Atmospheric Environment (v.39, #17)

Review of atmospheric metallic elements in Asia during 2000–2004 by Guor-Cheng Fang; Yuh-Shen Wu; Shih-Han Huang; Jui-Yeh Rau (3003-3013).
Metallic element transfer through the atmosphere is a significant part of the biogeochemical cycle of these elements. Natural and anthropogenic were two processes which can increase heavy metal concentrations in the atmosphere. Atmospheric particulates, especially secondary anthropogenic fine particles (PM2.5), have been influence human health. Generally speaking, the total daily mortality increases by approximately 1% for every 10 μg m−3 increase in PM10 concentration (Lippmann, 1998). This is why the PM10 and PM2.5 measurements are included in the US ambient air quality standards (US-EPA, 1987 for PM10; 1996 for PM2.5) (Querol et al., 2001).In recent years, since the great efforts made by Taiwan government towards the reduction of O3 and PM10 concentrations by controlling the emission rates of local pollutants sources, the frequency of exceeded PSI has gradually decrease the value of 4.9% in 1999 (Taiwan EPA, 2000).Urban populations are exposed to metals in suspended particles and these are often well above natural background levels owing to anthropogenic processes (Espinosa et al., 2002). This results in elevated metal concentrations that can pose an important risk to human health.Understanding emissions from traffic includes identification of the sources, which is also crucial for designing control measures. Road traffic involves numerous potential sources of metals, combustion products from fuel and oil, wear products from tires, brake linings, bearings, coach and road construction materials, and re-suspension of soil and road dust.The different sample collection devices, pretreatment and analysis methods were discussed in this study. The purpose of this study arranges the atmospheric metallic elements investigations in Asia regions. The data obtained here can also help to understand the sources, concentration, phase distribution and health impact of atmospheric metallic elements in Asian countries.
Keywords: Metallic element; Review; Metal; Asia;

An integrated measurement system was developed using high-speed photography, image recognition and PIV/PTV techniques to simultaneously measure the velocities and size distributions of fine atmospheric aerosols. The photographic system included a high-power (25 W) continuous laser, a microscope and a high-speed CCD camera (30–10,000 fps) could capture successive particle images on a microsecond time scale and micron space scale. The software system integrates improved image recognition algorithms for circular particles, various image-processing algorithms and advanced PIV/PTV algorithms. The information is extracted from the images in three steps: image preprocessing, particle recognition from images to obtain the diameters and positions of the particles, and the velocity vector calculation using PIV/PTV techniques. The system was tested using a computer-generated image. The results show that the particle extraction algorithms are robust and accurate. The system was also used to effectively measure fine atmospheric particles in the smoke from burning incense. Some problems are identified for future improvements.
Keywords: Velocity and size measurement; Aerosols; PIV/PTV; Circle detection; Image processing;

Characterization of exhaust particulates from diesel engine by Mukesh Sharma; Avinash Kumar Agarwal; K.V.L. Bharathi (3023-3028).
This study was set out to characterize particulate emissions from a mid-sized diesel engine in terms of metals, benzene soluble fraction, elemental and organic carbon. For this study, the exhaust particulates from Mahindra direct injection transportation diesel engine (40 hp) were collected at four different engine operating conditions, namely idle, 40%, 70% and full load for the above parameters. It was found that as the load increased from idling to full load, the metal content in particulates gradually decreased. The metal content in exhaust particulates was correlated with that of metal content in diesel. Benzene soluble fraction, which is a marker for carcinogenicity, also showed decrease in its level with increasing load. It was found that at idling, 67% of mass was soluble in benzene while at full load it was only about 24%. Elemental carbon content increased with increase in load showing a maximum of 48% at 70% load and a minimum of 25% at idling. The trend of organic carbon was quite opposite; it showed 35% organic carbon at idling, which decreased gradually with increase in load. This research has suggested that composition of diesel exhaust particles varies significantly with varying engine load conditions.
Keywords: Diesel particulate matter; Metals; Benzene soluble fraction; Elemental carbon; Organic carbon;

Export of atmospheric mercury from Asia by Daniel Jaffe; Eric Prestbo; Phil Swartzendruber; Peter Weiss-Penzias; Shungo Kato; Akinori Takami; Shiro Hatakeyama; Yoshizumi Kajii (3029-3038).
Emissions inventories indicate that Asian Hg sources are more than 50% of the global anthropogenic total. However, few measurements have been made to verify these large emissions. In this paper, we report on measurements of mercury from two sites during spring 2004 which received Asian outflow: Hedo Station, Okinawa (HSO), Japan and the Mt. Bachelor Observatory (MBO) in central Oregon, USA. At both sites, export of atmospheric mercury from China and East Asia was verified by the observations and meteorological data. The mean Hg0 concentration we observed at HSO was 2.04 ng m−3, which is significantly higher than the global background. Measurements of reactive gaseous mercury and particulate mercury at HSO found relatively little outflow of these compounds from Asia, generally less than 3% of the total mercury. By examining the correlation of Hg0 to carbon monoxide during outflow events, we have derived an enhancement ratio, which should reflect the ratio of these compounds in the source region. During outflow from Asia, the mean Hg0/CO molar enhancement ratio we observed at HSO was 6.2×10−7, which is nearly twice the expected ratio based on emissions estimates from China. During one episode of long-range transport from Asia to the MBO, we found a very similar ratio of Hg0/CO to that observed at HSO and a similar ratio was also reported by Friedli et al. (2004, Journal of Geophysical Research 109, D19 S25) downwind of Shanghai. Thus the ratio of Hg0/CO appears to be a good tracer of Asian industrial outflow. Using the Hg0/CO ratio and a recent inventory for CO emissions, we calculate Hg0 emissions from Asia of 1460 metric tons year−1, which is nearly two times the value in the Pacyna et al. (2003, Global mercury emissions. Presented at Long-Range Transport Workshop, Ann Arbor, MI, September 2003) inventory. Several hypotheses are proposed to explain this discrepancy, including an underestimation of the Chinese industrial Hg0 emissions, natural sources, significant re-emissions of previously deposited Hg and/or a higher ratio of Hg0/total Hg in the outflow from Asia than in the emission inventory.
Keywords: Mercury; Asia; Emissions; Carbon monoxide;

Plume dispersion from the MVP field experiment. Analysis of surface concentration and its fluctuations by Yimin Ma; Zafer Boybeyi; Steven Hanna; Kittisak Chayantrakom (3039-3054).
Surface concentration and its fluctuations from plume dispersion under unstable conditions in a coastal environment are investigated using the model validation program field experimental data. The goal of this study is to better understand plume dispersion under such conditions. Procedures are described to derive the plume surface concentration from moving vehicle measurements. Convective boundary layer scalings are applied and cumulative density functions (CDF) are studied.The results indicate that the relative concentration fluctuation intensity ( σ c / C ( y ) ) decreases with the normalized downwind distance ( X ) and that it is relatively small at the plume central line and largely increased at the plume edges, consistent with other field and laboratory results. The relation between σ c / C ( y ) at the plume centerline ( σ c / C ) and X for elevated sources can be described by σ c / C = a + b / X . The crosswind plume spread ( σ y ) is found to satisfy Deardorff and Willis's (J. Appl. Meteorol., 14 (1975) 1451) form of σ y / h = a 1 X / ( 1 + a 2 X ) 1 / 2 scaled with convective layer depth h . For elevated sources, the normalized crosswind integrated concentration ( C y ) is found to satisfy a relation of C y = 16 X - 3 / 2 , with Yaglom's (Izr. Atmos. Oceanic Phys., 8 (1972) 333) scaling rule on the free convective layer being applied.Empirical CDFs based on the gamma and the clipped probability density functions show agreements with the experimental CDFs, with the former being better than the latter when ( c - C ) / σ c > 0.5 . A new clipped-gamma CDF form is proposed based on the analysis of the present data, showing a better agreement.We suggest that a parameter u 0 * ( 12 - 0.5 h / L ) 1 / 3 , with combined efforts of surface friction velocity ( u 0 * ), Monin–Obukhov stability length ( L ) and unstable boundary layer height ( h ), replace the convective velocity scale ( w * ) under weak convective conditions in a coastal environment.
Keywords: Concentration fluctuation; CDF; Intermittency; Crosswind plume spread;

We analyse the air quality data measured at a green area of Buenos Aires City (Argentina) during 38 days in winter. We study the relationships between ambient concentrations of nitric oxide (NO), nitrogen dioxide (NO2), ozone (O3) and nitrogen oxides (NO x =NO+NO2). The variation of the level of oxidant (OX=O3+NO2) with the NO x is obtained. It can be seen that the level of OX at a given location is made up of two contributions: one independent and another dependent on the NO x concentration. The first one can be considered as a regional contribution, equivalent to the background O3 concentration and the second one as a local contribution that depends on the level of primary pollution. Local oxidant sources may include direct NO2 emissions, the reaction of NO with O2 at high-NO x levels, and the emission of species that promote the conversion of NO to NO2. The final category of emissions may include the nitrous acid (HONO) that is emitted directly in vehicle exhaust. Finally, we present a diurnal variation of the local as well as regional contributions and the dependence of the last one on wind direction.
Keywords: Urban pollution; Nitrogen oxides; Ozone; Oxidant; Regional pollution; Local pollution;

Mass distribution and concentrations of negative chemiions in the exhaust of a jet engine: Sulfuric acid concentrations and observation of particle growth by Thomas M. Miller; John O. Ballenthin; A.A. Viggiano; Bruce E. Anderson; Chowen C. Wey (3069-3079).
Measurements of negative-ion composition and density have been made in the exhaust of a J85-GE-5H turbojet, at ground level, as part of the NASA-EXCAVATE campaign. The mass spectrometer was placed 3 m from the exhaust plane of the engine. Measurements were done as a function of engine power in six steps from idle (50%) to military power (100%). Since the exhaust velocity changes with power, this also corresponds to a time evolution for ion growth. At 100% power most of the ions are HSO4 with minor amounts of HSO4 (H2O) n . With decreasing engine power the degree of hydration increases. In addition, ions with a 139-amu core dominate the spectra at lower engine power. The chemical identity of this ion is unknown. Observation of a small amount of NO3 core ions in the high-power spectra allows the determination of H2SO4 concentrations, which turn out to be a fraction-of-a-percent of the total sulfur in the fuel. Combining the present data with several previous composition measurements allows one to observe ion evolution from bare ions to ions with masses >8000 amu. Ion densities are derived and appear consistent with previous measurements used in modeling studies indicating that ion nucleation is a probable mechanism for volatile aerosol formation.
Keywords: Jet engine; Ion-induced nucleation; Aerosol growth; Sulfuric acid; Mass spectrometer;

The potential for ozone to cause an artifact during sampling of fine particulate mercury was investigated in southwest Detroit in July 2001 and July 2002. During the July 2001 sampling period, the use of KI and KCl denuders placed upstream to remove ozone and reactive gaseous mercury, Hg2+ (g), resulted in more particulate mercury (maximum 33 pg m−3) on a denuded filter when compared to an undenuded one. Additionally, a KCl denuded quartz filter also showed more particulate mercury (maximum 12 pg m−3) compared to an undenuded filter when ozone concentrations were most elevated. Further sampling in July 2002 used a repeated sampling design and resulted in more particulate mercury on undenuded filters when compared to denuded ones. Conditions in July 2001 were characterized by elevated concentrations of ozone during which there were three ozone action days declared by the State of Michigan whereas ozone concentrations in July 2002 were 22% lower. One possible explanation for the observations may be that elemental mercury, Hg0 (g), is oxidized to reactive gaseous mercury, Hg2+ (g), during sampling leading to an artifact. The potential for this type of artifact may be greater during periods of high photochemical activity and this must be considered when sampling during the warmer seasons when high levels of oxidants are likely to be present in an urban atmosphere.
Keywords: Atmospheric mercury; Annular denuders; Measurement challenges; Urban aerosol;

Long-term simulations of photo oxidant pollution over Portugal using the CHIMERE model by A. Monteiro; R. Vautard; C. Borrego; A.I. Miranda (3089-3101).
This work examines the performance of the CHIMERE photochemical model in simulating ozone and nitrogen dioxide in Portugal over a long-term summer period. The analysis focuses on comparisons against the available measurements during the 2001 summer season. The meteorological forcing of the model is given by the European Centre for Medium-Range Weather Forecasts (ECMWF), and the emission inventory used was obtained with a top-down methodology updated for the simulation year.Despite the coarse resolution of the meteorological data, the complex topography and coastal location of Portugal, the results obtained show that the modelling system is able to reproduce the nitrogen dioxide and ozone episodes that occurred during the simulated summer period. Mean error and correlation improve when considering the sum of photo oxidant instead of individual pollutants, indicating that a significant part of the model error is due to either the lack of representativeness of monitoring stations or to inaccuracies in the emission inventory.
Keywords: Regional; Modelling; Validation; Ozone; Nitrogen dioxide;

Surface layer turbulence parameters over a complex terrain by Osvaldo L.L. Moraes; Otávio C. Acevedo; Gervásio A. Degrazia; Domenico Anfossi; Rodrigo da Silva; Vagner Anabor (3103-3112).
The structure of the surface boundary layer over homogeneous conditions is nowadays well understood. However, there is no sufficient knowledge of the corresponding structure over complex topography. Another major issue in micrometeorological studies is the description of the turbulence with low wind speeds. Considering that almost all concentration of population and industry occurs in regions of complex terrain and that low wind speeds are the most critical ones to the dispersion of pollutants, additional studies are still essential in helping the understanding of the structure of PBL over such kind of terrain and such type of stability. This is the purpose of this paper in which a data set collected in a valley, using a sonic anemometer, is presented and discussed. Results reveal that the turbulence parameters in the convective and stable surface layer follow similarity relationships only if special conditions regarding wind direction and wind speed exist in relation to the surrounding terrain. In other words when the air masses travel for longer distances over a flat terrain.
Keywords: Turbulence parameters; Similarity theory; Complex terrain;

Source estimation of anthropogenic aerosols collected by a DRUM sampler during spring of 2002 at Gosan, Korea by J.S. Han; K.J. Moon; S.Y. Ryu; Y.J. Kim; Kevin D. Perry (3113-3125).
Continuous measurements of trace elements in size-segregated aerosols were performed by X-ray fluorescence analysis in order to assess the transport of natural and anthropogenic aerosol sources to Gosan site in springtime 2002. Various aerosol elemental components (such as Al, Ca, Si, K, Ti, and Mn) had low enrichment factors close to one and exhibited a similar temporal variation in coarse particle range (2.5–12 μm) indicating the dominant influence of soil dust. On the other hand, in the fine particle size range (0.09–2.5 μm), several anthropogenic elements in ambient particles (such as As, Se, Zn, Pb, and Cu) revealed high enrichment (100<EFs<10,000), and similar temporal variation was observed between these elements in partial size range implying that these elements could be emitted from the same anthropogenic sources.Factor analysis was conducted with the size- and time-resolved aerosol composition data in order to identify anthropogenic sources of aerosols during non-Asian dust (NAD) period. As a result, it was found that the coarse particles were mainly influenced by soil, seasalt, coal combustion, and metallurgical emission and fine particles were significantly attributed to several anthropogenic sources, such as oil combustion, nonferrous metal source, and municipal incineration. Especially, submicron particles (0.09–0.56 μm) were further impacted by motor vehicles and field burning emissions in addition to those sources.
Keywords: DRUM impactor; Aerosol size distribution; Trace element; Factor analysis; Source estimation;

Source apportionment of fine and coarse particulate matter in a sub-urban area at the Western European Coast by S.M. Almeida; C.A. Pio; M.C. Freitas; M.A. Reis; M.A. Trancoso (3127-3138).
Aerosol chemical composition data for PM2.5 and PM10–2.5, was acquired during the year 2001 at a sub-urban area located in the north outskirts of Lisbon. Principal component analysis (PCA) and multilinear regression analysis (MLRA) were used to identify possible sources of particulate matter (PM) and to determine their mass contribution. Seven main groups of sources were identified: soil, sea, secondary aerosols, road traffic, fuel-oil combustion, coal combustion and a Se/Hg emission source. In PM2.5, secondary aerosol and vehicle exhaust contributed on average, with 25% and 22% to total mass, respectively, while sea spray and soil represented, respectively, 47% and 20% of the coarse fraction mass loading. Maritime air mass transport has a significant role on air quality in the North of Lisbon. Maritime transport scenarios are very frequent and promote the decrease of anthropogenic and mineral aerosol concentrations. The highest PM levels were recorded during South Continental episodes. These episodes are characterized by high mineral aerosol contents, due to the transport of dust from the interior of Iberian Peninsula and the Sahara desert.After identifying the contribution of PM sources, it is possible to conclude that abatement strategies to improve local air quality should focus on traffic and on non-mobile combustion processes emitting sulphur and NO x , which conduce to the formation of secondary aerosols.
Keywords: Aerosol particles; Source apportionment; Receptor modelling; Sub-urban air; South western Europe;

Wet and dry deposition fluxes of mercury in Japan by Masahiro Sakata; Kohji Marumoto (3139-3146).
Wet and dry deposition fluxes of mercury were measured concurrently from December 2002 to November 2003 at 10 sites in Japan (nine sites for dry deposition measurements). Precipitation and dry deposition (gases and particles) were collected every half-month using an automatic wet and dry deposition sampler. This sampler is composed of a precipitation sampler and a water surface sampler for dry deposition. The results showed that wet deposition fluxes of mercury increase generally during summer when the precipitation amount increases. In contrast, the seasonal variations of the dry deposition fluxes were very small compared to those of the wet deposition fluxes. It was found that wet deposition plays a dominant role in mercury deposition in Japan, contributing fluxes ranging from 0.99 to 2.2 times (1.6 times on average at nine sites) those of the dry deposition fluxes. There was a high correlation between the annual wet deposition flux of mercury and annual precipitation amount ( r 2 = 0.73 , P < 0.01 ). Thus, about 70% of the variance of mercury wet deposition in Japan is explained by precipitation amount. This is probably because mercury wet deposition is dominated by the scavenging of reactive gaseous mercury (RGM) by precipitation via oxidation of gaseous elemental mercury (Hg0) in the gas and aqueous phases. In addition, a similar relationship between the annual wet deposition flux and precipitation amount ( r 2 = 0.55 , P < 0.001 ) was observed at the Mercury Deposition Network (MDN) sites in North America, suggesting that about 55% of the variance of mercury wet deposition is due to precipitation amount.
Keywords: Particulate mercury; Precipitation amount; Reactive gaseous mercury; Scavenging by precipitation; Water surface sampler;

The influence of background wind flow on the dispersion of carbon monoxide (CO) was investigated over a one-month period in two street canyons of differing geometry in the city of York, UK. Electrochemical sensors were used to measure CO concentrations at various heights and locations along each canyon. Six ultrasonic anemometers were used to measure the airflow and turbulence within one of the canyons. A seventh anemometer measured the above-roof (or background) winds at approximately twice the street building height. Bi-directional traffic flow was measured in each street using the Split, Cycle and Offset Optimisation Technique (SCOOT) system. The investigation indicates that differences in the street geometry and the background winds result in contrasting in-canyon wind flows and dispersion characteristics. Whilst there is evidence of a single cross canyon vortex in one street canyon for certain background wind directions, there are also a number of three-dimensional (3D) flow structures formed in the canyon due to the influence of flow channelling, both within the canyon and from adjoining side streets. For background winds with a strong perpendicular component relative to the canyon axes, the mean CO concentrations on the leeward side of the street canyons were on average, a factor of two greater than on the windward side. This feature can be attributed to the formation of vertical and horizontal vortices. However, when the concentrations were normalised by the background wind speed and total traffic flow, a three-fold difference exists between the windward and leeward sides of the street canyon. Evidence of flow channelling and helical flow regimes was also detected for background flows with a component parallel to the canyon axes. The difference in mean CO concentrations between the two street canyons was approximately two-fold, with the highest mean concentrations in the narrower canyon, which has a smaller cross-sectional area.
Keywords: Pollutant dispersion; Vortex; Wind flow; Urban;

The influence of traffic characteristics on the dispersion of carbon monoxide (CO) was investigated over a period of one month in two street canyons of differing geometry in the city of York, UK. Electrochemical sensors were used to measure CO concentrations at several locations along the two street canyons. Bi-directional traffic characteristics were measured in both streets using inductive detector loops of the Split, Cycle and Offset Optimisation Technique (SCOOT). The SCOOT data allowed comparison of free, unstable and congested traffic conditions with the spatial and temporal variability of CO concentrations in both streets. During free flow traffic conditions a significant linear relationship was found to exist between traffic flow and CO concentration in each street. The relationship fails during congested conditions, as the mean CO concentration was found to be more dependent on detector occupancy. In general, the highest mean CO concentrations occurred during high-occupancy conditions, indicating the likely influence of stop–start conditions and the accelerating driving mode that is associated with higher emission rates. By considering the measured above roof level wind direction, it was shown that during perpendicular winds, the leeward CO concentrations were a factor of at least two greater than on the windward side of the canyon. Overall, the greatest mean CO concentrations occurred during congested conditions with perpendicular above-roof level winds. The significance of the work for both emissions and dispersion modelling of traffic-related pollutants within urban areas is discussed.
Keywords: Emissions; Dispersion; Congestion; Acceleration; Driving mode; Network capacity;

Laser-induced fluorescence instrument for measuring atmospheric SO2 by Yutaka Matsumi; Hiroyuki Shigemori; Kenshi Takahashi (3177-3185).
We report on the development of a high-sensitive detection system for measuring atmospheric SO2 using a laser-induced fluorescence (LIF) technique at around 220 nm. Second harmonics of a tunable broad-band optical parametric oscillator (OPO) pumped by the third harmonic of a Nd:YAG laser is used as a fluorescence excitation source. The laser wavelength is alternatively tuned to the peak and the bottom wavelengths in the photoabsorption spectrum of SO2 at 220.6 and 220.2 nm, respectively, and the difference signal at the two wavelengths is used to extract the SO2 concentration. This procedure can give a good selectivity for SO2 and avoid interferences of fluorescent or particulate species other than SO2 in the sample air. The SO2 instrument developed has a sensitivity of 5 pptv in 60 s and S / N = 2 . The practical performance of the detection system is tested in the suburban area. The inter-comparisons between the LIF instrument and a commercial instrument using Xe flash lamp excitation for the fluorescence detection have been performed. The correlation between two instruments is measured up to 70 ppbv. A good linear relationship between the LIF measurements and commercial instrument measurements is obtained.
Keywords: Sulfur dioxide; OPO; Detection limit; Instrument development;