Air Pollution Modeling and Its Application XII

We have coupled a nonhydrostatic mesoscale model with a simple but comprehensive mineral aerosol source scheme, along with a spectral sedimentation scheme. We present a simulation of a Saharan dust transport event (4 days), including mass uptake estimates, 3D transport and dry deposition. The model is initialized with ECMWF data. Meteosat imagery is used to check the dust cloud uptake and trajectory.

In several studies during the past, the urban plumes have been extensively considered. In these studies, the spatial and temporal scales of episodic conditions have been described and emphasis was given to the formation and evolution of air pollution episodes within city limits (or in an area covering a few tens of kilometers around the city) and for a time period of one to two days. Moreover, the weather phenomena exhibiting strong diurnal variations (e. g. sea/land-breezes, upslope/downslope and drainage flows, orographic effects, heat islands etc.) were emphasized. The influence of the regional scale phenomena in such cases was not considered on a systematic manner. Actually, the role of phenomena with wavelengths larger than a few tens of kilometers was considered as not important for the formation of a specific air quality over the city of consideration. During the last few years, the influence of regional scale forcing on the formation of specific air quality conditions was found to be important. Kallos et al. (1993) reported that the regional scale phenomena should contribute significantly in the formation of specific air quality conditions in the Greater Athens Area (GAA). Luria et al. (1996) showed that significant degradation of the air quality in some areas should be attributed to regional scale transport phenomena. While the physicochemical properties of various urban plumes have been described at the urban scale with the aid of organized experimental campaigns and/or mesoscale and photochemical modeling (e. g. Ziomas, 1996), not enough attention was paid to the properties of the urban plume as it is passing to areas relatively far from its origin. Consequently, the urban plume impact on remote locations has not been extensively studied. Such phenomena should be considered as very important in some cases, especially in areas with specific characteristics like the Mediterranean Region.

In regional-scale air-pollution models probably no other source of uncertainty ranks higher than the current ability to specify clouds and soil moisture. Because modeled clouds are highly parameterized, the ability of models to predict the magnitude and spatial distribution of radiative characteristics is highly suspect and subject to large error. While considerable advances have been made in the assimilation of winds and temperatures into regional models (Stauffer and Seaman, 1990), the poor representation of cloud fields from point measurements at National Weather Service stations and the almost total absence of observations of surface moisture availability has made assimilation of these variables difficult if not impossible. Yet, the correct inclusion of clouds and surface moisture are of first-order importance in regional-scale photochemistry. Consider the following points relative to these variables.

The Fraser Valley is a complex topographic coastal environment which episodically experiences visibility degradation (and elevated aerosol concentrations) (Pryor et al., 1997). To examine concentrations and speciation of secondary inorganic aerosols and ozone in the transition between an oxidant event and period of elevated aerosol concentrations, numerical simulations were performed using a modified version of the ACDEP (Atmospheric Chemistry and Deposition) model. ACDEP is a lagrangian model which contains detailed and fully coupled gas-aerosol phase chemistry (Hertel et al., 1995). The modeling period is August 5–8, 1993, and the modeling domain is shown in Figure 1. Results calculated for two receptor sites are shown herein (locations specified in Figure 1). PIME is located in the western reaches of the valley directly east of the Vancouver metropolitan area, and CHIL is located in a region of mixed agricultural land, approximately 80 km from the Vancouver urban core. The modified ACDEP model was applied at a horizontal resolution of 5 km and with 10 layers in the vertical, increasing logarithmically from 2 m to 2 km. The meteorological parameterizations within ACDEP have been extended such that stability parameters and mixed layer depth are calculated using routinely available meteorological data.

The area around Milazzo, a city located on a narrow peninsula, north-east Sicily, is heavily polluted by the presence of power plants and refinery. In this work, the atmospheric pollution levels resulting from two circulation regimes are simulated. First, westerly synoptic winds are considered, as the most likely meteorological conditions (Graziani et al., 1997). Then, a less frequent but most dangerous situation with southerly winds is examined. In this meteorological condition, weak winds can develop that may favour the accumulation of pollutants in the most populated area of the city of Milazzo. The RAMS model (version 3b) is used to calculate the flow, in its non-hydrostatic configuration, due to the highly complex terrain. RAMS is initialised and nudged with data from ECMWF model and a multiple nested grid configuration is selected to solve both local flow and large scale circulation over the Mediterranean. The time-dependent RAMS output is used to drive the Lagrangian particle model MONTECARLO, written in terrain following co-ordinates. In the model, buoyancy of hot emissions in convective conditions is taken into account. Non-reactive SO₂ sources are only considered. Both meteorological and concentration model outputs are compared with observations performed during the 1996 Spring campaign (Cerutti et al., 1996).

The Katowice province in southern Poland experiences serious air pollution problems including deposition of heavy metals which is among the highest in Europe (Bartnicki et al. 1996; Olendrzyński et al., 1996). Heavy metals deposit onto surfaces at relatively low rates, however, due to their toxicity and accumulation in soils, long-term deposition needs to be evaluated. Long-range transport models applied to the whole of Europe cannot simulate high values of local deposition fluxes because of their low spatial resolution. A typical grid cell of the long-range models can cover the whole Katowice province. Therefore, it is necessary to apply a high resolution transport and deposition model linked to a three-dimensional mesoscale/regional meteorological model. Recent advances in computer technology, especially availability of modern workstations, allows one to integrate 3-dimensional mesoscale meteorological models over extended time periods, months or even years, to provide necessary input fields for long term dispersion modeling (Uliasz et al., 1996; Pielke and Uliasz, 1997). This paper presents selected results from the METKAT (heavy METals in the KATowice province) project launched by the International Institute for Applied Systems Analysis in cooperation with two Polish research institutions, Institute for Ecology of Industrial Areas (IEIA), Katowice, and Institute of Meteorology and Water Management (IMWM) in Warsaw, The goal of the project was to investigate high local deposition fluxes of arsenic (As), cadmium (Cd), lead (Pb) and zinc (Zn) in the Katowice province with the aid of mesoscale modeling. The extended description of the METKAT study is provided by Uliasz and Olendrzyński (1996).

Tropical biomass biogeochemistry is one of the most poorly understood on the earth. The tropics account for about 60% of the global annual net primary productivity and this enormous productivity is characterized by many chemical species which are emitted as gas or aerosols and can modify the global radiative balance. Biomass burning is associated with agricultural activity in the savannah, the destruction of tropical forests and the use of wood as “fuel”. They release into the atmosphere large quantities of CO2 and a variety of chemically active species such as CO, NOx, N₂O, CH₄, and others expressed in (1, 2). The biomass annually burned in the world represents about 1.8 to 4.7 GT. of carbon (3), savannah fires being the dominant component with about 1-1.6 GT. of carbon burned. Savannah fires alone contribute approximately 10% of the global CO emissions. This phenomenon is especially important in Africa. The contribution of African savannah fires to global emission of trace gas and aerosols has been estimated by (4). During dry season, pollution events are similar in magnitude to those observed in industrialised regions as observed in high levels of acid precipitation that were reported in these region (5).

Concentrations of surface ozone over Europe currently exceeds the critical levels over which damage to vegetation and health may occur in many locations. This is true also in northern Europe and over Sweden regarding critical levels for vegetation and also occasionally, during summer in southern Sweden, regarding critical levels for human health. Optimising measures to reduce these exeedances, both nationally and internationally, requires a better understanding of the importance of various types of precursor emissions and processes influencing the distribution of photochemical oxidants.

The Fraser Valley in the southwestern province of British Columbia, Canada is one of Canada’s highest ozone concentration areas where the National Ambient Air Quality Objective of 82 ppb per hour is exceeded several times a year under high pressure ridge conditions (Taylor, 1991; McKendry, 1994). Modeling efforts are aimed at supporting local regulatory agencies in their attempt to apply appropriate control strategies to the problem. The complex nature of the Fraser Valley offers a modeling challenge. The irregular terrain (Figure 1) generates complex mesoscale flows and modeling is further complicated by changing land use from the urban centre of Vancouver in the west to agricultural farmland in the east. Located on the east coast of the Pacific Ocean, the area is not influenced by upstream transport and thus provides a unique closed system in which to assess and verify model behavior.

Because photochemical grid models such as UAM-IV are being used to make policy decisions concerning emissions controls, it is important to know what confidence bounds we can place on the model predictions of, for example, how ozone will respond to changes in emissions. These are presently unknown. The factors influencing prediction error can be classified as input errors, as model formulation errors, or as random stochastic processes. In the present study we include among inputs such things as initial and boundary conditions, emissions, meteorological variables and chemical rate constants. Bias, imprecision and variability can contribute to the error. Formulation errors would include such things as inaccuracies in advection schemes, numerical solvers, process representations, and temporal and spatial resolution. Sometimes the distinction between input and formulation errors is not well drawn. The study described here has been limited by time and resource constraints to an examination of the prediction error associated only with input, not with model formulation, errors. We therefore implicitly assume, without justification, that the model physics and chemistry are correctly formulated. Since model formulation can influence not only simulation fidelity but how input errors are propagated through the model to output errors, we view the results of this study more as a methodological demonstration than a definitive uncertainty analysis.

The initialisation and the treatment of the boundary conditions of a mesoscale chemistry-transport-model, covering a limited area, are of great importance. The choice of the initial and boundary values can significantly influence the results of a simulation, so that they should be determined as well as possible (NAPAP, 1991). For this reason it is important to provide realistic conditions, if possible derived from current measurements. Unfortunately trace species in the troposphere, especially in the middle and upper free troposphere, are not observed continuously so that relatively little is known about background concentrations. Usually there are no current observations available which can be used as input data for episodic simulations. The available measurements show a high variability in the concentrations of the trace species. To analyse and to quantify the effects of a variation of the initial and boundary values for model results sensitivity studies were carried out with the European Air Pollution Dispersion modeling system (EURAD) using different initial and boundary scenarios. Therefore the literature has been reviewed and a set of initial and boundary values were derived based on available observation data. With regard to the formation of ozone the focus was set on reactive nitrogen species and hydrocarbons, which are important photooxidant precursor species. First a set of simulations with different scenarios representing free tropospheric conditions is calculated with a boxmodel version of the EURAD model in order to determine the non-linear dependencies of the gas phase chemistry. Furthermore a sensitivity study with the full three dimensional model is performed for a summersmog episode.

Photochemical air quality models increasingly are being used to understand the atmospheric dynamics of air pollutants and as the basis of emission control regulations. The response of the these model predictions to system parameters or emission controls provides valuable information for the strategy design to improve air quality. Such information can be pursued via sensitivity analysis, the systematic calculation of sensitivity coefficients, to quantitatively measure these dependencies. However, sensitivity analysis has not seen as wide of use as desired, in part because of the implementation complexity as well as computational limitations. For these reasons, sensitivity analysis has been applied primarily to subsystems of air quality models (e. g. Koda et al., 1974; Rabitz et al., 1983; Milford et al., 1992), or to limited aspects in air quality models (Cho et al., 1987). The “brute-force” method has been the most frequently used to determine model sensitivities, but it rapidly becomes less viable and prohibitively inefficient for a model when a large number of sensitivity coefficients needs to be computed.

Many urban areas in the Eastern United States have been classified to be in non-attainment for ozone, placing a high priority on finding cost-effective emission control measures for improving ambient ozone air quality. Recognizing the complexities associated with the nation’s ozone non-attainment problem, the 1990 Clean Air Amendments mandated the use of grid-based photochemical models for evaluating emission control strategies in urban areas having a serious or higher designation. Given the influx of elevated concentrations of ozone and its precursors into the urban-scale modeling domains (regional-scale transport), many states in the Eastern U.S. were unable to demonstrate ozone attainment for urban areas in their 1994 State Implementation Plans (SIPs) submittal using the urban-scale models. The 1994 SIPs were based on the UAM-IV photochemical model (Morris et al., 1990), which is an urban-scale model that reflects the state-of-science of the late 1980’s. Systems Applications International (SAI) recently developed the UAM-V, a regional-scale ozone air quality model, which contains some new features over the UAM-IV such as multi-scale modeling capability, grid nesting, plume-in-grid (PiG) treatment for point sources, etc. (SAI, 1995). Of particular interest is this model’s treatment of subgrid-scale processes relating to the transport, transformation, and interaction of elevated plumes with the ground-level plume.

Due to its high density of urban and industrial sources, the eastern United States often experiences widespread pollution episodes during the summer (Logan, 1985; Vukovich and Fishman, 1986). The effects of such continental emmissions on the oxiding capacity of atmosphere over the North Atlantic have been studied for many years. Most of the earlier works (Zeller et al., 1977; Kelleher and Feder, 1978; Spicer, 1982) presented evidence for the transport of plumes from the eastern seaboard of the United States out over 100 km or more of the North Atlantic. Measurements at Kejimbuijk National Park in Canada (Brice et al., 1988; Beattie and Wepdale, 1989), begun in 1979, demonstrate transport of these plumes to central Nova Scotia, located at a distance of than 500 km more. The 1993 North Atlantic Regional Experiment (N.A.R.E.) intensive provided further evidences for the transport of anthropogenic pollutants and ozone precussors (CH₄, CO,...) from the continent sources out over the Atlantic ocean (Fehsenfeld et al., 1996). We expect that many tropospheric photo-oxidants are generated by chemical reactions, in particular the hydrogen peroxide. This latter presents a real interest for acidification of the clouds and in gas-phase as a efficient source of OH.

The air pollution in a region is influenced by all atmospheric scales. Because it is not possible to resolve all scales in a single model, different models have been developed which describe the air pollution in certain regions, like the European scale model EURAD (EURopean Acid Deposition model (Ebel et al., 1989)) and the mesoscale model system KAMM/DRAIS (KArlsruher Meteorologisches Modell (Adrian and Fiedler, 1991)/DReidimensionales Ausbreitungs- und Immissions-Simulationsmodell (Schwartz, 1996)). Those effects which are smaller than the grid resolution of a model (subgrid effects) have to be considered by appropriate parameterizations or by nesting procedures. Effects resulting from scales which are even larger than the model domain are usually introduced into the model by the boundary conditions.

CALPUFF is a non-steady-state, multi-layer puff dispersion model that simulates the effects of time and space varying meteorological conditions. The puff modeling simulation results are compared with data obtained following a single 3-hour late-afternoon tracer release conducted on April 19, 1977 near Idaho Falls, Idaho, USA. Samplers were positioned on arcs at downwind distances of 3.2, 48 and 90 km. Low-level instrumented masts provided hourly values of wind and temperature at 17 sites within the experimental area. Hourly rawindsondes were available 600 m northwest of the release point. And hourly pibals provided winds aloft at three other sites within the area. In this discussion, alternative combinations of the available meteorological data were used to assess the differences to be seen in the simulation results. Analysis of the results suggests that the simulated lateral dispersion along each arc was best characterized when all of the surface and upper wind observations were used and that the position of the simulated maximum on each arc was poorly characterized regardless of data used.

The purpose of developing atmospheric transport-chemistry models is to create a tool, an instrument to investigate the underlying processes which lead to the chemical composition of the atmosphere. Modelling, in the sense both of developing, using and applying models is a scientific activity as it is focused on answering fundamental questions in the field of atmospheric chemistry.

According to Stokes law which is well fulfilled within the range of 0.1–30 μm, the velocity of gravitational deposition (sedimentation) and scavenging with precipitation for particles is proportional to their surface area, i. e. to D2 (D - particle size). With D less than 0.1 μm, their behavior is similar to that of “weightless” gas molecules and the velocities of surface dry deposition and the efficiency of particle captures by precipitation increasing.

The aim of the present project is to study the transport and photochemistry of regional air pollution in Greece using numerical models, a higher-order turbulence closure dynamic model coupled with an Eulerian photochemical dispersion model. The model used for the simulation of the three-dimensional wind field and the boundary layer structure is developed at the Department of Meteorology, Uppsala University (MIUU) while the calculations of pollutant concentrations were performed using the Urban Airshed Model (UAM).For the purposes of the present study, a new inventory for the biogenic emissions is prepared comprising hourly emissions from forest canopies and other vegetation as a function of temperature, sunlight (cloud cover), and the coverage of each vegetation category. The VOC speciation of the inventory is suitable as input to the UAM model. In addition, the industrial and traffic emissions were estimated for the whole region with a resolution of 10 × 10 km and used as input for the UAM.One summer day was selected for simulation during the period of the PAUR project, when the conditions were characterized by a moderate-to-strong synoptic forcing. The resulting meteorological conditions reveal a northerly flow prevailing over the Aegean sea which is known as Etesians. In the areas that are influenced by the Etesians, the development of local circulations is not favored and the flow field is rather homogeneous. The predicted concentrations of air pollutants are in good agreement with observations at rural sites.

Acid deposition is widely recognized as one of the most serious global atmospheric pollution problems. Regional scale, international actions to tackle this problem have been taken in Europe and North America. East Asian countries also face a potential regional scale, international acid deposition problem, and have recently started to expand their monitoring activities as a result.

In 1994, the UN Economic Commission for Europe (ECE) proposed that the future work on heavy metals should concentrate on priority elements: Pb, Hg, Cd, Cr, Ni, Zn, Cu, As and Se with particular focus on the first three metals. Concerning environmental impact of heavy metals, their concentrations, especially in remote areas, are usually too low to cause any serious adverse effects. However, the concentrations in remote parts of Europe can increase significantly during episodes of long range transport from anthropogenic sources. In addition, even low deposition fluxes accumulate in the soil and can reach the level at which heavy metals become mobile in the environment. Therefore, depositions and especially long-term cumulative depositions of heavy metals have to be monitored.

The PEM West-B. Pacific Exploratory Mission West: phase B. experiment was carried out from the beginning of February to the mid of March. 1994. The experiment explored chemical composition of the atmosphere over the tropical Pacific Ocean and the Pacific rim area of Asia. The experiment is valuable for understanding of long range transport/transformation of air pollutants released over Asian industrial area in late winter and early spring, and is useful for model validation in semi-global scale.

The subject of long range transport (LRT) of pollutants is in the early stages of study in Asia (Merrill et al., 1985; Kotamarthi and Carmichael, 1990; Arndt et al., 1996). The results from the PEM West A & B experiments have shown the widespread impact that long range transport of materials from the continental regions of east Asia has on the chemical composition of the troposphere over the western Pacific (Hoell et al., 1996). The projected growth in emissions for the region suggests that the long range transport of pollutants in east Asia will grow in importance over the next several decades.

Throughout this century the composition of the troposphere has been perturbed by anthropogenic emissions. These emissions are not uniformly distributed over the globe but are concentrated in the northern mid-latitudes in industrial regions. Biogenic emissions are largely emitted from the land and so too are biased towards the northern hemisphere. A modelling study of the global chemistry of the troposphere requires knowledge of how pollutants are carried away from their mainly continental source regions to the remote atmosphere and the southern hemisphere.

Saharan dust storms are the main source of the atmospheric dust in the Mediterranean region. Once injected to the atmosphere, dust may pass long distances under favourable meteorological conditions before it deposits to the ground or sea surfaces. Typically, several hundreds millions of tonnes of dust is transported away from sources annually (D’Almeida, 1986). Continuous presence of dust in the atmosphere causes diverse climatic and environmental effects. For example, dust modifies radiation properties of the air through absorption and scattering of the solar energy on dust particles (e. g. Chen et al., 1995). Recent estimate of Tegen and Fung (1994) shows that mineral dust may decrease the net radiation for about 1 Wm−2, revealing thus the fact that it could be a significant climate forcing factor. Another environmental effect of the dust process is dust deposition on the sea surface, which may significantly change the marine biochemical properties (e. g. Martin and Fitzwater, 1988; Kubilay and Saydam, 1995). Also, the atmospheric dust may significantly influence human activities: for example, it reduces the visibility, causing thus problems in the air and ground traffic; during dust storms, increased number of eye and respiratory organs infections is recorded, too.

The atmosphere contains a number of radiatively and chemically important trace gases (a. o. carbon dioxide (CO₂), carbon monoxide (CO), nitrous oxide (N₂O) and methane (CH₄)) whose concentrations are changing in the atmosphere, primarily due to human activities. These changing concentrations affect the radiative balance of our atmosphere and may thus lead to climate change. In order to compute reliable projections of the future evolution of the concentrations of these gases their natural and anthropogenic sources and sinks have to be known. Using a direct approach one can extrapolate locally measured fluxes to the entire globe. Because of the many necessary assumptions in the extrapolation, this direct approach, however, is subject to very large uncertainties. In contrast, one can apply an inverse approach, in which ambient observations of the atmospheric trace gas concentrations are used to constrain the surface fluxes. This requires a model of the atmospheric transport which provides the link between the surface fluxes and the concentrations at the monitoring sites.

As part of the European Tracer Experiment (ETEX) two successful atmospheric experiments were carried out in October and November, 1994. Perfluorocarbon (PFC) tracers were released into the atmosphere in Monterfil, Brittany, and air samples were taken at 168 stations in 17 European countries for 72 hours after the release. Upper air tracer measurements were made from three aircraft. During the first experiment a westerly air flow transported the tracer plume north-eastwards across Europe. During the second release the flow was eastwards. The results from the ground sampling network allowed the determination of the cloud evolution as far as Sweden, Poland and Bulgaria. Typical background concentrations of the tracer used are around 5 to 7 fl/l in ambient air. Concentrations in the plume ranged from 10 to above 200 fl/l.

The first European long-range tracer experiment (ETEX), jointly organised by the European Commission, the World Meteorological Organisation and the International Atomic Energy Agency, took place on October 23, 1994. The aim of the experiment was to simulate an emergency situation following a release of harmful material into the atmosphere, and to test both real-time and a-posteriori modeling capabilities of reproducing in space and time the tracer concentration field. An inert tracer (perfluoromethylcyclohexane, a perfluorocarbon compound) was released near Rennes, in Northwest France, for twelve hours starting form 16 h UTC, and sampled at three-hourly intervals by 168 ground sites up to a distance of about 2000 km. The synoptic situation at the beginning of the release was characterised by a deep low east of Scotland slowly moving North, maintaining a strong flow from west-southwest in the lower layers over the release site. This condition, together with the correct forecast of its evolution for the following three days, yielded a large number of sampling stations, sparse over Central and Northern Europe, detecting concentrations above background, and so determined the success of the experiment. The concentration data set constitutes the base for both real-time and a-posteriori model evaluations. We present here an a-posteriori intercomparison between two Lagrangian particle models (APOLLO, developed at ANPA, and MILORD, developed at CNR-ICG) simulations of ETEX.

Numerous numerical dispersion models have been developed to predict long range transport of hazardous air pollution in connection with accidental releases. When developing and evaluating such a model, it is important to detect uncertainties connected to insufficient characterization of the accidental release, the meteorological input data, errors in the formulation of the Numerical Weather Prediction (NWP) model, and errors in the formulation of the dispersion model.

In fall 1994 two long range tracer experiments were conducted in Europe over a distance of 1,800 km. For 12 hours an inert, non-depositing tracer was released at Rennes, Brittany, France. The releases took place at the surface and the tracer was sampled at 168 stations throughout Europe. The sampling stations were run by the National Meteorological Services and the whole experiment was sponsored by the EC, the WMO and the IAEA. 24 Institutions took part in the real-time forecasting of the cloud evolution with 28 long range dispersion models. They simulated surface concentration evolution at the locations where the tracer was sampled. The results of these calculations were compared for both experiments with the measurements, using a number of statistical parameters. The results of this comparison (see Archer et al., 1995 for details) indicated for the first release that a limited group of models (7–8) were capable to obtain a good reproduction of the cloud displacement throughout Europe for any time intervals between 24 to 60 hours after the release start. Large differences were however found when examining the predicted tracer concentration at a particular location. In this paper some results for the second release are also presented, similarities and differences with the first are evidenced and tentative explanations are proposed for the differences in model performance.

Until the definitive paper of Thomson (1987), arguably the most pressing problem in Lagrangian particle modelling was which form of the Langevin equation should be used for inhomogeneous, non-stationary, non Gaussian turbulence to ensure well-mixedness, i. e. to ensure that particle accumulations did not occur in regions of low turbulence. Since that time, the major topics of research in Lagrangian dispersion modelling have been associated with convectively unstable conditions. This review paper focuses on two recent topics: (1) the closure problem associated with specification of the probability density function (PDF) for vertical turbulent fluctuations, and (2) the most appropriate boundary conditions to apply at the ground and at the top of the convectively mixed layer, particularly when simulating the fumigation1 process. The reviewed studies emphasise the importance of turbulence and concentration data from a laboratory saline water tank (Hibberd and Sawford, 1994) for testing the closure schemes and the various methods for incorporating entrainment processes into stochastic models.

The purpose of this paper is to test the ability of two quite different models to simulate the combined spatial and temporal variability of Thermal Internal Boundary Layer (TIBL) depth and Mixed Layer Depth (MLD) in the complex terrain and coastline of the Lower Fraser Valley (LFV) of British Columbia, Canada during the course of one day. The models (the simple applied model of Gryning and Batchvarova (1996), and the Colorado State University Regional Atmospheric Modelling System (CSU-RAMS) described by Pielke et al. (1992)). will be tested by comparison with data gathered during a field study (called Pacific ‘93) of photochemical pollution in the LFV. The data utilized here are drawn from tethered balloon flights, free flying balloon ascents, and downlooking lidar operated from an aircraft flown at roughly 3500m ASL.

During the last several decades the surface frictional processes in the shear-free convective boundary layer (CBL) are considered conceptually in the spirit of the Prandtl (1932) theory of free convection, implying the ideas of (i) universal chaotic turbulence and (ii) local correspondence between turbulent fluxes and mean gradients. Accordingly the fluxes of heat and water vapour in the atmospheric surface layer are parameterized disregarding gross features of the CBL. Conventional practical tools are either the Monin-Obukhov similarity theory or simple downgradient turbulence closure models.

Of the planetary boundary layer (PBL) sublayers defined by Stull (1988, pp. 10–11), the nighttime residual layer (RL) may be the least understood. One reason for this is that, unlike the daytime convective boundary layer and surface based nocturnal boundary layer, the RL is in general decoupled from the surface. Consequently, turbulence in the RL does not scale with surface turbulent quantities, and derivation of accurate similarity relationships is difficult, if not impossible. Although qualitative success has been achieved in relating RL turbulence to the gradient Richardson number (Ri) (Mahrt et al., 1979; Lenschow et al., 1987; Kim and Mahrt, 1992), oftentimes these scalings quantitatively fail due to overly coarse vertical differencing in computing Ri (Padman and Jones, 1985). Study of RL turbulence is further complicated by the wide variety of atmospheric forcing mechanisms present during the night. Since nighttime turbulent intensities are small, forcing by radiation, baroclinity, topographic drainage flows, and gravity waves can become important. Variability in each of these leads to a seemingly infinite number of possible RL turbulent time/height structures (e. g., André et al., 1978; Garratt, 1985; Lenschow et al., 1978; Heilman and Takle, 1991; Weber and Kurzeja, 1991).

An algorithm for size-distributed atmospheric aerosols designed for the Northern Aerosol Regional Climate Model [NARCM] is applied to three versions of the Canadian climate models: GCM, RCM and FIZ-C/LCM. It incorporates the processes of aerosol generation, diffusive transport, transformation and removal as a function of particle size to simulate global and regional sea-salt aerosol spatial and temporal distributions. A comparison was made between observations and model predictions of sea-salt. Size-resolved aerosol properties such as transport patterns, fluxes and removals are obtained from the simulations. Since the sea-salt generation term is relatively well quantified, the comparison ensures that a reasonable parameterization of removal and transport schemes is used.

In recent years there has been a revival of interest in mercury as an atmospheric pollutant on local, regional and global geographical scales. An important indicator of increasing interest regarding mercury species in Europe’s atmosphere is the recent decision of the UN-ECE to prepare a protocol for heavy metals and persistent organic pollutants under its co-operative programme for monitoring and evaluation of the long range transmission of air pollutants (EMEP) with mercury as a priority substance. In North America, the 1990 U.S. Clean Air Act Amendments have identified mercury as one of the trace substances listed in the legislation as “hazardous air pollutants” because of its potentially significant effects on ecosystems and human health.

For the last fifteen years, the Office of Research and Development (ORD) of the U. S. Environmental Protection Agency (EPA) has been developing three-dimensional Eulerian based air quality models (AQMs) to study air quality problems, such as urban and regional tropospheric ozone and regional acid deposition. These AQMs simulate comprehensively atmospheric processes such as chemical transformations, transports, and removal of pollutants and their precursors. Model application experience with second generation air quality modeling systems has revealed several shortcomings such as slow execution speed, difficulty in implementing improved science algorithms in the model, and complexity in data exchange among system submodels. Byun et al. (1995) listed some of the shortcomings of the present AQM modeling systems in detail.

The CFD model CHENSI, based on the k-e two equation closure, including thermal buoyancy and scalar transport-diffusion, is first used to simulate the experiment of Cadle et al. (1976), allowing to deduce the turbulence produced by car motion. Then, a heuristic study of the streets capacity to ventilate the pollutants emitted by traffic, as a function of their geometrical and thermal properties, is presented from simulations of the flow and turbulent fields, the pollutant concentrations within the streets, and their fluxes to the atmosphere. Isothermal and non-isothermal simulations are presented, in the case when the street geometry corresponds in isothermal conditions to Oke’s (1988) skimming flow type. The wall heating is shown to induce different types of flow resulting in different pollutant distributions and net fluxes. The influence of the turbulence produced by cars on the pollutant dispersion is shown.

As a new constituent of the European Zooming Model (EZM) system, the multilayer model MUSE is designed to serve as an efficient tool for simulating transport and transformation of air pollutants in the urban scale and thereby in supporting local scale air quality management in the most cost effective way.Comparison of simulation results achieved with MUSE with corresponding results of the validated three-dimensional photochemical dispersion model MARS reveals that the model MUSE is capable of reproducing the spatial and diurnal variation of the major photochemical air pollutants.In order to investigate the effect of the chemical reaction mechanism on air quality predictions, three different reaction mechanisms ranging from the compact mechanism KOREM to the comprehensive mechanisms EMEP and RACM are compared. The latter mechanism is a revised version of the RADM2 mechanism, the improvement mainly focussing on the description of the RO₂ chemistry and biogenic emissions.The intercomparison reveals that despite of similar predicted ozone concentrations, the chemical mechanisms are still performing differently in many aspects. Thus, the choice of a suitable chemical reaction mechanism is mainly depending on the accuracy of the emission inventory as well as on the available computer memory and CPU time.

The Puff-Particle Model (PPM) combines the advantages of both, puff and particle dispersion models. In short, in this approach the centre of mass of each puff is moved along a ‘particle trajectory’, so trying to mimic the quickly changing turbulent flow field. However, particle models account for dispersion of turbulent eddies of all sizes (1 -particle statistics, absolute dispersion) while puff models use relative dispersion to describe the puff growth. Therefore, on combining these two approaches as described above, the dispersing effect of small eddies (smaller than approximately the puff’s size) is accounted for twice. A method is therefore presented to correctly simulate the relative dispersion of puffs within the framework of the PPM. It is based on removing the effect of the high-frequency part of the spectrum when using a ‘particle trajectory’ as the trajectory of the puff centre. It is shown on the basis of tracer data, that the correct treatment and interpretation of the two contributions to the dispersion process is crucial for reproducing experimental results to a good correspondence.

Urban vegetation can directly and indirectly affect local and regional air quality by altering the urban atmospheric environment. Trees affect local air temperature by transpiring water through their leaves, by blocking solar radiation (tree shade), which reduces radiation absorption and heat storage by various anthropogenic surfaces (e. g., buildings, roads), and by altering wind characteristics that affect air dispersion. During the summertime, trees predominantly reduce local air temperatures, but may increase within- and below-canopy air temperature due to reduced turbulent exchange with above-canopy air (Heisler et al., 1995). Reduced air temperature due to trees can improve air quality because the emission of many pollutants and/or precursor chemicals are temperature dependent. Decreased air temperature can also reduce ozone (O₃) formation (Cardelino and Chameides, 1990).

Between 1985 and 1991, the Chemistry of High Elevation Fog (CHEF) experiment was conducted on three mountains in southern Quebec, Canada. The CHEF project was linked with the Mountain Cloud Chemistry Project (MCCP) in the Appalachian Mountains of the USA. Measurements were made of liquid water content (LWC) as well as standard meteorological parameters (temperature, pressure, relative humidity, wind speed, wind direction and precipitation).

Burning is a locally, regionally, and globally important biospheric phenomenon. Globally, biomass burning is a major contributor of greenhouse gases and particulate matter to the atmosphere. Andreae (1991) estimated a value of 40% for the annual contribution of biomass burning to the total annual greenhouse gases emissions. In Portugal, the contribution of forest fire’s emissions to the total greenhouse gases emissions could reach about 10% (Miranda et al., 1994a).

Simulation of at mospheric diffusion in vertically inhomogeneous skewed turbulence, which is typical for the daytime boundary layer, is investigated. Particle trajectories are simulated by using the Langevm equation. The main objective is to introduce a formulation of the probability density function (PDF) of the vertical velocity fluctuation based on a fourth order expansion after Hermite polynomials (Gram-Charlier expansion). The statistics of the turbulence are introduced by forcing the moments of the Gram-Charlier expansion to equal the parametrized moments of atmospheric turbulence.

The objective of this work is to derive a model, simple yet effective, for the influence of microscale turbulence on nonlinear chemical reactions in a turbulent reactive plumes, such as the photochemical reactions between nitric oxides and ozone.

Plumes emitted from stacks are usually buoyant due to temperature differences to the ambient air. Such buoyancy effects may influence strongly the turbulent mixing of the plume with the ambient air. This is essential for reactive plumes, because the mixing of compounds distributed in the ambient air and the plume determines the occurrence of reactions. Furthermore, changes in the turbulent mixing of plume and ambient flow may affect considerably the spatial patterns of the distribution of tracers and consequently their ground concentration.

The application of comprehensive air quality models, such as the Urban Airshed Model, is a computationally demanding exercise that has to be preceded by the compilation of extensive input data sets for emissions, meteorology, and land use for the domain of interest. These computational requirements discourage the type of numerical experimentation required for a thorough analysis of emission reduction strategies.

ADMS-Urban is based on the Atmospheric Dispersion Modelling System (ADMS), a model developed in the UK by Cambridge Environmental Research Consultants, in collaboration with National Power and the Meteorological Office. Sponsors of the model include the UK regulatory agencies (Environment Agency and Health and Safety Executive). The original industrial source version of ADMS has already been well described elsewhere (Hunt et al., 1988 and Carruthers et al., 1992) and has been the subject of extensive validation (Carruthers et al., 1995 and Carruthers et al., 1997).

Airborne particulate matter is abundant in the atmosphere, with a typical spatial variability which reflects the relatively short atmospheric residence time on the order of a week. Natural aerosol particles originate from wind-blown dust in continental areas, from sea-spray over oceans, and from gas-to-particle conversions in the air. Man-made aerosol particles is mainly produced by gas-to-particle conversions, except for local scale dispersion of giant particles which is not included in our work.

Computational resources usually limit the application of three-dimensional models for climate simulation as well as simulation of transport, chemical transformation and deposition of atmospheric gases and particles with respect to horizontal resolution. But many features contributing to regional climate patterns and processes involved in the transformation of atmospheric pollutants occur on spatial scales which cannot be resolved by current global or even regional models. One approach to regionalize coarse grid numerical results is the so called ‘nesting’ technique: large scale phenomena are simulated by coarse grid models and the results are used to provide boundary conditions for a high resolution mesoscale model simulation over the region of interest. Computational resources thus permit a higher resolution for the limited area model and, therefore, a more accurate description of topographie and turbulent processes in the planetary boundary layer. A first development and application of the nesting procedure to climate simulation is described in Dickinson et al. (1989), a first nested grid mesoscale atmospheric chemistry model has been presented by Pleim et al. (1991).

A comprehensive atmospheric dispersion modelling system, designed for real-time assessment of nuclear accidental releases from local to European scale, has been established by integrating a number of existing preprocessors, wind, turbulence, and dispersion models together with on-line available meteorology. The resulting dispersion system serves the realtime on-line decision support system for nuclear emergencies RODOS (Ehrhardt 1996; Kelly et al., 1996; Ehrhardt et al., 1997) with a system-integrated atmospheric dispersion module. This module is called met-rodos, Mikkelsen et al. (1997).

A major, cooperative research project will be completed in 1997 from which an improved understanding will be gained about the dispersion of accidental, dense gas releases at industrial sites (i. e., high surface roughness) during low-wind stable meteorological conditions. The plans for this project were presented by Hanna and Steinberg (1995). Most previous research was limited to releases over smooth surfaces in nearly-neutral conditions (Hanna et al., 1993).

Increasingly, decisions taken for reducing urban pollution are based on some form of numerical modelling. Although the basic mechanisms for the formation of urban pollution are relatively well known, at least qualitatively, the evaluation of control strategies require accurate quantitative modelling of certain scenario, based on well documented past situations. This is the type of modelling that we will discuss here and which is quite different from the operational prediction of pollution conditions, for which the time constraint is crucial and the modelling usually simpler (and often statistical).

Urban air pollution continues to be a problem worldwide, and there is a critical need to develop cost-effective control strategies. Current strategies are designed using air quality models that describe the formation and transport of photochemical pollutants. Unfortunately, the emissions inventories that are used in airshed modeling and control strategy design have been widely underestimated. New methods are needed to improve the quality of emissions inputs. One approach is to solve the inverse problem using existing ambient data and a photochemical urban airshed model to determine the emission field. However, the high dimensionality of spatially and temporally resolved emissions fields proves to be the primary obstacle in solving this problem.

Concentrations of atmospheric trace constituents in the atmospheric boundary layer (ABL) are strongly affected by the meteorological conditions. One of the most important parameters to characterize the dispersion potential of the ABL is the mixing height (MH). In dispersion models, the MH is a key parameter needed to determine the turbulent domain in which dispersion takes place or as a scaling parameter to describe the vertical profiles of ABL-variables.

Mathematical models of atmospheric processes are a complicated matter for Quality Assurance (QA). As well as for measurements, QA procedure is oriented to numerical evaluation of the model precision, potential character of distortions and possibilities of their reduction. The procedure should not answer the question “how good this model is”. Normally after obtaining an extensive numerical information about the model quality, investigator takes a decision whether current model meets the requirements of a particular task.

In recent years comprehensive models have been used as an increasingly important tool for the understanding of “source-receptor” relationships of various atmospheric pollutants and for developing and testing emission control strategies. However one must realize that each process included in a comprehensive model can potentially introduce uncertainties in model results. Model uncertainties can arise from the formulation and numerical representation of these processes as well as from uncertainties in the required input parameters. The understanding of these model uncertainties is important in the process of model development, model evaluation and model application.

This paper describes the development of a modelling system for predicting the emissions, dispersion and chemical transformation of nitrogen oxides in an urban area. The system takes into account of all source categories, including stationary point and area sources, and vehicular sources. We also present a statistical comparison of predicted results and measured concentrations.

Several models exist in Europe aiming at quantifying the transport, transformation and deposition of air pollutants. The structure and complexity of the models vary and often depend on the scope of the applications. Within the EMEP-programme a focus has been on quantifying long term loads of acidity to the European Environment and consequently the relative simple two-dimensional Acid Rain model is employed (see Eliassen and Saltbones, 1983; Barret et al., 1995). Under the EUROTRAC programme the coupled acid rain and photochemical EURAD-model (European Regional Air Pollution Dispersion model) has been developed (Hass et al., 1993, Chang et al., 1987). The purpose of the EURAD model has been to investigate the processes leading to the formation of acidifying components and ozone on the European scale on a relatively short time-scale (days to weeks).

It is a well known fact, based on observations and modelling studies, that there are large spatial gradients over Europe in ozone characteristics. These differences are caused by the differences in strength of the phenomena which determine the ozone concentration at a specific location: the precursor emissions and the resulting chemical production, the meteorological situation and the dry deposition. Ozone measurements show the differences in ozone characteristics, but do not reveal the reasons for the differences.

Effective air-quality studies at the regional-mesoscale range require a good understanding of the main circulatory patterns of air pollutants within the study region. When the mesoscale conditions prevail, circulatory patterns will be mainly determined by the characteristics of the terrain in the area. A region’s unique orography often requires the development of unique air-quality squemes at a regional scale. That is evidenced in the frequent fact that air-quality remediation schemes developed for one region often prove useless when applied to other areas.

Clouds play an important role in our environment. Apart from their radiative properties, this is due to the fact that they capture, transport and modify atmospheric pollutants. To understand their role in our climate is a long-standing aim. Concerning the pure liquid clouds the work has already considerably advanced, as there exist numerous models that study, e. g., the fate of sulfur, ozone or aerosol scavenged by clouds (e. g. Hall, 1980; Chaumerliac et al., 1987 a, b; Flossmann et al., 1988; Flossmann, 1991). For temperature regions from −0°C to −20°C, however, ice particles coexist with liquid water drops in the clouds and, thus, interfere in the processing of pollutants. Concerning these so-called mixed phase clouds our understanding is less advanced. There exist few model approaches which, however, lack from the fact that little is known on the fate of scavenged material in drops once the drops freeze. In order to shed light on this problem the European project CIME (Cloud Ice Mountain Experiment) has been initiated to study this problem. To support this experiment, dynamical microphysical and chemical models have been prepared which will in a later phase be used to interpret and generalize the findings.

In this paper we present the outcomes of a meteorological and a photochemical smog simulation for the city of Perth, Australia. These simulations are designed to represent the conditions observed during the Perth Photochemical Smog Study (PPSS; PPSS 1996). Conducted during the summer of 1994, the study yielded an extensive data base suitable for the validation of both meteorological and photochemical models.

Since July 1995 German Law mandates vehicles without catalytic converters to be banned from traffic when measured ozone mixing ratios exceed 120 ppb and will be likely to exceed this value the day after. Therefore, daily ozone forecasts are needed. In order to replace currently used statistical models the DWD develops chemistry and transport models. One important restriction to be met is a time limit of 2 hours CPU-time for a three-day forecast on the DWD computers. On the Cray YMP4 available until this year this restricts the DWD to use a trajectory box model. A 3-D Eulerian model can only be used after an extensive up-grade of computer power which has just been started and will be continued during the next six years. The trajectory box model provides the additional advantage of giving insight into the couplings of emissions, transport, and chemistry due to its simple design. Modules tested in the box model could be implemented in the 3-D model later on.

Chiba prefecture is located in the eastern side of Tokyo Metropolitan Area(TMA), and it is adjacent to the city of Tokyo. In this area, air pollution is a serious environmental issue, which is caused by the pollutants emitted within and around the area. In order to prevent air pollution, a new Air Pollution Monitoring and Management System was introduced by the Chiba prefectural government in April, 1996. This system consists of real-time ambient air quality and source monitoring, data processing system and an air quality forecasting system. Almost 130 air monitoring stations transmit hourly air quality and wind data to the system center. Sixty-five major industries also transmit hourly emission intensity and related source data. The advanced air quality forecasting system can predict future ambient levels of photochemical oxidant(Ox) and nitrogen dioxide(NO₂) up to 48-hours based on the predicted meteorological conditions. Photochemical air pollution warnings are issued when the hourly values of Ox exceed 120ppb and the adverse meteorological conditions are predicted to sustain these levels. When air pollution forecasting warning is issued based on the results obtained from the forecasting system, the emergency information are sent rapidly to the municipalities and to operating factories in order to mitigate the effects of high air pollution.

Project MOHAVE is a large monitoring, modeling, and data analysis study whose main goal is to assess the effects of the Mohave power plant (MPP), a large coal-fired facility in southern Nevada, upon visibility in the southwestern United States, in particular at Grand Canyon National Park (Pitchford and Green, 1997). Additional goals of Project MOHAVE include estimating the effects of other sources upon visibility in the southwestern United States, and evaluating atmospheric transport and dispersion models, and receptors models. One of the key design features of Project MOHAVE was the release of perfluorocarbon tracers (PFTs) at the Mohave power plant and other locations during a 30 day winter intensive study and a 50 day summer intensive study. Tracer and aerosol measurements were made at over 30 locations (mostly 24-hour average concentrations); upper air measurements were made with radar wind profilers, sodars, and rawinsondes; optical monitoring included transmissometers, nephelometers, and time-lapse photography.

This paper addresses the problem of uncertainty analysis of the predictions of an atmospheric dispersion model. The proposed method uses Bayesian conditioning of the predictions on the available observations and enables consideration of the uncertainties which influence model predictions such as uncertainty of observations and limitations of the model. As an illustration, the proposed methodology is applied to a simple, Gaussian short range atmospheric model (Clarke, 1979). This is commonly referred to the R91 methodology, after the name of the report on which it is based, and shall be referred to as such throughout this paper. The choice of the model was dictated by its availability, simplicity and widespread use. However, the proposed approach has much wider application and can be used for different air dispersion models such as UKADMS (Carruthers et al., 1992) or any other model with comparatively short times of computations and for which observations are available. The simplifications used in dispersion modelling are very substantial and comprehensive data are very limited. This means that it is important to make best possible use of the information which is available -in the form of distributions, value ranges, specialist opinions and common sense. Also many uses in nuclear safety rely on the concept of risk and it is important that uncertainty modelling is applied at all.

Realizing the need for a dispersion model which explicitly considers non-steady state emissions and arbitrarily complex, space-time varying meteorological conditions, and the desirability of having a model which can yield the probability distribution function (PDF) of concentrations rather than simply the ensemble mean, the German Federal Environmental Agency (UBA) funded the development and preliminary testing of an advanced atmospheric dispersion model, referred to as the Kinematic-Simulation-Particle (KSP) Model. The KSP modeling system (Yamartino et al.; 1993–6, and Strimaitis et al., 1995) includes dry and wet removal processes, many other of the processes (e. g., stack and building downwash, plume rise, partial lid penetration) included in the CALPUFF (Scire et al.; 1990, 1995) puff model, and computes the space-time average and percentile concentrations required by German and EC legislation; however, KSP also has the intrinsic capacity to predict the short-term concentration fluctuations that are critical to the assessment of odor and hazardous chemical exposure problems. A key model input involves the 3-d gridded mesoscale wind fields and accompanying 2-d fields specifying boundary layer quantities. These data may be provided by an external wind field/PBL model, or, in the case where only single-point measures of wind and stability are available, an internal flow and turbulence generator creates a vertical profile which emulates the dispersion conditions specified in the German regulatory guideline (TA-Luft). The current KSP model also accesses data produced by the CALMET diagnostic model (Scire et al., 1990; and U.S. EPA, 1995), and an interface to the GESIMA prognostic model has been developed by Reimer et al. (1994) that allows the modeled TKE budgets to be used in the computation of turbulence fields. Expansion of this interface to permit METRAS (Schluenzen et al., 1994) generated winds and turbulence to be utilized is also envisioned.

At the previous ITM in Baltimore, the author presented a paper entitled “Toward the establishment of a common framework for model evaluation” (Olesen, 1996). It described various ongoing activities aimed to improve model evaluation methods. Further, it discussed problems about the way that evaluation is normally practised, and suggested ways to improve matters.

It is known (Thomson, 1987) that Ito’s type stochastic models (LS) satisfy the well-mixed condition and hence are physically consistent. An Eulerian probability density function (PDF) of the turbulent velocities, as close as possible to the actual atmospheric PDF, must be prescribed in order to specify the model. Unfortunately these models have a unique solution in one-dimension only (Sawford and Guest, 1988). For this reason the present study will focus on one-dimensional diffusion simulation.

An outlet portal of a city road tunnel is one of crucial point sources of pollutants to be taken into account when modeling dispersion of pollutants within a city. Namely in case of local scale modeling a correct knowledge of velocity and concentration profiles at the outlet portals may play a non-negligible role. And these profiles are established in a dominant way by the traffic density and speed. In the first part of the paper we are going to present a new Lagrangian - Eulerian approach to moving vehicles througout a traffic tunnel and as a result to give values of flow rate entrained into and out of the tunnel and velocity and concentration profiles at the outlet tunnel portal. In the second part of the paper, dispersion of pollutants from the Plabutsch highway tunnel in Graz (Austria) is predicted for various boundary conditions.

Because of widespread concern about the effects of the exposure of urban populations to a large number of air pollutants, a method allowing a quantitative evaluation of the number of excess cancer cases caused by individual substances is of great interest. It makes it possible to compare the effects of different compounds, more difficult to distinguish in epidemiological studies, and it can be a useful tool to evaluate the impact of possible abatement strategies.

Health effects in relation to airborne suspended particulate matter (SPM) receive nowadays much attention in Europe and North America. The inhalable fraction of SPM consists mainly of particles with sizes below 10 μm (PM10) and is thought to be originating from antropogenic activities mainly. These activities lead partly to direct particle emissions and partly to the production of secondary particles such as ammonia-sulphate and ammonia-nitrate.

The paper presents the results of research work started at RISØ National Laboratory, Denmark and continued at “Petroleum-Gas” University of Ploiesti, Romania regarding application of the real-time episode dispersion model RIMPUFF to the Ploiesti - Prahova industrial area with the motivation to understand and quantify severe air pollution scenarios of ground level air concentrations resulting from the different sources of petro-chemical origin located in the industrial area.

The land use classification is usually produced by hand using different detailed maps which have been produced by aerial picturing or direct processes. The use of satellite images has been incorporated widely in the recent years. We present a land use classification which has been made by using a LANDSAT-5 image (October. 1987). The ANA (San José et al., 1994) mesoscale air quality system is used to study and compare the results on ozone concentrations at surface level when the automatic land use classification by using the satellite image is used and when the handmade classification is used. We present results on four different Madrid monitoring stations for ozone for both land use classifications. Two important results can be concluded: the changes on the ozone concentrations due to different land use classifications are important and can lead to different and wrong conclusions. The satellite information when applied the automatic land use classification seems to be more precise and accurate than the handmade land use classification. A full and complete statistical study should be done in the future however computational power limitations are an important obstacle for such a experiment.

The aim of this study is twofold: First, to gain perspective for the assessment of the spatial distribution of one of the air pollutants, sulfur dioxide, in the region by the use of a statistical modelling scheme, kriging; and second, to show the decrease of sulfur dioxide concentrations over İstanbul by elaborating on the reasons of this decrease in terms of the meteorological factors.

Estimates of fluctuations in concentration resulting from multiple sources are complicated by the non-independence between the contributions to the concentration from each of the sources. Such fluctuations are important for toxic, inflammable and odorous releases and, with new standards for common air pollutants such as SO₂ being introduced which are based on averaging times less than 1 hour, are becoming important for routine emissions too.

The European Tracer Experiment (ETEX) is jointly organised by the European Commission, the International Atomic Energy Agency and the World Meteorological Organization. The project was established to evaluate the validity of long-range transport models for real time application in emergency management and to assemble a database which will allow the evaluation of long-range atmospheric dispersion models in general (Klug et al., 1993) ETEX main objectives were to conduct a European-scale tracer experiment, and test the capability of meteorological institutes responsible for emergency response to forecast in real time atmospheric dispersion (Girardi et al., 1997). Following this, the Atmospheric Transport Model Evaluation Study II (ATMES II) will simulate the tracer experiment with the same meteorological input data for all models.

In case of nuclear or chemical accidents, local authorities have to be rapidly informed on pollutant immissions which could harm the population. They need to know the extent of affected area, appearance time of the pollutant, stagnation duration and levels of forecasted concentrations.

Inverse dispersion modelling means to derive information such as source strengths of emissions from measured concentration and/or deposition data of trace constituents, using a dispersion model. If source-receptor relationships are linear, they can be determined from a single model run, and the inversion corresponds to solving a linear system of equations. Examples for this approach are Enting and Newsam (1990), Brown (1993), and Hein et al. (1996). Non-linear source-receptor relationships require an iterative solution using an adjoint model. The calculation of large source-receptor matrices is very time consuming. Therefore, simple trajectories have been used widely to derive information on sources from pollution measurements. So-called potential source contribution functions have been calculated, e. g., by Zeng and Hopke (1989); similar techniques were employed Seibert and Jost (1994) and Stohl (1996). These are statistical methods, not based on dispersion models, and not exploiting the full information contained in the data sets. However, trajectories can be viewed as primitive Lagrangian dispersion models, and source-receptor matrices can be derived from them for use in formal inverse modelling. One has to be aware, however, that systematic shortcomings (e. g., neglection of precipitation scavenging or vertical exchange) will lead to systematic errors in the results, a problem shared with the statistical approach.

The Gaussian plume model concept is still important for estimating ground-level concentrations due to tall stack emissions and is usually suitable for regulatory use in air quality models. Improved dispersion algorithms in updated Gaussian models calculate the dispersion parameters σ_y and σ_z in terms of distinct scaling parameters for turbulence. In this work, we use the convective similarity theory and Taylor’s statistical diffusion theory to derive a general expression for the dispersion parameters in a planetary boundary layer dominated by convection. The novelty is use Taylor’s theory to construct a relation for the dispersion parameters that is described only in terms of the universal characteristics of the turbulent field in a convective boundary layer (CBL), eliminating in this manner, the necessity of dispersion parameters fitting from diffusion experiments.

Last years, a great number of dispersion models were developed. They possess different features and need different computer resources. Recently, the PC-oriented Eulerian multi-layer model EMAP (Eulerian Model for Air Pollution) was created in Bulgaria and applied to different pollution problems. The vertical diffusion block of the model uses a 2nd order implicit scheme including dry deposition as a bottom boundary condition, realised on a non-homogeneous (log-linear) staggered grid. The experiments with EMAP show that, if the concentration at the first computational level is used for calculation of the dry deposition flux, the deposited quantity changes when the height of this level is changed. It is obviously that the roughness level concentration is necessary as to calculate properly dry deposition. It is not possible to have a model level at this height, because roughness usually changes from one grid point to another. On the other hand, because of the steep gradients in the surface layer (SL), many levels must be introduced near the ground for adequate description of pollution profiles. This would increase memory and time requirements without any practical need. Usually, the first computational level is placed high enough above roughness. As a result, a good estimate for the roughness level concentration is necessary, determined on the base of the calculated concentrations. The problem becomes much more complex when surface sources of pollution have to be treated. Such are the processes of evaporation and re-emission of tracer under consideration. A proper parametrization of the diffusion processes in the surface layer can avoid these difficulties. An effective SL parametrization, based on similarity theory, was elaborated and tested by Syrakov and Yordanov (1996a, b). It allows to have the first computational level at the top of this layer. Here, an upgrade of this parametrization is presented, taking into account the presence of a continuous surface source.

The particulate matter suspended in the atmosphere is strongly linked to numerous air pollution problems. These include the -direct and indirect- influences on the radiative budget of the atmosphere, the potential for adverse health effects, the influence of particles on the long-range transport of air pollutants and the interaction of aerosol particles with clouds. Therefore air quality models have to take into account particle formation, transport and deposition with respect to aerosol chemistry as well as aerosol dynamics. Due to the strong interactions between the gas phase and the aerosol phase, the aerosol model has to be fully coupled to a gas phase model. The Modal Aerosol Dynamics modelling technique as a time and memory efficient method provides a suitable approach to realise that purpose even in complex three-dimensional Eulerian models.

A new generation of the atmospheric dispersion models is under development in the USA and Europe (Berkowicz et al., 1986; Perry et al., 1994). These Gaussian models employ parametrizations of dispersion coefficients and other variables with physical parameters rather than with stability categories. Seemingly more direct approach to introduction of physical parameters into dispersion models can be based on solution on the advection-diffusion equation. Results of such an approach are presented in this paper; ways of generalization of the developed flat-terrain model for account of building-downwash effects are also discussed here.

Air pollution forecasting models are a helpful tool for controlling pollution around sources such as large thermal power plants. Recently we developed a neural network - based, short-term SO₂ pollution forecasting model for measuring sites around the Sostanj Thermal Power Plant. The most important problems that should be solved in order to improve the model performance are feature determination and pattern selection. We developed several methods to solve these two problems.

A complex expert system for air pollutant dispersion modeling ScalEx (Scale up Expert) is currently being developed at Faculty of Technology and Metallurgy in Belgrade. The project goal is to develop a powerful software system as an engineering toolbox for modeling air pollutant dispersion under various conditions and terrain configurations. ScalEx is planned to be an integrated software system of the new generation, putting together expert system, advanced database technologies, hypermedia, conventional numeric routines and neural networks (Pocajt, 1996).

The application of a dispersion model of pollutants released from near surface and elevated continuous point sources is presented. The model is based on the bidimensional semiempirical equation, with vertical profiles of wind and eddy diffusivity for the atmospheric boundary layer. The suggested model makes use of a continuous description of the dispersion processes in the different regimes of the atmospheric boundary layer and needs meteorological input parameters that can be estimated from routine measurements1. The model is used to simulate the dispersion of non-buoyant, non-depositing releases from several source heights in a variety of atmospheric stability and surface roughness conditions. The observational data were obtained in tracer experiments carried out at Copenhagen (Denmark)2. Lillestrom (Norway)3, Hanford (USA)4 and Cabauw (The Netherlands)5. The predicted cross-wind integrated concentrations were compared with the observed ones. All concentrations were normalized by source strength. Current quantitative measurements and techniques of model evaluation were obtained and applied6,7.

Soot-like components of atmospheric aerosol is known to be one of an important factor of air pollution. This matter is especially high priority near the sources of industrial pollution and the composition of atmospheric aerosol would be the most quantity here. In contrast to usual background atmospheric aerosol, solar radiation is especially absorbed by soot-like components in passing through the atmosphere.

A meteorological pre-processor can handle input data of various kind: routine meteorological data, data from small meteorological masts of 10 or 30 meters high, heat flux measurements and so on... Does the quality of the pre-processor output depend on the set of input parameters used? This was investigated by comparing a pre-processors output for different input parameters with measurements from a high meteorological tower.

The Gaussian type air pollution dispersion and deposition model AEROPOL is developed at the Tartu Observatory, Estonia (Kaasik and Rõõm, 1997). The multiple reflections with partial adsorption from the underlying surface and capping inversion, as well as the initial rise of a buoyant plume with the inversion penetration, are included. The wet deposition is included considering different features of rain and snow. The concentrations of pollutants, dry and wet deposition fluxes could be computed. The computing code is written in Turbo Pascal and designed for operational use in Windows’95.

We present numerical results of the boundary layer model used for regulatory purposes in Finland, and compare these with measurements from field campaigns and routine measurements. The parameterisation schemes used in the dispersion models of the Finnish Meteorological Institute (FMI) are based on the energy flux method of van Ulden and Holtslag (1985), and the parameterisation of the boundary layer height is based on classical boundary layer models with a separate treatment for convective and stable conditions.

COST Action 710 is a voluntary program of the European Committee for the “Harmonization in the Preprocessing of Meteorological Data for Dispersion Models”. The idea of carrying out an international benchmark exercise was endorsed by the Management Committee of the Action at the end of 1995 year. As responsible scientist for this exercise D. J. Szepesi of Hungary was appointed.

Sometimes there is a need for determination of air pollution levels at remote places far from local sources with rather good spatial distribution evaluation. It is useful to use diffusive sampler where the sorbed amount must be proportional to the ambient concentration of the gas. The gas is transported to the sorbent by molecular diffusion and the diffusion rate is made constant by placing the impregnated filter inside a tube1.

Exceeding of the limit mass rate values admitted at interfering emission sources for the same type of pollutant can determine the growth of concentration level in the imssion area over the admissible limits.