Atmospheric Environment (v.35, #5)

Neural networks and periodic components used in air quality forecasting by M Kolehmainen; H Martikainen; J Ruuskanen (815-825).
Forecasting of air quality parameters is one topic of air quality research today due to the health effects caused by airborne pollutants in urban areas. The work presented here aims at comparing two principally different neural network methods that have been considered as potential tools in that area and assessing them in relation to regression with periodic components. Self-organizing maps (SOM) represent a form of competitive learning in which a neural network learns the structure of the data. Multi-layer perceptrons (MLPs) have been shown to be able to learn complex relationships between input and output variables. In addition, the effect of removing periodic components is evaluated with respect to neural networks. The methods were evaluated using hourly time series of NO2 and basic meteorological variables collected in the city of Stockholm in 1994–1998. The estimated values for forecasting were calculated in three ways: using the periodic components alone, applying neural network methods to the residual values after removing the periodic components, and applying only neural networks to the original data. The results showed that the best forecast estimates can be achieved by directly applying a MLP network to the original data, and thus, that a combination of the periodic regression method and neural algorithms does not give any advantage over a direct application of neural algorithms.
Keywords: Nitrogen dioxide; Self-organizing maps; Multi-layer perceptron; Model comparison; Residual;

The present paper demonstrates the applicability of EPR spectrometry for separate estimation of soot (EC) and polycyclic aromatic hydrocarbons (PAH) in aerosols. The content of EC is obtained directly because of their paramagnetic properties whereas diamagnetic PAH, adsorbed on the soot, are converted to paramagnetic forms by oxidation over silica/alumina catalyst. In order to fulfill the goal of our study at this stage only few samples of aerosols are investigated after being collected at four different locations: near and distant to motorway, office and cafeteria. The obtained results show that the quantities of soot and PAH in all cases are μg m−3. However, their content varies depending on the place of sample collection. The following order of decreasing soot quantity is found: motorway>urban air>cafeteria>office whereas for PAH the order is cafeteria≅motorway>urban air>office. The obtained results are discussed in the light of the pollution sources at the sampling places.
Keywords: Aerosols; Elemental carbon; PAH; Determination; EPR;

A model for concentration fluctuation moments has previously been developed for a scalar dispersing in a turbulent flow (see Mole et al., 1997). This model assumes that the mean concentration is known as a function of space, and of time t, and that higher moments of concentration for a dispersing cloud can be determined fully by a further two parameters α(t) and β(t) (see Chatwin and Sullivan, 1990a). A closure is used which enables a coupled pair of first-order differential equations for α and β to be written down. Here attention is restricted to cases when the mean concentration is self-similar, with a spatial scale L(t). (It is assumed that L−1   dL/dt→0 as t→∞, so cloud growth must be slower than exponential.) It is shown that there is a constant α s , dependent on the spatial form of the mean concentration, such that αα s as t→∞ when α s >0, and α→∞ when α s <0 (and β→0 in all cases). In the former case the asymptotic analysis shows that α−αs∝(L−1   dL/dt)1/2 and β∝(L−1   dL/dt)1/2 . In the latter case it shows that α∝(L−1   dL/dt)−1 and β∝L−1   dL/dt. These results are supported by the numerical solutions for a variety of cases. Some of the corresponding results for the concentration moments are compared with experimental measurements for line sources in wind tunnels.
Keywords: Turbulent diffusion; Modelling concentration moments; Line sources;

Monitoring of PM10 and PM2.5 around primary particulate anthropogenic emission sources by Xavier Querol; Andrés Alastuey; Sergio Rodriguez; Felicià Plana; Enrique Mantilla; Carmen R Ruiz (845-858).
Investigations on the monitoring of ambient air levels of atmospheric particulates were developed around a large source of primary anthropogenic particulate emissions: the industrial ceramic area in the province of Castelló (Eastern Spain). Although these primary particulate emissions have a coarse grain-size distribution, the atmospheric transport dominated by the breeze circulation accounts for a grain-size segregation, which results in ambient air particles occurring mainly in the 2.5–10 μm range. The chemical composition of the ceramic particulate emissions is very similar to the crustal end-member but the use of high Al, Ti and Fe as tracer elements as well as a peculiar grain-size distribution in the insoluble major phases allow us to identify the ceramic input in the bulk particulate matter. PM2.5 instead of PM10 monitoring may avoid the interference of crustal particles without a major reduction in the secondary anthropogenic load, with the exception of nitrate. However, a methodology based in PM2.5 measurement alone is not adequate for monitoring the impact of primary particulate emissions (such as ceramic emissions) on air quality, since the major ambient air particles derived from these emissions are mainly in the range of 2.5–10 μm. Consequently, in areas characterised by major secondary particulate emissions, PM2.5 monitoring should detect anthropogenic particulate pollutants without crustal particulate interference, whereas PM10 measurements should be used in areas with major primary anthropogenic particulate emissions.
Keywords: Atmospheric particles; PM10; PM2.5; Ceramic emissions; Ambient air quality monitoring; Primary particles; Mediterranean basin;

Using a specially designed closed chamber, the time evolution of the size distributions of airborne cat allergen (Fel d 1) was measured at three heights during the first 55 min following resuspension of dust from a carpet using a vacuum cleaner. By analyzing these data, indoor removal rates were estimated, along with the initial concentrations present. These parameters were then used to model the cumulative dose of allergen in various regions of the respiratory tract following a disturbance event. Because most of the resuspended Fel d 1 mass initially consisted of coarse particles, 77% by mass of the total cumulative dose in the first 30 min after a disturbance was predicted to deposit in the nose.
Keywords: Size distributions; Bioaerosols; Indoor sources; Dust resuspension; Domestic activities;

Currently, legislation is being considered to reduce NH3 emissions in the UK. The major sources of NH3 and their relative contributions are well known, however, the processes that control the rates of emission are still poorly defined. A series of wind-tunnel experiments has been carried out to determine the effects of various management practices on NH3 losses. The tunnels were modified to enable NH3 emission and subsequent deposition to the adjacent swards in the field to be measured. The wind-tunnels were used to examine the effects of herbage length, cutting and N status on rates of NH3 fluxes, which together with the prevailing environmental conditions affected the rates of NH3 emission and deposition. Results showed that between 20 and 60% of the NH3 emitted was deposited within 2 m. Compensation points of between 1.0 and 2.3 μg m−3 were calculated for the grass sward.
Keywords: Ammonia; Compensation point; Nitrogen; Urine; Pollution; Wind-tunnels;

We present a plume rise model which can be applied to situations with arbitrary wind fields and source exit directions and to both dry and wet plumes. The model is an integral model which considers plume properties averaged over the plume cross section. It is validated by means of water tank, wind tunnel, and field experiments (stacks and cooling towers).
Keywords: Plume rise; Entrainment; Condensation; Turbulence; Dispersion;

Uncertainties in predicted ozone concentrations due to input uncertainties for the UAM-V photochemical grid model applied to the July 1995 OTAG domain by Steven R Hanna; Zhigang Lu; H Christopher Frey; Neil Wheeler; Jeffrey Vukovich; Saravanan Arunachalam; Mark Fernau; D Alan Hansen (891-903).
The photochemical grid model, UAM-V, has been used by regulatory agencies to make decisions concerning emissions controls, based on studies of the July 1995 ozone episode in the eastern US. The current research concerns the effect of the uncertainties in UAM-V input variables (emissions, initial and boundary conditions, meteorological variables, and chemical reactions) on the uncertainties in UAM-V ozone predictions. Uncertainties of 128 input variables have been estimated and most range from about 20% to a factor of two. 100 Monte Carlo runs, each with new resampled values of each of the 128 input variables, have been made for given sets of median emissions assumptions. Emphasis is on the maximum hourly-averaged ozone concentration during the 12–14 July 1995 period. The distribution function of the 100 Monte Carlo predicted domain-wide maximum ozone concentrations is consistently close to log-normal with a 95% uncertainty range extending over plus and minus a factor of about 1.6 from the median. Uncertainties in ozone predictions are found to be most strongly correlated with uncertainties in the NO2 photolysis rate. Also important are wind speed and direction, relative humidity, cloud cover, and biogenic VOC emissions. Differences in median predicted maximum ozone concentrations for three alternate emissions control assumptions were investigated, with the result that (1) the suggested year-2007 emissions changes would likely be effective in reducing concentrations from those for the year-1995 actual emissions, that (2) an additional 50% NO x emissions reductions would likely be effective in further reducing concentrations, and that (3) an additional 50% VOC emission reductions may not be effective in further reducing concentrations.
Keywords: Photochemical modeling; Model uncertainty; Monte Carlo uncertainty methods;

The 1977 and 1990 Amendments to the Clean Air Act call for visibility and atmospheric deposition monitoring throughout the United States. We compare sulfate and nitrate particle mass concentrations measured by two regional air quality networks, the Interagency Monitoring of PROtected Visual Environments (IMPROVE) network and the Clean Air Status and Trends Network (CASTNet), or CASTNet Deposition Network (CDN). The intent of this comparison is to quantify bias that may be introduced from differences in the respective network's sampling protocols. A number of sampling protocol differences exist between the two networks that may lead to sampling bias, particularly for particle NO3 . Observed differences between particle SO4 2− mass concentrations reported by the two monitoring networks are generally small, yet statistically significant at many comparison sites. Differences between particle NO3 mass concentrations are substantial, statistically significant at nearly all comparison sites, and the bias magnitude varies by geographic region. Differences in particle NO3 , based on data from monitoring sites selected for this comparison, are 40% in the west, 56% in the interior desert/mountain region, and −9% in the east, expressed as the IMPROVE mean subtracted from the CDN mean, as a percent of the IMPROVE mean. Comparisons are made using data from 23 locations where monitoring sites from IMPROVE and CDN are within approximately 50 km.
Keywords: Aerosol monitoring; CASTNet; Measurement bias; Dry deposition; Visibility;

Seasonal variations in VOC emission rates from gorse (Ulex europaeus) by C Boissard; X.-L Cao; C.-Y Juan; C.N Hewitt; M Gallagher (917-927).
Seasonal variations of biogenic volatile organic compound (VOC) emission rates and standardised emission factors from gorse (Ulex europaeus) have been measured at two sites in the United Kingdom, from October 1994 to September 1995, within temperature and PAR conditions ranging from 3 to 34°C and 10–1300 μmol m−2  s−1, respectively. Isoprene was the dominant emitted compound with a relative composition fluctuating from 7% of the total VOC (winter) to 97% (late summer). The monoterpenes α-pinene, camphene, sabinene, β-pinene, myrcene, limonene, trans-ocimene and γ-terpinene were also emitted, with α-pinene being the dominant monoterpene during most the year. Trans-ocimene represented 33–66% of the total monoterpene during the hottest months from June to September. VOC emissions were found to be accurately predicted using existing algorithms. Standard (normalised) emission factors of VOCs from gorse were calculated using experimental parameters measured during the experiment and found to fluctuate with season, from 13.3±2.1 to 0.1±0.1 μg C (g dwt)−1  h−1 in August 1995 and January 1995, respectively, for isoprene, and from 2.5±0.2 to 0.4±0.2 μg C (g dwt)−1  h−1 in July and November 1995, respectively, for total monoterpenes. No simple clear relation was found to allow prediction of these seasonal variations with respect to temperature and light intensity. The effects of using inappropriate algorithms to derive VOC fluxes from gorse were assessed for isoprene and monoterpenes. Although on an annual basis the discrepancies are not significant, monthly estimation of isoprene were found to be overestimated by more than a factor of 50 during wintertime when the seasonality of emission factors is not considered.
Keywords: Biogenic emissions; Ulex europaeus; Seasonal variations; Emission factors; VOCs; Long term;

Single-day scenarios are used to calculate incremental reactivities by definition (Carter, J. Air Waste Management Assoc. 44 (1994) 881–899.) but even unreactive organic compounds may have a non-negligible effect on ozone concentrations if multiple-day scenarios are considered. The concentration of unreactive compounds and their products may build up over a multiple-day period and the oxidation products may be highly reactive or highly unreactive affecting the overall incremental reactivity of the organic compound. We have developed a method for calculating incremental reactivities for multiple days based on a standard scenario for polluted European conditions. This method was used to estimate maximum incremental reactivities (MIR) and maximum ozone incremental reactivities (MOIR) for ethane and dimethyoxymethane for scenarios ranging from 1 to 6 days. It was found that the incremental reactivities increased as the length of the simulation period increased. The MIR of ethane increased faster than the value for dimethyoxymethane as the scenarios became longer. The MOIRs of ethane and dimethyoxymethane increased but the change was more modest for scenarios longer than 3 days. MOIRs of both volatile organic compounds were equal within the uncertainties of their chemical mechanisms by the 5 day scenario. These results show that dimethyoxymethane has an ozone forming potential on a per mass basis that is only somewhat greater than ethane if multiple-day scenarios are considered.
Keywords: Ozone; Ozone formation potential; Incremental reactivity; VOC oxidation mechanisms;

Methyl bromide (MeBr) is commonly used for fumigating structures and commodities. Emission of MeBr during such treatments is environmentally detrimental because of the reaction of MeBr with stratospheric ozone. In this study we evaluated adsorption of MeBr and methyl iodide (MeI) – a potential MeBr replacement, on five commercial activated carbons, and studied water-initiated catalytic decomposition of adsorbed fumigants. All carbon samples showed great adsorption affinity to MeBr and MeI, with the adsorption capacity for MeI several times greater than that for MeBr on the same carbon. For the same fumigant, adsorption was affected by the type of carbon and the concentration of fumigant. Water initiated decomposition of both fumigants, liberating Br or I as a transformation product. The rate of decomposition increased with increasing temperature, and was also influenced by the carbon type. The half-life of MeBr or MeI on Centaur, a catalytically modified carbon, was <2 h at 80°C. The rapid decomposition of MeBr and MeI on wet carbons at elevated temperatures may be used to detoxify these fumigants after adsorption on activated carbons.
Keywords: Methyl bromide; Methyl iodide; Methyl halides; Activated carbon; Adsorption; Fumigation; Fumigant; Catalytic decomposition; Emission control;

Evaluation of the CAR-FMI model against measurements near a major road by Jaakko Kukkonen; Jari Härkönen; Jari Walden; Ari Karppinen; Kaisa Lusa (949-960).
A field measurement campaign was conducted near a major road in southern Finland from September 15 to October 30, 1995. The concentrations of NO, NO2 and O3 were measured simultaneously at three locations, at three heights (3.5, 6 and 10 m) on both sides of the road. Traffic densities and relevant meteorological parameters were also measured on-site. We have compared measured concentration data with the predictions of the road network dispersion model CAR-FMI, used in combination with a meteorological pre-processing model MPP-FMI. In comparison with corresponding results presented previously in the literature, the agreement of measured and predicted datasets was good, as measured using various statistical parameters. For all data (N=587), the index of agreement (IA) was 0.83, 0.82 and 0.89 for the measurements of NO x , NO2 and O3, respectively. The IA is a statistical measure of the correlation of the predicted and measured time series of concentrations. However, the modelling system overpredicts NO x concentrations with a fractional bias FB=+13%, and O3 concentrations with FB=+8%, while for NO2 concentrations FB=−2%. We also analyzed the difference between model predictions and measured data in terms of meteorological parameters. Model performance clearly deteriorated as the wind direction approached a direction parallel to the road, and for the lowest wind speeds. The range of variability concerning atmospheric stability, ambient temperature and the amount of solar radiation was modest during the measurement campaign. As expected, no clear dependencies of model performance were therefore detected in terms of these parameters. The experimental dataset is available for the evaluation of other roadside dispersion models.
Keywords: Dispersion; Traffic; Pollution; Road; Model; Measurement campaign; Nitrogen oxide;

The estimation of organic gas vapour pressure by Paul A. Makar (961-974).
The base 10 logarithm of vapour pressure for each of 39 structural classes of hydrocarbons and substituted hydrocarbons is a linear function of carbon number, and a cubic polynomial of temperature. These functions may be used to estimate the vapour pressures of compounds for which no laboratory measurements are available. Comparison of the resulting vapour pressure predictions with field measurement data of organic particulates has suggested that vapour pressures lower than approximately 5E−05 mm Hg are necessary for low-solubility compounds to partition significantly to particulates. Soluble compounds may enter the particle phase at much higher vapour pressures. A hierarchy of species’ likelihood to partition to the particulate phase, based on vapour pressure, has been suggested for the compounds studied, from diacids being the most likely to partition, to iso-alkyl ketones as the least likely to partition.

Nitrogen oxide fluxes between corn (Zea mays L.) leaves and the atmosphere by Daniel P. Hereid; Russell K. Monson (975-983).
In the United States, fertilized corn fields, which make up approximately 5% of the total land area, account for approximately 45% of total soil NO x emissions. Leaf chamber measurements were conducted of NO and NO2 fluxes between individual corn leaves and the atmosphere in (1) field-grown plants near Champaign, IL (USA) in order to assess the potential role of corn canopies in mitigating soil–NO x emissions to the atmosphere, and (2) greenhouse-grown plants in order to study the influence of various environmental variables and physiological factors on the dynamics of NO2 flux. In field-grown plants, fluxes of NO were small and inconsistent from plant to plant. At ambient NO concentrations between 0.1 and 0.3 ppbv, average fluxes were zero. At ambient NO concentrations above 1 ppbv, NO uptake occurred, but fluxes were so small (14.3±0.0 pmol m−2  s−1) as to be insignificant in the NO x inventory for this site. In field-grown plants, NO2 was emitted to the atmosphere at ambient NO2 concentrations below 0.9 ppbv (the NO2 compensation point), with the highest rate of emission being 50 pmol m−2  s−1 at 0.2 ppbv. NO2 was assimilated by corn leaves at ambient NO2 concentrations above 0.9 ppbv, with the maximum observed uptake rate being 643 pmol m−2 s−1 at 6 ppbv. When fluxes above 0.9 ppbv are standardized for ambient NO2 concentration, the resultant deposition velocity was 1.2±0.1 mm s−1. When scaled to the entire corn canopy, NO2 uptake rates can be estimated to be as much as 27% of the soil-emitted NO x . In greenhouse-grown and field-grown leaves, NO2 deposition velocity was dependent on incident photosynthetic photon flux density (PPFD; 400–700 nm), whether measured above or below the NO2 compensation point. The shape of the PPFD dependence, and its response to ambient humidity in an experiment with greenhouse-grown plants, led to the conclusion that stomatal conductance is a primary determinant of the PPFD response. However, in field-grown leaves, measured NO2 deposition velocities were always lower than those predicted by a model solely dependent on stomatal conductance. It is concluded that NO2 uptake rate is highest when N availability is highest, not when the leaf deficit for N is highest. It is also concluded that the primary limitations to leaf-level NO2 uptake concern both stomatal and mesophyll components.
Keywords: Nitrogen oxides; Dry deposition; Flux; Stomatal uptake; Corn plants;

We evaluated the loss of HNO3 within a Teflon-coated aluminum cyclone of an annular diffusion denuder atmospheric sampling system (ADS) under simulated marine conditions. To simulate marine environment, the cyclones were pre-coated with NaCl aerosol droplets. Loss of vapor-phase HNO3 within the NaCl-coated cyclone was generally greater than 30% at relative humidities (RH) of 60 and 80% and as large as 67% when the cumulative HNO3 dosages were lower than 3 μg. In contrast, there was little loss of HNO3 (<8%) in cyclones with no NaCl coating at RHs ranging from 0 to 80%, at HNO3 air concentrations of 4.3±1.6 μg m−3, and at cumulative HNO3 dosages of greater than 5 μg. However, at lower HNO3 cumulative dosages (<3 μg), losses in the non-coated cyclones were strongly influenced by RH, ranging from 9% in dry air to 58% at 80% RH. The enhanced loss of HNO3 in the NaCl-coated cyclone was most likely caused by the reaction between HNO3 and NaCl on the cyclone wall.
Keywords: HNO3 gas measurement; HNO3 inlet loss; Cyclone; Denuder; Marine atmospheric sampling;