5-trimethyl benzene were added as a tracer for OH. Formation yields of acrolein of (52±7)%, 1,2-epoxy-3-butene of (3.1±0.5)% and OH radicals of (13±3)% were observed. In addition, the rate coefficient of the gas-phase reaction of ozone with 1,2-epoxy-3-butene was measured both directly and relative to propene, finding an average of (1.6±0.4)×10−18  cm3  molecule−1  s−1, respectively, at 296±2 K. The results are briefly discussed in terms of the effect of atmospheric processing on the toxicity of 1,3-butadiene.
Keywords: 1,2-Epoxy-3-butene; Ozone rate constant;

The clinical symptoms of patients and other evidences of a gas poisoning accident inside an industrial building strongly suggested an abrupt influx of engine exhaust from a construction vehicle which was operating outside in the open air. But the obviously high level of gas concentration could not be well explained by any conventional steady-state gas diffusion models. The author used an unsteady-state continuous Puff Model to simulate the time-wise changes in air stream with the pollutant gas being continuously emitted, and successfully reproduced the observed phenomena. The author demonstrates that this diffusion formula can be solved analytically by the use of error function as long as the change in wind velocity is stepwise, and clarifies the accurate differences between the unsteady- and steady-states and their convergence profiles. Also, the relationship between the Puff and Plume Models is discussed. The case study included a computational fluid dynamics (CFD) analysis to estimate the steady-state air stream and the gas concentration pattern in the affected area. It is well known that clear definition of the boundary conditions is key to successful CFD analysis. The author describes a two-step use of CFD: the first step to define the boundary conditions and the second to determine the steady-state air stream and the gas concentration pattern.
Keywords: Unsteady-state puff model; Plume model; CFD;

Concentrations of submicrometre particles from vehicle emissions near a major road by J Hitchins; L Morawska; R Wolff; D Gilbert (51-59).
As part of a program of study to assess the exposure risks related to particulate matter in the outdoor environment, number concentrations of particles from vehicle emissions were measured at increasing distances from a major road. Particles in the size range from 0.015 to 0.697 μm were measured with the scanning mobility particle sizer (SMPS) and in the size range from 0.5 to 20 μm, with the aerodynamic particle sizer (APS). In addition to number concentration measurements, an approximation of PM2.5 fraction was obtained using a DustTrak (simple photometer). The measurements conducted at distances from the road ranging from 15 to 375 m showed, that for conditions where the wind is blowing directly from the road, the concentration of fine and ultrafine particles decays to around half of the maximum (measured at the closest point to the road) at a distance of approximately 100–150 m from the road. For the wind blowing parallel to the road, the reduction to half of the concentration occurs at 50–100 m. There is no effect on total particle number concentration at a distance greater than 15 m from the road when the wind is blowing towards the road and away from the sampling points. Total number concentrations of larger particles measured were not significantly higher than the average values for the urban environment, and decrease with distance from the road, reaching about 60% at 150 m from the road for wind from the road. PM2.5 levels also decrease with distance to around 75% for wind from the road and to 65% for wind parallel to the road, at a distance of 375 m.
Keywords: Ultrafine particles; Fine particles; Particle number concentration; PM2.5; Horizontal profiles; Traffic; Vehicle emissions;

This large-eddy simulation study investigates the effects of an idealised surface inhomogeneity of sensible heat flux on dispersion of a passive plume emitted from elevated sources into an urban convective boundary layer (UCBL). The results show that when Deardorff's translation is made to introduce a wind with its direction aligned with the centreline of park blocks, dispersion of such a passive plume is strongly affected by turbulent structure associated with the patchy pattern of surface heat flux. In comparison with a case with a homogeneous surface, when the point source is located above the central line across the park areas, the mean plume height Z c is lower, the vertical dispersal parameter S z is smaller, the surface concentration C 0 is generally higher, and pollutant is less dispersed in the vertical direction. When the point source is aligned with the built-up area, however, the opposite situation occurs. The difference in surface concentration at a same downwind distance can be as large as 100% among the cases over a homogeneous surface, the location above built-up surface and the location above park surface in the idealised urban area. The extent to which the patchy pattern of surface heat flux influences the plume dispersion depends on two parameters: d/Z i and b/p, where d is the distance between the centres of two adjacent parks, b is the width of built-up area, p is the size of park, and Z i is the UCBL height. This study suggests that dispersion of a passive plume in a UCBL has different behaviour from that for a homogeneous surface in the previous studies. Estimate of pollutant concentration by the existing methods or models that are based on observations over a homogeneous surface cannot distinguish such differences revealed by the results in this paper.
Keywords: Large-eddy simulation; Plume dispersion; Convective boundary layer; Idealised patchy urban surface; Inhomogeneous heat flux;

Particle size distributions of polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/Fs) and polycyclic aromatic hydrocarbons (PAHs) were measured at a rural location in northeastern Bavaria during summer. The complete size range of the atmospheric particles was sampled using cascade impactors and a rotary impactor.On an average 92% of the hepta- and octachlorinated PCDD/Fs and 97% of the PAHs with five or more rings, which were all more or less completely particle bound in the atmosphere, were associated with particles with aerodynamic diameters (d ae)<2.9  μm. The substance loading decreased steadily with increasing particle size. On average only 3.4 and 1.2% of the Cl7–8DD/F and the PAHs with five or more rings, respectively, were found on large particles with d ae>8.6 μm. The temporal variation of the size distributions was low. Despite similar measured size distributions the washout ratios of the PCDD/Fs and PAHs calculated using simultaneously collected precipitation samples were different, with those of the PAHs lying consistently lower than those of the PCDD/Fs. This may have been due to differences in the particle size distributions which were not resolved with the impactors used.In the case of the semivolatile compounds for which significant fractions were present in the gas phase, 60–95% of the particle bound substance mass was always found on particles with d ae<2.9 μm. However, the concentrations did not decrease continuously with increasing particle size. Rather, increased shares were found in the particle size range 0.95–2.9 μm for the PAHs and 8.6–26 μm d ae for both the PCDD/Fs and the PAHs. It is proposed that the secondary maxima in the 8.6–26 μm size range are attributable to resuspended biogenic material that had accumulated the contaminants from the gas phase, whereas the high levels of the semivolatile PAHs in the 0.95–2.9 μm size range were due to combustion source characteristics.

Reduction of nitrate losses from filter and impactor samplers by means of concentration enrichment by Ming Chih Chang; Constantinos Sioutas; Seongheon Kim; Henry Gong; William S. Linn (85-98).
Sampling errors (artifacts) have greatly affected the precision of the quantitative analysis of volatile species, such as particulate ammonium nitrate. This work presents the effect of the enrichment in concentration of particulate nitrate in reducing volatilization losses in impactors and Teflon filter samplers. During the performance characterization of an ambient fine particle concentrator developed by Sioutas et al. (1995a, Environmental Health Perspectives 103, 172–177, 1995b, Inhalation Toxicology, 7, 633–644, 1977, Journal of Aerosol Science 28, 1057–1071) losses of ambient ammonium nitrate from denuded and undenuded Teflon filter samplers as well as the microorifice uniform deposit impactor (MOUDI) were evaluated in Los Angeles, CA, an area where ammonium nitrate constitutes a major component of ambient fine particulate matter. The field study data were compared to those predicted theoretically for a given set of gas and particulate nitrate concentrations, temperature and relative humidity. Both theoretical and experimental results indicated that the ratio of nitrate gas-to-particle concentration affects significantly the volatilization loss, with higher volatilization losses occurring at higher gas-to-particle concentration values. The concentration enrichment of particulate-phase nitrate resulted in reducing evaporation losses from the MOUDI from 20–50% to less than 10%. Losses of nitrate from denuded Teflon filters were reduced from 60–95% to less than 30%, and for undenuded Teflon filters from 30–80% to less than 5%. Our study concluded that nitrate losses from impactor, denuded and undenuded Teflon filter samplers could be virtually eliminated by placing the sampler downstream of a particle concentrator with a small cutpoint (i.e., 0.1 μm).
Keywords: Nitrate loss; Denuders; Filter samplers; Impactors; Ammonium nitrate; Particle concentrator;

The rate coefficient for the photolysis of nitrogen dioxide (J NO2 ) is used in ambient photochemical air quality models as this photodissociation leads to the formation of tropospheric ozone. This review examines the effectiveness and efficiency of four methods for the calculation of J NO2 . The nature of J NO2 is reviewed briefly with regard to temperature, pressure, zenith angle, altitude, albedo, clouds, atmospheric turbidity and atmospheric molecular absorption. Comparison of the four methods against a radiation transfer model, for three clear-sky days at different latitudes in Australia, shows that any of the methods may be used to approximate a clear-sky diurnal profile of J NO2 , as the maximum variation between all models at noon was found to be 15%. However, more care needs to be taken when calculating other photolysis rate coefficients, such as J O3 , as these are more susceptible to other variables such as the ozone column concentration. A simple procedure for calculating cloudy-day J NO2 is also illustrated.
Keywords: Actinic flux; Radiation; Nitrogen dioxide; Troposphere; Rate constant J NO2 ;

The anthropogenic contribution to isoprene concentrations in a rural atmosphere by Stefan Reimann; Pierluigi Calanca; Peter Hofer (109-115).
Atmospheric hydrocarbons are continuously monitored at the rural site of Taenikon, Switzerland. As expected for a rural area, highest isoprene concentrations are found in summer. However, elevated concentrations are also measured on some occasions in winter, in particular during events with long-lasting surface inversions, temperatures constantly below 0°C and snow covering the vegetation. During such events, concentrations of isoprene are strongly correlated with those of 1,3-butadiene, a substance that is mainly due to human activities. For these periods, a molar ratio between the concentrations of isoprene and those of 1,3-butadiene of 0.42 is observed. This value, together with the concentrations of 1,3-butadiene, is used to estimate the anthropogenic fraction of the atmospheric isoprene for the whole of 1997. It is found that the fraction is close to 100% in January–February and again in November–December. On the other hand, as early as March, a considerable amount of the observed isoprene appears to be of biogenic origin, although isoprene emissions by trees are negligible. The relative anthropogenic contribution is minimal in midsummer, when biogenic emissions are highest. For this time of the year, the anthropogenic contribution is largest during the early morning hours, in agreement with the traffic peak on nearby country roads.
Keywords: Non-methane hydrocarbons; Isoprene, 1,3-butadiene; Motor vehicle exhaust;

A comparative review of inorganic aerosol thermodynamic equilibrium modules: similarities, differences, and their likely causes by Yang Zhang; Christian Seigneur; John H Seinfeld; Mark Jacobson; Simon L Clegg; Francis S Binkowski (117-137).
A comprehensive comparison of five inorganic aerosol thermodynamic equilibrium modules, MARS-A, SEQUILIB, SCAPE2, EQUISOLV II, and AIM2, was conducted for a variety of atmospheric concentrations of particulate matter (PM) constituents, relative humidities (RHs), and temperatures. Our results show that although the PM compositions and concentrations predicted by these modules are generally comparable under most conditions, significant discrepancies exist under some conditions, especially at high nitrate/chloride concentrations and low/medium RHs. As a consequence, the absolute differences in total PM concentrations predicted by these modules under all simulation conditions are 7.7–12.3% on average and as much as 68% for specific cases. The PM predictions are highly sensitive to changes in the molar ratios of ammonium to sulfate, nitrate to sulfate, and sodium chloride to sulfate, relative humidity, and temperature. The similarities and differences in simulation results predicted by the five modules are analyzed and the likely causes for these differences are discussed in detail. Recommendations are provided regarding the relative advantages of these modules, possible improvements of their performance, and applications in three-dimensional PM modeling studies.
Keywords: Inorganic particulate matter; Thermodynamic modeling; Multi-phase equilibrium; Module comparison; Sensitivity;

Comparison of two particle-size spectrometers for ambient aerosol measurements by Th. Tuch; A. Mirme; E. Tamm; J. Heinrich; J. Heyder; P. Brand; Ch. Roth; H.E. Wichmann; J. Pekkanen; W.G. Kreyling (139-149).
There is an ongoing debate on the question which size fraction of particles in ambient air may be responsible for human health effects observed in epidemiological studies. Since there is no single instrument available for the measurement of the particle-size distribution over the full range of the fine fraction (diameter <2.5 μm) of the atmospheric aerosol, two instruments, the mobile aerosol spectrometer (MAS) and the electrical aerosol spectrometer (EAS), have been tested in a side-by-side comparison measuring ambient aerosol for a time period of six weeks in spring 1996 in the city of Erfurt, Germany. Furthermore, total particle number concentration measured by a condensation particle counter (CPC) and mass concentrations PM10 and PM2.5 were determined. Both spectrometers, MAS and EAS, are based on electrical mobility measurements for particles <0.1 μm and <0.5 μm, respectively, while MAS applies optical particle spectrometry and EAS applies again electrical mobility analysis for particles up to 2.5 and 10 μm, respectively. Both instruments proved to be reliable during this comparison providing data availability of >94%. To compare the spectral data, particle numbers were integrated within three size ranges: 0.01 – 0.1, 0.1 – 0.5, 0.5 – 2.5 μm. Hourly mean number concentrations of each size range observed during the six week comparison was: 2.6×104±19500 (2.48×104±1.79×104), 3.1×103±1.5×103 (4.1×103±2.0×103), 50±45 (1.9×102±1.2×102) cm−3 for MAS (EAS), respectively. Both aerosol spectrometers followed the variations of the ambient aerosol in a similar manner and yielded almost identical results for particle number concentrations of particles with diameters smaller than 0.5 μm. Furthermore, the total particle number concentration derived from MAS and EAS measurements (29000±20000; 29000±19000 cm−3) is well comparable with the number concentration derived from an integral counting CPC (31100±22000 cm−3). The results of this side-by-side comparison suggest that MAS and EAS together with PM2.5 measurements are suitable to reliably characterize size-distribution parameters of number and mass concentration of ambient aerosols.
Keywords: Aerosol spectrometers; Ambient aerosol; Particle number distribution;

A new technique based on the nebulization/reflux principle (Cofer chamber mist) has been used to sample atmospheric dimethylsulfoxide (DMSO). DMSO is reduced to DMS by sodium borohydride (NaBH4) and DMS is subsequently analyzed by gas chromatography. Recovery and reproducibility of the analysis are of the order of 90% and better than 10%, respectively. This analytical method has been also successfully applied to DMSO determinations in rainwater and aerosols.
Keywords: Atmosphere; DMSO; Rainwater; Aerosol; Atmospheric sampling; Analysis; Mist chamber;

With the aid of three atmospheric aerosol equilibrium models, we quantify the effect of metastable equilibrium states (efflorescence branch) in comparison to stable (deliquescence branch) on the partitioning of total nitrate between the gas and aerosol phases. On average, efflorescence branch concentrations of aerosol nitrate are 11% greater than those of the deliquescence branch at low aerosol nitrate concentrations (<8 μg m−3), whereas for higher aerosol nitrate concentrations (>8 μg m−3), deliquescence branch concentrations are 3% greater. In the low aerosol nitrate range, approximately 40% of the time deliquescence and efflorescence branch concentrations of aerosol nitrate have differences greater than 20% implicating the importance of considering both branches of aerosol behavior in this region. The largest differences between the two equilibrium states occur at several sets of conditions: at temperatures above 295 K and mid-range rh (60%), at mid-range temperatures (290–300 K) and low rh (<40%), and for sulfate-to-aerosol nitrate molar ratios of less than 0.5 and greater than 1 at low rh (<40%). In these two regions, average differences of 1–2 μg m−3 between deliquescence and efflorescence branch concentrations of aerosol nitrate are estimated. The potential existence of efflorescence branch aerosols in Southern California, where pollutant levels are high, appears to have a small effect on total nitrate partitioning. However, for areas characterized by moderate-to-low pollutant levels such as the Northeastern US, a significantly larger effect is predicted. The implications of these findings for modeling studies are discussed.
Keywords: Deliquescence; Efflorescence; Metastable equilibrium; Partitioning; Aerosol nitrate;