main-content

## Über dieses Buch

The 20th International Technical Meeting on Air Pollution Modelling and Its Application was held in Valencia, Spain, during late 1993. At this conference, a new record of abstracts was submitted, a new record of scientists participated, and a new record of countries was represented. This clearly indicates society's continuous and growing interest in, as well as importance of, the complexities associated with the modelling of air pollution. The conference addressed the following main subjects: integrated regional modelling, global and long-range transport, new modelling developments, accidental releases, and model assess­ ment and verification. In addition, two project-oriented workshops were organized as part of the conference. The many contributing authors and scientists taking active part in the discussions following the papers, have made this proceeding a record of the current status in the field of air pollu­ tion modelling. We want to express our gratitude to their efforts. We also wish to extend our gratitude to the sponsors that made this conference possible. In addition to financial support from NATOjCCMS the conference received contributions from CEAM, the European Asso­ ciation for the Science of Air Pollution, Danish Center for Air Research, and Ris0 National Laboratory. A special grant was given by NATOjCCMS to facilitate attendance of scientists from Central and Eastern Europe. We also wish to express our gratitude to Rosa Salvador and Pilar Zamora of CEAM, who laboriously organized the conference pre-proceedings, and to Anne N0rregaard and Ulla Riis Christiansen of Ris0 National Laboratory, who seved as conference secretariat.

## Inhaltsverzeichnis

### Experimental and Numerical Study of High Ozone Events in the Alpine-Region

In the alpine region as well as in other European areas the ozone concentration appears to be increasing (Warmbt, 1977). During this century the ozone levels in the atmospheric boundary layer have been almost doubled (Volz and Kley, 1988). After Puxbaum et al. (1991) the ozone concentration in the Zillertal (a small valley in the Austrian Alps) has been increased by a factor of 3 since 1950. Beside this long term trend high ozone values of 100 ppb and more are increasingly observed during pollution episodes. These episodes are associated with high pressure systems, when the meteorological condition, in particular warm temperatures, clear skies and low wind speeds, encourage the photochemical formation of ozone. Calms and inversions causing reduced ventilation occur more frequently in mountain valleys than over flat terrain (Dreiseitl and Weber, 1991). Therefore the meteorological conditions in the Alps conduce the ozone production.

Jutta Graf, Hans Schlager, Monika Krautstrunk

### Match: A Mesoscale Atmospheric Dispersion Model and Its Application to Air Pollution Assessments in Sweden

The MATCH (Mesoscale Atmospheric Transport and CHemistry) model has been developed as a tool for air pollution assessment studies on different geographical scales. It has been used as a basis for decision making concerning environmental protection within Sweden or subregions of Sweden.MATCH is an Eulerian atmospheric dispersion model, including physical and chemical processes governing sources, atmospheric transport and sinks of oxidized sulfur and oxidized and reduced nitrogen. In its standard configuration the model has three levels in the vertical, where the second layer is variable in height and follows the mixing height. The mixing height is derived from standard meteorological fields and varies in both space and time.Using the MATCH system, air pollution contributions from different source types like traffic, industry, shipping, farming etc can be obtained. Using a combination of air and precipitation chemistry measurements and the MATCH model, the contribution of air pollution and deposition from long range transport can be quantified in the model region. A first tentative comparison between the EMEP and MATCH calculations indicate that the relations between local/regional contributions and long range transport are similar for deposition of oxidized and reduced nitrogen, but for oxidized sulfur the MATCH system gives a somewhat smaller influence from local/regional sources compared to what is obtained from EMEP.

Christer Persson, Joakim Langner, Lennart Robertson

### Recent Applications of the Rams Meteorological and the Hypact Dispersion Models

The Regional Atmospheric Modeling System (RAMS) developed at Colorado State University is being widely applied in regulatory, operational forecasting and emergency response programs (Pielke et al., 1992). Its two-way nesting capability facilitates use of the model over relatively small regions employing a very fine innermost grid while at the same time treating the influences of inhomogeneous and time variable synoptic fields within which the mesoscale perturbations develop. RAMS has been applied using horizontal mesh sizes (Δx) from as large as 100 km to only 1 meter. The model predicts the basic state variables (U,V,W wind components, pressure, temperature and mixing ratio) as well as a wide variety of derived products, including planetary boundary layer (PBL) depth. RAMS also drives an advanced dispersion code called HYPACT(Hybrid Particle and Concentration Transport). HYPACT disperses emissions from a variety of source types using both Lagrangian particle and Eulerian methodologies (Lyons et al., 1993). Two recent RAMS applications will be summarized.

Walter A. Lyons, Roger A. Pielke, William R. Cotton, Craig J. Tremback, Robert L. Walko, M. Uliasz, Jose Ignacio Ibarra

### Application of the Mesoscale Dispersion Modeling System to Investigation of Air Pollution Transport in Southern Poland

Recent advances in computer technology have opened the door for a broad application of sophisticated numerical models for air pollution dispersion in complex terrain. Very promising opportunities for intensive air quality studies and a real time modeling are provided by Lagrangian particle dispersion models linked to 3-D mesoscale meteorological models (Pielke et al., 1991; Lyons et al., 1993; Uliasz, 1993). A real revolution in mesoscale dispersion applications has been introduced by powerful and affordable workstations which can be dedicated to specific tasks. It allows one not only design and perform short case studies but to use these models for extended periods of time as well. An example of such a computationally intensive application on modern workstations is a project MOHAVE (Uliasz et al., 1993) where daily meteorological and dispersion simulations for the southwestern United States are being performed for a year long study. The modeling methodology developed for the project MOHAVE is being applied to regions in southern Poland with very serious air pollution problems.

Marek Uliasz, Magda Bartochowska, Anna Madany, Henryk Piwkowski, Jan Parfiniewicz, Maciej Rozkrut

### Effects of the Selected Domain in Mesoscale Atmospheric Simulations and Dispersion Calculations

The role of the various-scale circulations (synoptic, regional, sub-regional, mesoscale, microscale) in the dispersion of air pollutants, released from different type of sources, is considered as very important because the 3-D meteorological and turbulence fields show significant temporal variations (e.g. Pielke et al., 1983; Glendening et al., 1986; Segal et al., 1988; Ulrickson and Mass, 1990; Kallos, 1990 among others). This is especially true in areas with significant physiographic variations. Today, allot of mesoscale models are able to provide meteorological and turbulence fields adequate for use in dispersion and/or transformation models (Pielke, 1988; Seigneur, 1989). The combined use of such models is considered as superior from the existing regulatory models because they should provide a more realistic description of the dispersion characteristics and the spatial and temporal variations of the concentration fields of air pollutants on more accurate manner. Despite this fact, the use of such models may lead in erroneous calculations for several reasons. The most important reasons are the mishandling from the models the physical processes taking place under certain circumstances and of course the improper use of them. The improper use may be due to the violations of physical rules which are not necessarily associated with the algorithms and the chosen initial and boundary conditions.

George Kallos, Pavlos Kassomenos

### Summer Episodes of Pollution Dispersion over the Coastal Area of Israel - A Numerical Study

The rising need to control industrial emissions over urban areas in Israel and to monitor the dispersion of pollutants (Graber, 1981) has stimulated extensive experimental and simulation studies in this field (Graber et al., 1984). The dependence of the dispersion of pollutants on mesoscale meteorological conditions was investigated with the aid of a recently developed model of air pollution transport over Israel (Tokar et al., 1993) and a mesometeorological model based on the PSU/NCAR mesoscale numerical model known as MM4 (Anthes et al., 1987). This model is suitable for forecasting flows with characteristic wavelengths of 10 to 2500 km under a variety of meteorological conditions. The models were implemented over a 324×270 km region with the Hadera power plant -the largest in Israel - at its center. Topography was included in the model. The MM4 model with 15 height levels and 6 km horizontal resolution was adapted to fit the Israeli conditions. The pollution transport model was applied for a smaller domain of 108×90 km with 2 km of horizontal grid spacing. The transport model uses the MM4 calculations as its input data.

Joan Goldstein, Yakov Tokar, Yakov Balmor, Ed Glaser, Pinhas Alpert

### Numerical Simulation of Meso-Meteorological Circulations in the Lisbon Region

During the last 5 years, the University of Aveiro has directed significant research towards the analysis of regional air quality and problems related to the transport of airborne pollutants over Portugal (Coutinho et al., 1989; Borrego et al., 1991). One of the main conclusions of these studies was the necessity to consider mesoscale atmospheric circulations in studying atmospheric dispersion patterns in Portugal. Mesoscale meteorological models seem to be the most adequate tool to represent those circulations through the mathematical simulation of atmospheric physical processes.

Miguel Coutinho, Alfredo Rocha, Carlos Borrego

### Problems of Modelling the Long-Range Transport of Reduced Nitrogen

As negotiations on a new sulphur protocol under the UN Economic Commission for Europe reach an advanced stage, increasing attention is being paid to the contribution of nitrogen compounds to acidification. In this context European emissions of nitrogen as ammonia (estimated at almost 8 million tonnes of N over the EMEP map area- EMEP 1993), originating mainly from livestock, are comparable with those as NOx (7 million tonnes as N over the same map area). However as a reactive gas emitted at or near ground level from a distribution of sources, there are special problems in modelling the subsequent dispersion and atmospheric transport out to long distances. There are also large uncertainties about the nature and magnitude of emissions, and the interaction of ammonia with other atmospheric species. This paper illustrates some of these difficulties by application of a Lagrangian model TERN (Transport over Europe of Reduced Nitorgen) to selected episodes.

Helen M. ApSimon, Brian M. Barker, Serpil Kayin

### Coupling the Photochemical, Eulerian Transport and ‘Big-Leaf’ Deposition Modelling in a Three Dimensional Mesoscale Context

The purpose of this paper is to show the most important features of the air dispersion modelling focused on the photochemical, Eulerian transport and deposition modelling. The model we describe will take into account most recent advances on atmospheric photochemistry, it will account with the source and receptor oriented approaches and finally will focus on the deposition processes based on a landuse classification for accounting the canopy resistance effects.

Roberto San José, Luis Rodríguez, Magdalena Palacios, Javier Moreno

### Gem: A Lagrangian Particle Model for the Dispersion of Primary Pollutants in Urban Canyons-Sensitivity Analysis and First Validation Trials

The objective of the work is to develop a model able to describe the microscale concentration field of primary pollutants in urban topography and particulary within street canyons. This kind of topography influences wind and turbulence fields that become very irregular (De Paul and Sheih, 1985; Nakamura and Oke, 1988). So, concentration field presents complex patterns (Hoydysh and Dabberdt, 1988).

Guido Lanzani, Matteo Tamponi

### Verification of Urban Scale Time-Dependent Dispersion Model with Subgrid Elements, in Oslo, Norway

Results from monitoring of air pollution concentrations in cities in Norway have shown that nitrogen dioxide (NO2) is one of the compounds which most often, and to the largest extent, exceeds current air quality guidelines (Hagen, 1992; Larssen, 1993). This is the case both in city streets and in the urban atmosphere in general. In Norway, the highest NO2 concentrations occur during the winter months, in connection with “episodes” with poor dispersion. In the general urban atmosphere, high 24-hour average values are of greatest concern relative to Air Quality Guideline (AQG), while in the street atmosphere, very high peak (hourly) concentrations may be the most important problem.

Steinar Larssen, Knut Erik Grønskei, Frederic Gram, Leif Otto Hagen, Sam-Erik Walker

### Observation and Simulation of Urban-Topography Barrier Effects on Boundary Layer Structure Using the Three-Dimensional TVM/URBMET Model

Urban areas affect prevailing mesoscale and synoptic flow patterns due to a variety of physical processes, including urban heat island induced accelerations, surface roughness induced decelerations, and building barrier effects. Analysis of data collected over New York City (NYC) has shown that the city is capable of significantly altering the speed and/or direction of movement of thunderstorm cells, sea breeze fronts, and synoptic fronts. Analyses of the above effects, plus additional analysis (during periods without such features) of surface flow patterns, surface convergence fields, tetroon vertical velocities, and double theodolite velocity fields all point to the urban barrier effect as the most significant factor altering mesoscale flow over NYC (Bornstein and LeRoy1).

R. Bornstein, P. Thunis, G. Schayes

### Wind Flow and Photochemical Smog in Thessaloniki: Model Results Compared with Observations

The EUMAC Zooming Model is applied to simulate wind flow and pollutant transport and transformation in the Greater Thessaloniki Area (GTA). The simulation results agree fairly well with observations during the Thessaloniki ‘81 Field Measurement Campaign. Both the calculation and the observation show that the ozone levels in the GTA are decisively affected by the sea breeze circulation and the nighttime inversion.

Nicolas Moussiopoulos, Athena Proyou, Peter Sahm

### A Three Dimensional Hemispheric Air Pollution Model Used for the Arctic

There have been a considerably scientific interest of the air pollution in the Arctic. The reason for this is that the arctic area is rather sensitive to the pollution, and the pollution there is a finger print of the air pollution for the whole Northern Hemisphere. In the last 20 years several groups have done measurements in the Arctic (see, for example, AGASP, 1984; Arctic Air Chemistry, 1985, 1989, 1993). All these measurements shows that the air pollution have a great seasonally variation with relative high concentrations during the winter and the early spring with a maximum normally in march and very low concentrations during the summer. The air pollution is also characterized by an episodic nature, because the origin of the pollution is only due to the long-range transport, and a deep vertical distribution, because of the very cold and stable Arctic atmosphere.

Jesper Christensen

### Model Simulation of the Atmospheric Input of Trace Metals into the North, Baltic, Mediterranean and Black Seas

Since the middle of this century, energy generation, industrial production and transportation have caused serious environmental contamination by trace elements including heavy metals. The rate of contamination can vary from place to place as a function of source densities and intensities of heavy metals flux as well as meteorological conditions. Aerosols containing heavy metals can be transported far away from their sources by advection before being deposited.

Joseph Alcamo, László Bozó, Jerzy Bartnicki

### Model of Long-Range Pollutants Transport and Acidity of Precipitation for Baltic Region

In the present paper the investigation of the long-range pollutants transport includes the formation of the acid precipitation. For this purpose a proper combination of models which describe the different stages of this complex phenomenon is realised [1,2].

D. Syrakov, M. Kolarova, D. Perkauskas, K. Senuta, A. Mikelinskene

### Modelling the Long-Range Transport and Deposition of Persistent Organic Pollutants over Europe and Its Surrounding Marine Areas

For a long time it has been recognized that the atmosphere is an important pathway for the transport of pollutants from industrialized and densely populated areas to ecosystems near to or far from these areas. In particular, the important role of the atmospheric pathway in the pollution of Europe’s marginal seas by land-based sources was identified more than a decade ago (Cambray et al.1979; Van Aalst et al.,1982).,In case of persistent organic pollutants (POPs) transport through the atmosphere may even lead to hemispheric or global distribution of the pollutants.

J. A. van Jaarsveld, W. A. J. van Pul, F. A. A. M. de Leeuw

### A New Model for Routine Calculation of Long-Range Transport of Air Pollutants

At the Western Meteorological Centre (MSC-W) for the European Monitoring and Evaluation Programme (EMEP), LRT-models have been applied for many years based on official estimates of emissions and real-time meteorological data (Eliassen and Saltbones, 1983; Hoy et al., 1988; Iversen, 1990 and 1993; Sandnes, 1993; Simpson, 1993). So far, only “1.5-level” models have been used with a Lagrangian estimation of horizontal transport. Although showing considerable skill for long-term averages, several shortcomings connected with the crude vertical treatment and coarse horizontal mesh width of 150 km are quite evident. As the work under the UN/ECE Convention for Long-Range Transboundary Air Pollution increases the demand for detailed information about atmospheric transport, these model weaknesses should be diminished as far as possible, without destroying the balance between the level of detail and the need for long-term calculations covering several years. Effects introduced by steep topography such as the Alps are only coarsely resolved by the 150 km grid, and calculations of ecosystem-specific dry deposition also require as good a horizontal resolution as possible. Furthermore, EMEP is supposed to keep budgets of transboundary transport, and these are quite uncertain for countries only resolved by very few grid-squares. A calculation of the subgrid-scale deposition after emission release involving time-resolved characteristics of the atmospheric boundary layer (ABL), is best taken care of by a certain minimum vertical resolution of the ABL. A significant part of the wet deposition in regions about 1000 km away from the main emission sources probably stems from free troposphere just above the ABL, hence some vertical resolution of the lowermost portion of the free troposphere above ABL is also required. This is even more important for models for ozone for which the free troposphere in some regions acts as a reservoir. A scale analysis of the budget-equations of atmospheric constituents on regional scales (L ~ 250–2500 km) shows that horizontal and vertical advection terms are of the same order of magnitude.

Trond Iversen, Erik Berge

### Atmospheric Mercury Species over Europe. Model Calculations and Comparison with Observations from the Nordic Air and Precipitation Network for 1987 and 1988

Mercury exists in the atmosphere in different physical and chemical forms, whose properties and interactions with their surroundings determine its transport and transformations, as well as its removal mechanisms such as wet and dry deposition to the earth’s surface. It is now generally accepted, that more than 90 % of atmospheric mercury is in the vapor phase while the remainder is associated with particulate matter. The chemical speciation of mercury in the atmosphere is still under discussion although there is evidence for the total predominance of elemental mercury. Recent data indicate that, besides elemental mercury, methylmercury species are present in ambient air in minor quantities. In the emissions from high-temperature combustion facilities without flue gas cleaning systems divalent inorganic mercury compounds in gaseous or particulate form have been identified, but far from sources these species are at very low or undetectable levels.

G. Petersen, Å Iverfeldt, J. Munthe

### A Boundary Layer Parameterization for Global Dispersion Models

In this note we present the essentials of a simple algorithm which simulates the main features of a dry, horizontally homogeneous atmospheric boundary layer, a more comprehensive presentation of these ideas are found in Nappo and van Dop (1993). The algorithm or parameterization scheme is designed only for use in global-scale air chemistry and transport models. Details of dispersion in the convective and stable boundary layers are not addressed; instead, we specify the concentration profiles for these PBL conditions. It is this simplicity that makes this algorithm attractive for parameterizing subgrid-scale boundary-layer fluxes in numerical global-scale transport/chemistry models. The algorithm was developed for the Moguntia (Latin word for Mainz, FRG) global chemistry model (Zimmerman, 1987) which uses monthly-averaged meteorological input data. The algorithm can be used over land and sea surfaces provided grid-cell averaged convective mixing heights and stable PBL heights are given. Trace-gas removal and surface concentrations, including diurnal variations, are satisfactorily described. However, the effects of isolated thunderstorms and frontal systems are not parameterized.

Carmen J. Nappo, Han van Dop

### Analysis of Mid-Tropospheric Carbon Monoxide Data Using a Three-Dimensional Global Atmospheric Chemistry Numerical Model

Carbon monoxide is an important atmospheric trace species. It has long been recognized as a major contributor to urban air quality and in high concentrations is known to adversely affect health (Seinfeld, 1986). CO is the third most abundant carbon-containing species in the atmosphere and its reaction with hydroxyl radical (OH) represents a 2000–3000 Tg/yr (1 Tg = 1012 g) source of carbon dioxide. On a global basis, through its reaction with OH, CO plays a significant role in the troposphere’s overall oxidative capacity (Crutzen and Zimmerman, 1991). Furthermore, depending on the local abundance of nitrogen oxides, CO can participate in reactions that either increase or decrease the formation of tropospheric ozone (Logan et al., 1981).

Richard C. Easter, Rick D. Saylor, Elaine G. Chapman

### The Influence of Aqueous Phase Chemistry on Long Term Ozone Concentrations over Europe

Clouds affect tropospheric ozone levels through aqueous phase chemical reactions and by altering radiation transfer and mixing. Cloud droplets can act as important sinks for ozone and ozone precursors such as oxidized hydrocarbons and NOΧ (NO+NO2) through aqueous oxidation and/or wet removal. Moreover the presence of liquid water separates soluble and insoluble species which again can effect the ozone formation. For instance NO, which is relatively insoluble will be separated in the presence of cloudwater from the relatively soluble HO2, reducing effectively the formation of ozone in the gasphase.

J. Matthijsen, P. J. H. Builtjes, M. G. M. Roemer

### A Parameterization of the Effect of Clouds on Photodissociation rates; Comparison With Observations

Many molecules in the atmosphere dissociate under influence of incident sunlight, a process called photodissociation. This photodissociation typically initiates a sequence of chemical reactions (radical formation) and is therefore a key process in simulating the chemistry of the atmosphere.

Michiel van Weele, Jordi Vilà-Guerau de Arellano, Peter G. Duynkerke

### Evaluation of Radiative Flux and Tropospheric Chemistry Under Global Climate Change Scenarios

A detailed one dimensional radiation model coupled with a comprehensive regional scale photochemical model(STEM-II) is used to evaluate ground level ultraviolet radiation and changes in photochemical oxidant concentrations under potential global change scenarios. The simulations are conducted for a representative metropolitan region in eastern United States. The simulations involve perturbations to radiative flux based on trends of stratospheric ozone loss. The impact of increased radiative flux on tropospheric ozone is analyzed. In addition, the effects of tropospheric pollution levels on ground level radiative flux are studied. For the urban environment increased tropospheric ultraviolet radiation leads to an increase in photochemical production while ground level radiative fluxes are significantly reduced as a result of increased levels of tropospheric ozone production.

Kevin C. Crist, Gregory R. Carmichael, Kuruvilla John

### Similarity and Scaling for Convective Boundary Layers (Extended Summary)

Convective boundary layers (CBLs) are observed in the atmosphere and ocean, reproduced in wind tunnels, and investigated numerically with large-eddy simulation models. A host of data on vertical distribution of turbulent statistics and mean flow characteristics have been collected. The problem arises as to how we can compare the data representing CBLs of different geometrical scales, in different media, and characterized by different buoyancy forces and velocity shears.

Sergej Zilitinkevich

### A Fast Lagrangian Particle Model for Use with Three-Dimensional Mesoscale Models

In recent years, the air quality modelling field has seen the development of Lagrangian particle dispersion models driven by wind and turbulence predictions from three-dimensional mesoscale models. The Lagrangian models are based on a Langevin equation for the velocity of a fluid particle, and can take a number of different forms, depending mainly on the type of turbulence being simulated. We briefly review these forms, concentrating on the theoretically correct version for simulating dispersion in a convective boundary layer (CBL).

W. Physick, P. Hurley

### A Random Walk Model for Atmospheric Dispersion in the Daytime Boundary Layer

In this paper we present a Random Walk Model (RWM) for atmospheric dispersion in the daytime boundary layer with a continuous transition between neutral and convective conditions. A RWM describes the dispersion of a passive tracer in a flow, whose turbulent structure is known. Particle trajectories are generated using a stochastic model for Lagrangian velocities, and from those the mean concentration field is calculated.

Caterina Tassone, Sven-Erik Gryning, Mathias Rotach

### Applied Model of the Height of the Daytime Mixed Layer Including the Capping Entrainment Zone

Dilution of air pollutants by vertical mixing is mainly confined within the atmospheric boundary layer, which consists of the fully mixed layer and the capping entrainment zone. A fairly accurate estimate of the height to which pollutants are mixed is of main importance in air pollution monitoring and assessment studies.

Ekaterina Batchvarova, Sven-Erik Gryning

### Applications of the Mixed Spectral Finite-Difference (MSFD) Model and Its Nonlinear Extension (NLMSFD) to Wind Flow over Blashaval Hill

In Walmsley et al. (1990), a comparison was made among four models for surface-layer flow in complex terrain. The models were applied to Blashaval Hill, Scotland, the site of a 1982 field experiment. In the present study we compare the results of two additional models for the same terrain.

John L. Walmsley, Wensong Weng, Stephen R. Karpik, Dapeng Xu, Peter A. Taylor

### Impact of a Fully Spectral Microphysical Scheme Upon Gas Scavenging in a Mesoscale Meteorological Model

The process of wet deposition is the main mechanism for removing trace gases and aerosol particles from the atmosphere (Iribarne and Cho (1989)). They have shown that modeling cloud chemistry and wet deposition is a very complex task because dynamical and microphysical processes are linked together and interact strongly with chemical processes. Dynamical processes are responsible for the transport of water, gas and aerosol particles, but the chemical concentrations in aqueous phase are determined by the microphysical history of the droplets, including condensational and coalescence events. Modeling studies such as those of Hales (1989) and Ferretti et al. (1992) use a bulk microphysical parameterization like the Kessler one (1969), which treats droplet spectra as a single well mixed ideal solution of droplets that maintain vapor liquid equilibrium with surrounding trace gases. However, experimental results of Noone et al. (1988) and theoretical results obtained by Flossmann et al. (1987) show a dependency between drop size and the chemical concentration inside the drops.

N. Huret, N. Chaumerliac, S. Cautenet

### A Numerical Study of DMS-Oxidation in the Marine Boundary Layer

During the last decade, dimethyl sulfide (DMS) has been invoked as an important source of non-anthropogenic sulfur (Andreae et al., 1983; Nguyen at al., 1983). Produced by marine phytoplankton, DMS is transferred to the marine boundary layer (MBL) where it is oxidized to sulfur dioxide (SO2),methanesulfonic acid (MSA) and to sulfuric acid (H2SO4). The relationship between DMS emission, cloud condensation nuclei (CCN) and cloud albedo is a problem of climatic interest (Charlson et al., 1987) to be treated on the mesoscale due to the fact that the processes involved in DMS-oxidation in the MBL have timescales typically of the order of some hours or less.

Karsten Suhre, Robert Rosset

### A Comparison of Fast Chemical Kinetic Solvers in a Simple Vertical Diffusion Model

The photochemical reaction mechanisms used in regional air quality models usually consider 20 to 100 pollutant species. The equations resulting from these chemical mechanisms are nonlinear, highly coupled and extremely stiff depending on the time of the day. Therefore, the simulation time of the models is determined to a large degree by the computational burden associated with the solution of the chemistry equations.

Oswald Knoth, Ralf Wolke

### Modelling Flux-Gradient Relationships for Chemically Reactive Species in the Atmospheric Surface Layer

In atmospheric surface layer studies, the turbulent flux of variable A is approximated by the formula (Stull, 1988; Garratt, 1992): 1$$\overline {wa} = - \frac{{\kappa zu * }}{{{\phi _a}(\zeta )}}\frac{{\partial {\rm A}}}{{\partial z}}$$

Jordi Vilà-Guerau de Arellano, Peter G. Duynkerke, Karl F. Zeller

### Neural Network Techniques for SO2 Episode Prediction

Air quality has become a major issue during the last decades for a great number of cities and a matter of concern for both citizens and scientists. The quality of life has deteriorated in urban regions due to high pollution levels and a large number of people suffer from pollution effects on health. The prediction of air pollution episodes is very useful in that it enables the local authorities to give a warning against high pollutant concentrations or to take limitation measures on the emission sources. hi the past, efforts for pollution level forecasting have been reported using mathematical models (simulating the physical process) or statistical methods (Zannetti, 1990). The mathematical model operation requires high computational capacity, meteorological data and emission inventory. The required meteorological data are very difficult to collect and process, while the creation of an emission inventory is not an easy task.

S. J. Perantonis, N. Vassilas, G. T. Amanatidis, S. J. Varoufakis, J. G. Bartzis

### Source Footprint Analysis for Scalar Fluxes Measured in Flows over an Inhomogeneous Surface

Any source near the ground could potentially contribute to the vertical scalar flux measured downwind, depending on the distance from the source and the height of the measurement, as well as the flow and turbulence characteristics of the atmospheric boundary layer. The `footprint’ is a measure of the relative importance of upwind sources to the vertical flux measurements at a given point.

Ashok K. Luhar, K. Shankar Rao

### Multiple Master Length Scales Derived from a Statistical Diffusion Theory

The second-order closure models have been claimed to be one of the most appropriated tools to simulate numerically the planetary boundary layer. They reconcile the numerical amenability with a more basic physical description of the turbulent process. Several second-order closure models simulate appropriately the convective boundary layer(CBL). Simulating the stable planetary boundary layer(SBL) on the basis of a second-order closure model is a more difficult task. This certainly explains why SBL has been numerically simulated less frequently than the corresponding CBL(Moeng and Wyngaard,1989). The worst difficulty has to do with the proper choice of a turbulent master length scale(Mellor and Yamada,1982) that leads to an adequate parameterization of the undetermined terms in the equations for evolution of the second-order moments. In the SBL turbulence length scales are also relatively small and limited ultimately by the local Obukhov length rather than SBL Depth H. Recently Lacser and Arya(1986) have shown that the schemes which incorporate the local Monin-Obukhov length as a stability limit on the turbulent master length scale predict shallower and more stable boundary layers. Detailed comparisons between model predictions and data from the Cabauw mast have shown that Delage’s(1974) mixing-length formulation performed better than the other schemes in describing vertical profiles and temporal behavior of the mean and turbulent variables.

G. A. Degrazia, A. P. de Oliveira, O. L. L. Moraes

### Development of a Lagrangian Stochastic Model for Dispersion in Complex Terrain

A particle model SPRAY suitable for dealing with the atmospheric dispersion of buoyant emission in complex terrain has been developed. It is an extension of the model LAMBDA (Anfossi et al. 1991, Brusasca et al. 1989 and 1992) based on the Langevin equation, designed to simulate the dispersion on flat terrain, which was validated in various atmospheric conditions. In this last model the vertical profiles of wind and turbulence, defined at the source locations, were kept constant in all the computational domain.

G. Brusasca, G. Tinarelli, D. Anfossi, E. Ferrero, F. Tampieri, F. Trombetti

### Recurrence of Extreme Concentrations

We have known for many years that atmospheric dispersion modeling should include not only predictions of mean concentrations of pollutants, but also information about possible extreme events, i.e. probabilistic statements about deviations from the predicted mean concentrations. One way of complying with this requirement is to predict, as a function of time and space, the variance together with the mean of the concentration. With an assumption about the form of the probability density it will then be possible to calculate the probability that, at any given time, the concentration will exceed a particular value. Relevant discussions can be found in Chatwin and Sullivan (1993) and references therein.

Leif Kristensen

### Major industrial Hazards: The SEVEX Project - Source Terms and Dispersion Calculation in Complex Terrain

In 1982, the European Economic Community has adopted a Directive on the prevention of major accidents of industrial activities (“SEVESO Directive”). In this frame, Public Authorities have namely to delineate “risk areas” and organise “external emergency planning”. For this purpose, the Southern Region of Belgium (Walloon Region) supports a multidisciplinary approach known as the “SEVEX Project” (SEVeso EXpert). The project results from the close co-operation of three Belgian university departments, each being involved in a specific part of the computation of data needed to assess the effects of a major industrial hazard. The determination of the source has been done by the Faculté Polytechnique de Mons, the mesoscale wind field has been developped by the Université Catholique de Louvain, and the dispersion calculation was performed by the Université de l’Etat à Liège.

C. Delvosalle, J-M. Levert, F. Benjelloun, G. Schayes, B. Moyaux, F. Ronday, E. Everberg, T. Bourouag, J. P. Dzisiak

### A Mesoscale Boundary Layer Meteorological Model for Inhomogeneous Terrain

Limited area mesoscale meteorological models are very useful tools for regional weather forecast on short temporal and spatial scales, and for air pollution studies. With particular interest in air pollution transport and dispersion in the boundary layer and for responding to air pollution emergencies in coastal regions, a mesoscale model restricted to the lower troposphere and capable of providing boundary layer meteorology simulations within a short time has been developed. This model is programmed for a microcomputer and is in research mode. Daggupaty and Sahota (1991) report results with an earlier version of this mesoscale model by assuming the land area of the model domain as one homogeneous land type. In the present, paper we discuss the model results by considering land heterogeneities.

S. M. Daggupaty, R. S. Tangirala, H. Sahota

### Use of DMI-Hirlam for Operational Dispersion Calculations

During the latest decades there has been a. rapid progress in the ability to forecast weather. This progress has been possible because of the development of advanced numerical weather-prediction (NWP) models running on the most powerful computers available. Especially, there has been progress in the capability to make numerical forecasts in the range from one day to about a week ahead in the Northern Hemisphere which is dominated by the travelling weather systems. On longer time scales, there has been some progress, but severe difficulties are met due to the intrinsic chaotic nature of the atmosphere.

Jens Havskov Sørensen, Leif Laursen, Alix Rasmussen

### Experimental Evaluation of a PC-Based Real-Time Dispersion Modeling System for Accidental Releases in Complex Terrain

The local-scale real-time dispersion modelling system LINCOM/RIMPUFF is evaluated using data from the complex terrain Guardo experiments carried out in Northern Spain November 1990.

Søren Thykier-Nielsen, Torben Mikkelsen, Josep M. Santabàrbara

### Radioactive Dispersion Modelling and Emergency Response System at the German Weather Service

Following the Chernobyl accident in 1986, the German Weather Service (DWD) was called upon by law to perform emergency response calculations in case of nuclear incidents. It was already charged with supervising atmosphere and precipitation for radioactive substances and their transport, and a monitoring network comprising measurements and trajectory calculations was operational.

Barbara Fay, Hubert Glaab, Ingo Jacobsen, Reinhold Schrodin

### The Embedding of the Lagrangian Dispersion Model LASAT into a Monitoring System for Nuclear Power Plants

The Lagrangian dispersion model LASAT (LAgrangian Simulation of Aerosol Transport) has its origin in a research model developed in 1980. One key problem then was the correct treatment of the dispersion in inhomogeneous turbulence fields. When the problem of the drift velocity in the case of inhomogeneous turbulence was solved1,2, the model was used as the reference for calibrating simpler dispersion models of Gaussian type3. In 1989, the model was implemented on a PC and became available for other users.

Lutz Janicke

### A Transport and Dispersion Model Performance Evaluation Using the Results of a Tracer Experiment in Complex Terrain

Within the framework of the TRANSALP project, a short range tracer experiment was carried out on October 3, 1991 under calm, purely thermic wind conditions over a terrain that is representative in irregularity for the northern prealpine region of Switzerland. TRANSALP is a project in the framework of the environmental research program EUROTRAC of the European Community and is designed to investigate the transport and diffusion of air pollutants in the planetary boundary layer over complex terrain.

R. Lamprecht, D. Berlowitz

### An Emergency Response and Local Weather Forecasting Software System

Recent advances in computer technology have now placed supercomputer power on the desktop for a small fraction of the price. Many traditional supercomputer applications have benefited greatly in the move from the realm of the supercomputer center to more direct local control of the end user. Two of the atmospheric applications that have and will continue to benefit greatly from these advances in computer technology is in the arenas of local weather forecasting and emergency response systems.

Craig J. Tremback, Walter A. Lyons, William P. Thorson, Robert L. Walko

### Comparison of Models for Aerosol Vaporisation in the Dispersion of Heavy Clouds

The dispersion of heavier-than-air aerosol clouds is of great relevance to the analysis and assessment of hazards in the chemical and petrochemical industries. There are many simple mathematical models for heavy gas dispersion, but two-phase dynamics are either ignored or treated in a very simple way, without any convincing scientific justification. In fact essentially all of those models which do treat two-phase phenomena rely on the “homogeneous equilibrium” assumption, whether they be for jet dispersion, for gravity dominated dispersion, or for any other aspect of a two-phase release.

J. Kukkonen, M. Kulmala, J. Nikmo, T. Vesala, D. M. Webber, T. Wren

### MDGP: A New Numerical Model for Dense Gas Dispersion. Sensitivity Analysis and First Validation Trials

Modern industries produce and use different substances that are dangerous for the human health because they can be toxic, flammable or explosive. Some of these substances, if accidentally released in the atmosphere, have a density greater than that of the air. So the result is often the formation of a cloud at ground level that can travel over the populated area near the industrial plant.

Roberto Bellasio, Matteo Tamponi

### Evaluation of the Atmospheric Release Advisory Capability Emergency Response Model for Explosive Sources

The Atmospheric Release Advisory Capability (ARAC) at the Lawrence Livermore National Laboratory (LLNL) uses a modeling system to calculate the impact of accidental radiological or toxic releases to the atmosphere anywhere in the world (Sullivan et al., 1993). Operated for the U. S. Departments of Energy and Defense, ARAC has responded to over 60 incidents in the past 18 years, and conducts over 100 exercises each year. Explosions are one of the most common mechanisms by which toxic particulates are injected into the atmosphere during accidents. Automated algorithms with default assumptions have been developed to estimate the source geometry and the amount of toxic material aerosolized. The paper examines the sensitivity of ARAC’s dispersion model to the range of input values for explosive sources, and analyzes the model’s accuracy using two field measurement programs.

Ronald L. Baskett, Robert P. Freis, John S. Nasstrom

### Improving the Science of Regulatory Dispersion Models for Short-Range Applications

Atmospheric dispersion models used in assessing environmental impact and regulatory issues are often labeled “regulatory.” Such models are typically of a simple form due to their repeated use and the fast turnaround required. In the United States for example, models often are run for a year of hourly meteorological inputs to determine the highest or “worst-case” concentration for a source. Despite the simplicity requirement, however, regulatory models must be based on sound physical principles to generate confidence in their results.

J. C. Weil

### European Coordinating Activities Concerning Local-Scale Regulatory Models

During the 19th ITM in Ierapetra, a round-table discussion on harmonization within atmospheric dispersion modelling took place. Since then, there has been a number of activities within Europe related to this issue. In particular, two workshops dealing with topics discussed at the Ierapetra meeting have been held - one of them in May 1992 at Risø, Denmark, and the other in August 1993 in Manno, Switzerland. The workshop at Risø was entitled “Objectives for Next Generation of Practical Short-Range Atmospheric Dispersion Models”, and that in Manno “Intercomparison of Advanced Practical Short-Range Atmospheric Dispersion Models”.

H. R. Olesen

### UK Atmospheric Dispersion Modelling System Validation Studies

The UK Atmospheric Dispersion Modelling System (UK-ADMS) is a new PC based dispersion model for continuous releases, or releases of finite duration, based on an up to date understanding of the boundary layer and dispersion and, including the complex effects of underlying complex terrain, buildings and coastlines. The scientific approach used has been described at previous ITM’s (Hunt et al 1990, Carruthers et al 1993). After extensive validation and testing the system has now been released. The purpose of this paper is to summarise the main features of the system and present important aspects of the validation.

D. J. Carruthers, C. A. McHugh, A. G. Robins, D. J. Thomson, B. Davies, M. Montgomery

### The Use of Simultaneous Confidence Intervals to Evaluate Carbon Monoxide (CO) Intersection Models

The United States (U.S.) Environmental Protection Agency (EPA) has recently evaluated the performance of eight modeling techniques in simulating carbon monoxide (CO) concentrations at six intersections in New York City (EPA, 1992). The eight intersection modeling techniques evaluated include: CAL3QHC (1985 Highway Capacity Manual Modified CAL3Q Model), FHWAINT (Federal Highway Administration (FHWA) Intersection Model), GIM (Georgia Intersection Model), EPAINT (EPA Intersection Model), CALINE4 (California Line Source Model), VOL9MOB4 (MOBILE4 Modified Volume 9 Technique), TEXIN2 (Texas Intersection Model), and IMM (Intersection Midblock Model). The New York City database includes hourly meteorological, carbon monoxide, and traffic observations for six intersections in the city. This paper describes a method developed by EPA for aggregating component results of model performance into a single performance measure which is then used to compare the overall performance of the modeling techniques.

Donald C. DiCristofaro, David G. Strimaitis, Thomas N. Braverman, William M. Cox

### A Study of the Dispersion of Air Pollutants Released from Major Elevated Sources Located Near Athens, Greece

In the area around Athens is a number of elevated sources mainly power plants, refineries, cement plants etc. These industrial installations are using fuels of different types and characteristics. Almost all the examined installations are located near the coast and therefore the dispersion characteristics show significant spatial and temporal variations.

Pavlos Kassomenos, George Kallos, Maria Varinou, Anastasios Papadopoulos

### Evaluation of Atmospheric Dispersion Models During the Guardo Experiment

In November 1990, an intensive field tracer experiment was carried out in the Guardo valley (northern Spain). The main purpose of the tracer study was the atmospheric characterization of dispersion patterns over complex terrain. To do so, an evaluation exercise based on different modelling platforms was developed. The selected modelling tools embraced a set of different dispersion techniques: (1) a straight-line gaussian plume model, (2) a gaussian puff model drived by a simple interpolated wind field, (3) a particlein-cell model drived by a mass-consistent wind model and (4) a lagrangian particle model drived by a prognostic mesoscale model.

José I. Ibarra

### Dispersion Modelling and Observations from Elevated Sources in Coastal Terrain

This study is focussed on the Central Coast region of New South Wales, on the east coast of Australia, where Pacific Power Inc. operates three power stations, situated on small lakes within a few kilometres of the coast. One of the objectives of the study was to assess the suitability and value of our modelling system to the air quality assessment needs of Pacific Power in the region, including the evaluation of green-fields sites for future power stations. The dispersion of plumes from all power stations throughout the year is not only influenced by the terrain blocking and channeling of the synoptic winds, but also by mesoscale wind systems such as sea breezes and drainage flows. Our Lagrangian Atmospheric Dispersion Model (LADM) predicts winds and turbulence and uses these to simulate the transport and diffusion of emissions from discrete sources, for impact distances ranging from hundreds of metres to a few hundred kilometres. In this paper we compare LADM results to available observations in the far field.

J. A. Noonan, W. L. Physick, J. N. Carras, D. J. Williams

### Sensitivity Analysis of the Urban Airshed Model to Wind Fields Derived From the Regional Oxidant Model, Diagnostic Wind Model, and the URBMET/TVM Mesoscale Model

The United States Environmental Protection Agency has recommended the use of the Urban Airshed Model (UAM) with the Carbon Bond-IV chemical mechanism in evaluating the efficacy of emission control strategies and in determining controls necessary to reduce ambient ozone concentrations in urban areas to the level of the National Ambient Air Quality Standard (NAAQS) for ozone. UAM applications to urban areas in the northeastern United States utilize a one-way nesting of the UAM with the Regional Oxidant Model (ROM) for deriving the wind fields for the UAM. However, diagnostic and prognostic meteorological modeling techniques are being explored as alternatives for the development of wind fields for the UAM. In this study, we present the sensitivity of the UAM-predicted ozone concentrations to the wind field derived from three different approaches. The three techniques are, (1) interpolation of ROM wind fields to the UAM using the ROM-UAM Interface system, (2) interpolation of the observed surface and upper-air data using a diagnostic model (DWM), and (3) interpolation of winds from a mesoscale planetary boundary layer model (URBMET/TVM).

G. Sistla, S. T. Rao, R. D. Bornstein, F. Freedman, P. Thunis

### Assessment and Verification of Different Types of Dispersion Models in Complex Terrain

In connection with the evaluation and selection of calculational models, which are well suited for real time dispersion simulation of airborne radioactivity in complex terrain, 10 dispersion models of 4 different types are evaluated, compared and verified by the analysis of 2 different experimental episodes of the tracer experiment SIESTA. The result of the analysis had to be assessed with respect to the performance of the models, which was done qualitatively by visualization and quantitatively by statistical analysis as well as by soft-/hardware characteristics. Specific conclusions had to be drawn with respect to the final selection of a dispersion model operational in complex terrain and under real time conditions. 2 conditions which still present a real challenge for modern scientific computing. More general conclusions are drawn.

Volker R. D. Herrnberger, Patrick Doria

### Examination of the Efficacy of Voc and NOx Emissions Reductions on Ozone Improvement in the New York Metropolitan Area

Ozone is not directly emitted into the atmosphere, but is instead a secondary pollutant that is formed from a variety of atmospheric reactants in the presence of sunlight. The magnitude of ozone concentrations in urban areas, for example, along the eastern sea-board of the United States depends upon the transport of ozone and its precursors into the domain, precursors emitted within the domain, the rate at which chemical reactions take place, deposition of pollutants within the domain, and transport and diffusion of pollutants out of the domain.

Kuruvilla John, S. T. Rao, Gopal Sistla, Nianjun Zhou, Winston Hao, Kenneth Schere, Shawn Roselle, Norman Possiel, Richard Scheffe

### Deposition of Gases and Particles in the PBL: Evaluation of the Influence of a Vertical Resolution in Atmospheric Transport Models

The deposition processes determines the lifetime of a variety of gases and particles in the atmosphere. Computed depositions depend highly on the parameterisation and model concept. A vertical resolution of the tested model influences not only directly the description of dry and wet deposition processes, but also indirectly, as the chemical reactions are affected by the vertical distribution of gases and particles in the PBL. This indirect influence is due to the common assumption of full mixing in the PBL in models of no or only coarse resolution in the vertical. The effect of the vertical model resolution on estimations of concentrations and depositions of gases and particles is the subject of this paper.

Ole Hertel, Jesper Christensen, Erik Runge, Ruwim Berkowicz, Willem A. H. Asman, Kit Granby, Mads F. Hovmand, Øystein Hov

### Designing an Air Quality Network for Brisbane Using a Mesoscale Model

Brisbane is a subtropical city of one million people situated on the east coast of Australia and flanked by mountains to the west and south. The city has very few major industrial sources but is rapidly expanding. Fifteen years of multi-parameter monitoring at three inner suburban sites have shown a considerable inter-annual variability of ozone concentrations at greater than background levels, with maximum readings usually occurring between midday and early afternoon, and from late winter through to late autumn. On all of these days it is likely that areas of greater ozone concentrations exist elsewhere in the region later in the day. It is the aim of this study to identify those areas and to design an expanded meteorological and air quality network. Data from such a network will be invaluable to environmental and planning authorities.

W. Physick, P. Best, K. Lunney, G. Johnson

### An Operational System for Emergency Response to Large Scale Releases of Pollutants in the Atmosphere

The Canadian Meteorological Centre (CMC) has been designated by the World Meteorological Organization (WMO) as a Regional Specialized Meteorological Centre (RSMC) for the provision of products and guidance on atmospheric transport-dispersion-deposition of pollutants to member countries during Environmental Emergencies. The Canadian Meteorological Centre is the national meteorological centre for Canada. CMC uses a Global Spectral Model (T119, 21 levels) for medium range forecast, and a variable resolution Regional Finite Element Model (50 km horizontal resolution over North America, 23 levels) for short range forecast. In order to fulfil this international commitment as well as other national responsibilities for environmental emergency response, CMC has implemented in its operations a system for quick response to requests of products and guidance on the large scale movement of pollutants in the atmosphere.

Réal D’Amours, Michel Jean, Joseph-Pierre Toviessi, Serge Trudel

### Real-Time Application Software “Tracer Imager Package” (TRIP)

Operationally running computer programs at the IfM continously collect and archives actual meteorological using the worldwide desseminated WMO-bulletins (GTS), geostationary METEOSAT satellite data (PDUS quality), and orbital NOAA satellite data. After an extended diagnostic cycle the resulting 3D-analyses on isentropic surfaces are used to calculate trajectories by a mixed dynamic/kinematic approach (REIMER, 1991). Starting from selected source positions in forward mode or immission locations in backward mode container files are prepared that contains the synoptic 3D spreading of air particles. Also short term trajectories are calculated at each gridpoint of the analysis area to determine broad scale vertical velocity. These container files then can be interactively displayed on a personal computer using TRIP (“TRacer Imager Package”) which is the user interface of TRAMPER (Tropospheric Realtime Applied Meteorological Procedures for Environmental Research). The resulting consecutive images simulate the motion of particles or other meteorological variables that change over time. TRIP so acts as a viewer of accompanied analysis data of the given weather situation too.

Wieland Weiß, Eberhard Reimer

### Vertical Diffusion Parameter in the Atmospheric Boundary Layer

This paper presents relations between the mean vertical displacement ($$\bar z$$) of pollutants and their mean horizontal displacement (x$$\bar x$$ ) in the atmospheric boundary layer. A modification of the lagrangian similarity theory applied to atmospheric diffusion extended to the atmospheric boundary layer is developed. According to Pasquill and Smith (1983) the similarity theory of Monin-Obukhov applied to atmospheric diffusion states 1$$\frac{{d\bar x}}{{d\bar z}} - \frac{{\bar u\left( {c\bar z} \right)\bar z}}{{k\left( {\bar z} \right)}}$$ where ū is the mean wind speed, c depends on stability and K is the vertical material diffusivity. Considering the eddy diffusivity for momentum and the form of the friction velocity introduced by Yokoyama et al.(1979) a form of ū(c$$\bar z$$)for the atmospheric boundary layer is developed. Substituting ū(c$$\bar z$$) in Eq.(1), equating K to the heat diffusivity given by Yokoyama et al.(1979) and integrating, the following expressions can be obtained (where z1 is the boundary layer depth, L is the Monin-Obukhov length, θ=$$\bar x$$/z1; α=$$\bar z$$/z1; αo = zo/z1; k=0.41, β=6.9 and γ=0.92): - neutral and stable conditions: (η=z1/L ≥ 0)

Laura E. Venegas

### Trajectory Analysis of High-Alpine Air Pollution Data

The EUROTRAC subproject ALPTRAC (High Alpine Aerosol and Snow Chemistry Study) is devoted to the investigation of air and snow pollution at high Alpine sites. The aerosol surface concentration is continuously recorded at Jungfraujoch (3450 m a.s.1., 7° 59’E, 46° 32’N) in the Swiss Alps and Sonnblick (3106 m a.s.1, 12° 57’E,47° 03’N) in the Austrian Alps with a time resolution of 30 min with an epiphaniometer (Gäggeler et al., 1989; Baltensperger, et al. 1991). The measurements showed a pronounced seasonal cycle with mean summer concentrations more than one order of magnitude higher than mean winter concentrations, and the occurrence of episodes with especially high or low concentrations (Seibert et al., 1993). While the seasonal cycle is mainly to be explained by the atmospheric stability, the short-term variations are caused by synoptic-scale transports. These have been investigated using isobaric back trajectories at 700 hPa with a length of 72 h, computed twice daily for a period of three years (July 1990 - June 1993). Due to technical problems at Sonnblick, only 925 trajectories were available for the analysis; most of the missing data fall on winter.

P. Seibert, H. Kromp-Kolb, U. Baltensperger, D. T. Jost, M. Schwikowski

### An Eddy Diffusivity Model from a Theoretical Spectral Model for the Stable Boundary Layer

Much has been written in recent years concerning modeling and parameterizations of the planetary boundary layer (PBL), particularly in formulations that require turbulent fluxes of momentum, heat and moisture.These turbulent fluxes are primarily responsible for shaping the structure of the PBL and from the mathematical point of view are related with the closure problem. Thus the solution of the PBL equations require some type of closure schemes.

Osvaldo L. Moraes, Gervásio A. Degrazia, Amauri P. Oliveira

### On the Relationship Between Synoptic Scale Parameters and Pasquill Stability Classes for the Purposes of Air Pollution Modeling

Quite often, for the tasks connected with environmental impact assessment it is necessary to estimate pollutant concentrations on the basis of standard meteorological data such as sunshine, cloud amount, and wind velocity using the Pasquill stability classes. They are usually determined from the atmospheric characteristics estimated near the ground surface using a table classification with 10 m wind velocity and insolation intensity (or cloud cover at night). These meteorological data are not allways available and they are difficult for prognostication. For the purposes of the prognostical estimation of the atmospheric diffusion in case of emergency releases when regulatory type models are used it is useful to find the relation between the synoptic scale parameters obtained from the numerical weather prediction and the Pasquill stability classes.

Dimiter Yordanov, Dimiter Syrakov, Maria Kolarova

### A Lagrangian Model of Long-Range Transport of Sulphur

A Lagrangian receptor-oriented one-layer model has been developed in order to simulate a synoptic-scale transport of airborne sulphur. Mass-balance equations for sulphur dioxide and particulate sulphate were integrated along 3-day backward trajectories arriving at 0000 and 1200 GMT at selected receptor points. Advective winds used in trajectory calculations were based on the wind profile from radiosonde reports, taking an average over the ground based layer up to an 850 hPa level. These vertically averaged winds were thereafter objectively analysed using the ‘1/r2 aligned’ technique (Kahl and Samson, 1986), which gives the greatest weight to observations upwind and downwind of the interpolation point. Sulphur emissions were taken from EMEP inventory (Sandnes and Styve, 1992). They were assumed to vary linearly over the year, with the maximum and minimum occuring in January (multiplication factor=1.3) and July (multiplication factor=0.7), respectively. Background concentrations proposed by Szepesi and Fekete (1987) were used. Other model parameters had diurnal variation and they were estimated from routine synoptic observations, where 1/r2 spatial and linear temporal interpolation techniques were employed. Parameters were calculated at the beginning of each 3-hour time step. Mixing height varied from 500 to 2000 m, depending on stability conditions. Dry deposition velocities of sulphur dioxide (vd) and particulate sulphate (wd) over the ground and transformation rate of sulphur dioxide to particulate sulphate were taken as proposed by Renner et al. (1985). Over the sea vd=0.8 cm/s and wd=0.1 cm/s were taken. Wet deposition rates for both pollutants depended on precipitation intensity and mixing height (Eliassen et al., 1988).

Zvjezdana Klaić

### Local Background Air Pollution in Response to Coastal Circulation

Background pollution is a stationary initial pollution which should be taken into account in every monitoring system and control strategies. Its definition and consequently the methods of its evaluation are not generally adopted and may vary from one country to another.

### Lagrangian Stochastic Dispersion Modelling for Varying Boundary Layer Stabilities

In a given flow field the particle velocities are modelled as a Markov process, i.e., through the following stochastic differential equation 1$${u_i} = {a_i}\left( {\operatorname{x} ,\operatorname{u} ,t} \right)dt + {b_{ij}}d{\xi _j},{\text{ }}d\operatorname{x} = \operatorname{u} \cdot dt$$

Mathias Rotach, Sven-Erik Gryning, Caterina Tassone

### Wake Flows over Mountainous Areas

The wind field near the earth’s surface is strongly influenced by the topography which can be observed in all scales of atmospheric motions above all in the mesoscale. Often observed phenomena are secondary circulations which are known to increase the efficiency of transport processes of momentum, heat, humidity and also of airborne material. Therefore it seems to be important to describe circulations induced by the topography with a good accuracy in air pollution modeling as well as in regional climatology. Flow processes influenced by the topography can be thermally or dynamically induced, whereby in nature both mechanisms work often together. Here only dynamically induced phenomena like wake flows with stable stratification are discussed. An overview about the phenomena is given by Etling (1989).

### Fluid Particle Motion in Inhomogeneous and Non-Gaussian Turbulence

It has been proven that modelling turbulent dispersion in the Lagrangian framework is a successful and flexible approach. The effects of inhomogeneity, instationarity and nonGaussianity of turbulence have to be taken into account (Thomson, 1987; Sawford, 1991) and there are questions concerning the modelling of buoyancy (van Dop, 1991) and chemical reactions. As shown recently (Heinz and Schaller, 1993), equations for particle motion, particle potential temperature and mass fractions of chemical components can be determined completely consistent with the exact ones for the means and variances, assuming the approximations of Kolmogorov and Rotta. Moreover, the influence of non-linear terms to the systematic particle motion is discussed for locally Gaussian-distributed fluctuations. However, there arises the question in which extent the consistence with the exact transport equations for the first and second moments determines particle motion, which influence arises from non-Gaussian distributed fluctuations for instance for transport coefficients like the diffusion coefficient or the Lagrangian timescale. To get more insight in the contributions of nonGaussianity and inhomogeneity transport coefficients are calculated here for different non-Gaussian distributed vertical velocity fluctuations coinciding in it means and variances.

Stefan Heinz

### Mesoscale Modelling of the Atmospheric Input into Coastal Waters

Coastal waters and the oceans are not only polluted from rivers and by direct input but also from the atmosphere by a considerable amount. In the paper examples are given for the input of SO2, NO2, and Pb into the German Bight in a high temporal and spatial resolution. The input figures are calculated from a mesoscale model. The modelled meteorological situations are representative for typical weather conditions in May when the atmospheric input might become an important factor to control the growing of phytoplankton.

K. Heinke Schlünzen, Klaus Bigalke, Ulrike Niemeier

### Regional Scale Transport Model of Atmospheric Acid Compounds, and Its Application for Hungary

A regional scale operational model for the simulation of air and precipitation quality originating from potentially acidifying substances has been worked out (Fekete, 1986; Fekete and Szepesi, 1987; Fekete and Gyenes, 1993). The poster shown at the 20th ITM will present the results of the Hungarian regional scale model for 1987.

Katalin E. Fekete

### Regional Scale Modeling Case Studies for Aerosol Transport Over Hungary

Aerosol samples were daily taken simultaneously on Nuclepore filters in suburban Budapest and at the regional background air pollution station (K-puszta) 80 km far from Budapest. The samples were analysed among other elements for V, Cr, Co, Ni, Cu, Zn, As and Pb by means of PIXE method. The cases with constant air mass flow between the source and the receptor determined on the basis of radiosonde observation for 925 hPa level carried out in Budapest and Szeged were chosen for modeling consideration. K-puszta station is located between these two cities. A simple source-receptor model of Lagrangian-type was used for the simulations. Budapest was considered as an area source for the elements mentioned above. Aerosol concentration data measured in suburban Budapest were applied as input parameters altogether with the actual mixing heights, precipitation intensities, dry deposition velocities and washout ratios. It was assumed that there are no anthropogenic sources between the two points - except for Pb where road transport has been taken into consideration. Dry and wet cases were separately examined. The aerosol concentrations computed for K-puszta were compared to the data measured there.

Ernő Mészáros, László Bozó, Ágnes Molnár

### Fluxes over Complex Terrain, Analytical Evaluation

It is long been recognized that land and sea contrasts produce sea breezes. More recently, it has become evident that landscape variations which result in spatial gradients in surface heat flux can produce mesoscale circulations as strong as sea breezes. However, it remains to determine how small a horizontal variation in surface heating can still produce a significant mesoscale circulation. It order to address this question, in this paper we present results for one Fourier component. While only the linear response can be evaluated using this tool, such models have been shown to be effective at identifying dominant physical processes associated with sea breezes (Rotunno, 1983; Dalu and Pielke, 1989, 93). This approach can also be used to assist in the development of a parameterization of mesoscale effects generated by spatial surface sensible heat variability for use in larger scale models. The primitive equations can be reduced to an equation for the streamfunction: 1$$\left[ {{{\left({\frac{\partial }{{\partial x}} + {k_p}} \right)}^2} + k_o^2} \right]\frac{{{\partial ^2}\varphi }}{{\partial {z^2}}} + \left[ {{{\left( {\frac{\partial }{{\partial x}} + {k_p}} \right)}^2} + {l^2}} \right]\frac{{{\partial ^2}\varphi }}{{\partial {x^2}}} = - \frac{1}{{{U^2}}}\frac{{\partial Q}}{{\partial x}}$$2$$\begin{gathered}\frac{{\partial \varphi }}{{\partial z}} = u {\text{and}} \frac{{\partial \varphi }}{{\partial x}} = - w; {l^2} = \frac{{N_o^2}}{{{U^2}}} - \frac{{{U_{zz}}}}{U}; p = s + \lambda ; {k_p} = \frac{p}{U} + \frac{{K{k^2}}}{U} {\text{and}} {k_o} = \frac{f}{U} \hfill \\h = {h_o}\frac{1}{2}\left[ {1 + {a_1}\sin \left( {kx} \right) + {a_2}\cos \left( {kx} \right)} \right] = {h_o}\frac{1}{2}\left[ {1 + a\sin \left( {kx + b} \right)} \right] \hfill \\\end{gathered}$$3$${a_1} = \frac{{\lambda \left( {\lambda + K{k^2}} \right)}}{{{{\left( {\lambda + K{k^2}} \right)}^2} + {k^2}{U^2}}}; {a_2} = - \frac{{\lambda kU}}{{{{\left( {\lambda + K{k^2}} \right)}^2} + {k^2}{U^2}}}; \mathop {\lim }\limits_{k,U \to \infty } {a_1},{a_2} = 0$$4$${h_a} = \frac{1}{2}{h_o}\left( {1 + a} \right) Q = \frac{{{Q_o}q\left( s \right)}}{2}\left[ {1 + a\sin \left( {kx + b} \right)} \right]He\left( {{h_a} - z} \right)$$

G. A. Dalu, M. Baldi, R. A. Pielke, G. Kallos

### Lagrangian Model Simulation of 3-D Concentration Distribution in Complex Terrain

Most of the up to date dispersion simulations are discussed using ground level concentration (g.l.c.) values only. However it is straightforward that the spatial distribution of the concentration should be carefully computed, in order to assess the capability of a given model to correctly simulate the atmospheric dispersion. Therefore, the vertical and horizontal concentration profiles at various downwind distances and at different levels above ground should also be computed and compared to the corresponding observed values if any, even if this is generally thought of as very severe test for any kind of model.

G. Tinarelli, D. Anfossi, G. Brusasca, E. Ferrero, U. Giostra, M. G. Morselli, F. Tampieri, F. Trombetti

### Treatment of Transport in MOGUNTIA

At the IMAU, the program MOGUNTIA has been implemented. MOGUNTIA is a climatological three-dimensional global tropospheric transport model, designed by Zimmermann (1988). The current resolution amounts to l0o x l0o x 100 hPa. The upper layer is located at 100 hPa. The model is used for extensive tropospheric chemistry calculations in which the effects of heterogeneous chemistry and aerosols can be taken into account. The global character of the model has proved to be valuable for example to calculate budgets of ozone (Dentener and Crutzen, 1993; Crutzen and Zimmermann, 1991).

Maarten C. Krol

### Methodology for Mapping Local Scale Deposition of Acidifying Components over Europe

Current methods and models for estimating deposition of acidifying components differ largely in the used horizontal resolution i.e. varying from 5x5 km (Erisman,1992) up to 150x150 km (EMEP, Iversen et al. 1991). In describing the effects of acidification on the level of ecosystems, acid loads should be available at least on the size of ecosystems. No deposition maps on this resolution are available leaving a serious gap in estimating exceedences of critical loads in Europe.

W. A. J. van Pul, J. W. Erisman, J. A. van Jaarsveld, F. A. A. M. de Leeuw

### Ground Level Concentrations of Ozone, Oxidant, PAN and Precursors in the Netherlands During the Last Two Decades and the Relation with the LOTOS-Model

PeroxyAcetylNitrate (PAN), ozone (O3), nitrogen oxides (NOx) and hydrocarbons (C2-C4) are measured at several ground level stations in The Netherlands from 1973 onwards. The hydrocarbon measurements at Delft and Moerdijk show the existence of large industrial sources. Analysis of the annual mean values of the compounds has brought about the following results:• In the seventies levels of C2-C4, NOx and PAN have generally gone up. For O3 and Ox no significant trend is observed.• In the eighties levels of C2-C4 and NOx have slightly gone down. For PAN, O3 and Ox no significant trend is observed.The effect of a reduction in precursors on PAN and oxidant levels is estimated with the LOTOS model. The model calculations over April 1985 show, for The Netherlands, that a 30% emission reduction of NOx and VOC results in a 6% reduction of PAN and no reduction of Ox.

P. Esser, M. G. M. Roemer, P. J. H. Builtjes, R. G. Guicherit, Th. Thijsse

### Meteorological Aspect of Chemical Composition of Precipitation/Deposition - Long Range Pollution Transport in a Mountain Region

The acidification of forest ecosystems in the southwestern Poland (Sudety Mountains) has been a topic of intensive research and modeling efforts in recent years. It is generally believed that this problem is a consequence of transport and deposition of acidic atmospheric pollutants, both gases and particles. In the period from 1988 to 1992 sampling was underway at five locations of the mountain legion (810 ÷ 1490 m a.s.l.)1,2. The main objective of the network was to measure routinely the daily concentration: SO2, NH4+, NO2, $$SO_{4 acr}^ =$$, total suspended particles, sulphates in air and the following species: pH value, SO4=, C1-, N(NO3-), N(NH4+), K+, Na+, Ca++, Mg++ in precipitation/deposition (rain, snow, rime). Moreover, meteorological parameters were registered: windspeed (V) and direction, humidity (H), temperature (T), the circulation type, the amount of precipitation and trajectory wind roses. Air acidic species episodes and precipitation/deposition concentrations of ionic species are found to be strongly dependent upon meteorological circumstances. On a basis of freq-uency histogram of daily average values (concentration species, meteorological parameters), the range of values wem determined for which frequency of occurence was 55 ÷ 80%.

G. Kmieć, A. Zwoździak, K. Kacperczyk, J. Zwoździak

### Sea Breeze in Summer, Along the West Coast of Portugal

Along the west coast of Portugal, in a narrow strip of 20 to 35 Km, more than 70% of the portuguese population is concentrated, and more than 90% of air pollutants in Portugal are released in this region. The atmospheric circulations detailed study as well as the thermal structure in the lower troposphere are very important because they are determinant factors in the air pollutants transport and dispersion conditions. During the dry period, since middle May to middle October, it is frequent the formation of a thermal low pressure over the central region of the Iberian peninsula; in association with this low pressure, a persistent ridge over the north of the peninsula is connected with the Azores anticiclone. Under these conditions, along the west coast of Portugal, especially during the months since June to September, the sea breeze regime is a direct consequence of the differential heating of the surface air, which reaches more than 20 oC between the land and the sea, in the horizontal scale along the coastal strip.

Renato A. C. Carvalho, Victor Prior

### Researches on Dispersion of the Pollutants Emitted into the Atmosphere from a Nuclear Power Plant Under Rugged Conditions in View of Validating Varied Mathematical Models

The air concentrations field determination of the exhausted noxes through sources of low emissions height, from the energy systems (NPP,TEPP) or industrial plants situated on a complex roughness soil (hills, valleys, high buildings, and so on) are of major concearn for selecting a mathematical model. In this case both the influence of the buildings in the emission area on the plume rise (Δh) and of the complex soil roughness on the air currents circulations are manifested. These influences are reflected through the local changes of the air dispersion parameters as transport support.

Traian Pop, Livia-Mihaela Pop

### Long-Term Average Air Pollution over Cities: Operative Calculation Technique for Elevated Sources

Long-term average field of pollutant concentrations in the atmospheric surface layer is one of the criteria of air quality in different urban districts as well as a basis for accepting adequate city layout decisions. An applied technique for calculating such a field formed by urban multiple elevated point source emissions is proposed. Superposition principle for single source average fields is used. The technique takes into account a seasonal joint frequency for wind velocity, wind direction, thermal stratification and precipitation gradations, seasonal changes in emission rates, plume rise depending on seasonal emission characteristics and meteorological gradations, dry deposition of different spesies and washout. An original method of the stratification parameter calculation based on routine meteorological data is proposed. The whole algorithm is implemented as a set of FORTRAN routines.

Irina A. Krotova, Larissa Melikhova

### Influence of the Traffic Conditions on the Air Quality of Barcelona During the Olympic Games ‘82

During the last Olympic Games ‘82 in Barcelona (July 25th - August 9th), 1) the city council of Barcelona adopted special traffic restriction measures in some areas (several streets were closed and there were parking limitations) and 2) some traffic jams were expected between 1100and 1300LST, and between 1900 and 2100 LST in several highways going into the city. The aim of this contribution is to present the results of a numerical simulation to forecast the influence that these traffic conditions could have on the air quality over the city with respect to the normal situation existing before the Olympic Games.The simulations have been performed for a typical summer non-workable day (24 hours) within a 39x39x6-km3 region with a horizontal grid spacing of 1 km x 1 km.Concentration fields for NOx and O3 have been calculated for selected scenarios on the Olympic Games ‘92 in Barcelona by applying the Eulerian dispersion model for reactive species MARS.1 For the numerical modeling of photochemical oxidant formation the chemical transformation of pollutants should be considered together with their transport in the atmospheric boundary layer. 3D wind fields necessary to run the model were calculated by the non-hydrostatic mesoscale model MEMO2 and the emission inventory calculated was done by the model EMITEMA-EIM.3This work has been supported by the Servei de Control del Medi de l’Area Metropolitana de Barcelona, IBM España, S.A., and Comisión Interministerial de Ciencia y Tecnología (CICYT) grant no. NAT91-0987.The following discussion refers to differences in the pollutant concentration levels between the situation forecast during the Olympic Games ‘92 and the normal situation existing before the Olympics.Because of the traffic restrictions, surface concentration patterns obtained by simulation show lower ozone concentrations. From 900to 1400 LST, a spread area of lower ozone concentrations appears (minus 5–20 ppbv O3), which is originated in Barcelona Downtown Area (BDA) and driven towards the sea. Due to the sea breeze the afternoon ozone concentration patterns also show lower ozone concentrations (minus 5–10 ppbv O3) in the North-western part of the modelling domain and in the Llobregat valley. It can be said that the decrease of emissions due to traffic restrictions leads to decreased ozone concentrations in widely spread areas, whereas significant decrease of NO and NO2 concentrations (minus 5–20 ppbv), are located almost only near BDA, except in the morning hours (until 1000 LST), when differences of NO and NO2 concentrations are also observed above the sea. During the whole day lower NO2 concentration levels in BDA are promoted by decreased NO2 formation which is caused by less NO and O3 reactions.As to the influence of traffic jams, it can be said that at the beginning of the traffic jams higher ozone and lower NO concentrations can be seen only in a small area surrounding the concerned highway sections. Noon concentration patterns show lower NO concentrations (minus 5–20 ppbv) in the surroundings of concerned highway sections and higher ozone concentrations (plus 3–10 ppbv) in more widely spread regions. At 1300and 1400 LST the sea breeze is stronger and thus transports air with higher ozone concentrations (plus 3-10 ppbv) along the Llobregat valley. Ozone concentrations in Barcelona and above the sea are slightly higher in the 1300 LST concentration pattern. These ozone concentration levels are almost fully dilluted by fresh air in the afternoon. The influence of the evening traffic jams can be seen in the concentration patterns between 1900 and 2100 LST. Again higher ozone concentrations (plus 3-20 ppbv) are observed near the concerned highway sections, but also in the valley of Llobregat and above the sea.As conclusion, the influences of both lower emissions due to traffic restrictions in BDA and higher VOC emissions caused by traffic jams on some highway sections, are almost spatially separated, except for a small region in BDA where decrease of ozone due to traffic restrictions is almost counterbalanced by an increase of ozone formation due to traffic jams. In the evening lower ozone concentrations in the Llobregat valley because of traffic restrictions are overcompensated by higher ozone concentrations due to the traffic jams.These preliminary results are now being validated.

J. M. Baldasano, M. Costa, L. Cremades, Th. Flassak, M. Wortmann-Vierthaler

### Evaluation of the Three-Dimensional Distribution of Dense Gas Concentration Estimated by Numerical Models

Generally, the evaluations of heavy gas models have consisted in checking the performance in predicting the plume centerline concentration at ground level and the width of the plume (see, for example, Hanna et al., 1991), but they have not considered the vertical distribution of dense gas. Experimental studies have shown that vertical distributions of gas inside a dense-than-air cloud is an important factor to describe the dispersion of the cloud impacting an obstacle. In this paper, the three-dimensional distributions of dense gas concentration have been computed from the outputs of the SLAB and HEGADAS models which were run for some field experiments of dense gas release in the atmosphere. The resultant distributions have been statistically compared.

Fernando Martín, Inmaculada Palomino, Begoña Aceña

### Influence of the Topography on the Long-Term Average Concentration Computed by Dispersion Models

Although the straight-line Gaussian plume models are still recommended for regulatory use for routine release of nuclear installations, it is well known that they could be inaccurate under certain complex meteorological or/and topographical conditions. The development of computer capabilities, measurement instruments and modeling approaches made possible the use of more realistic models.

### Coupling the Photochemical, Eulerian Transport and ‘Big-Leaf’ Deposition Modelling in a Three Dimensional Mesoscale Context

In this paper we show the different modules of an air dispersion modelling study for the Madrid Area. We show the general background of a non-hydrostatic meteorological mesoscale model as a key module for obtaining the wind, temperature and humidity fields by using a k-Theory turbulence model. Because we solve the pressure in the vertical axe, we have to solve explicitly the Helmohltz equation because the complexity of the terrain (in a general sense) request to use the so-called ”terrain following coordinates”. This approach introduces non linear terms in the first and second order spatial coordinates which complicates the solution of the Helmohltz equation, Pielke (1984). The solar radiation module is also essential for knowing the radiation -in all the wave lengths-at every cell of the three dimensional domain. This radiation -global, diffuse, scattered and net-allows to parameterize the complex photochemical reactions which are undergoing in the atmosphere engined by the solar radiation. This photochemical package works in parallel with the prognostic and diagnostic meteorological model by using different time steps.

Roberto San José, Luis Rodríguez, Magdalena Palacios, Javier Moreno

### Test of Mesoscale Numerical Model on Swiss Midlands

Mountain and valley systems exert a particularly strong influence on air flow, through channelling effects of the topography, and through the generation of mountain and valley breezes. The purpose of this limited study was to see how far a mesoscale numerical model can simulate regional effects due to the complex topography of the Swiss Midlands, and to compare these results with measurements carried out in the summers of 1990 and 1991 during the two first campaigns of the POLLUMET (POLLution and METeorology) project.

D. Schneiter, C. Thurre

### Application of the Abatement Strategies Assessment Model, ASAM, to Abatement of SO2 Emissions in Europe

In preparing a new international protocol on sulphur dioxide emissions under the auspices of the UN Economic Commission for Europe, the Task Force on Integrated Assessment Modelling has been investigating scenarios for future abatement. The ASAM model (see below) has been used together with 2 other models (RAINS and CASM) to derive emission reductions required in different countries to reach given objectives. The calculations are based on estimates of atmospheric transport between different European sources and receptors derived with the EMEP model of the Norwegian Meteorological Institute.

Helen M. ApSimon, Rachel F. Warren

### Review and Evaluation of the RATCHET Model Used for the Hanford Dose Reconstruction Project

The Regional Atmospheric Transport Code for Hanford Emission Tracking (RATCHET) long-term, trajectory puff air dispersion computer model was developed by the Pacific Northwest Laboratory as part of the Hanford Environmental Dose Reconstruction (HEDR) project. The purpose of HEDR is to estimate monthly-averaged radiation doses that individuals could have received from operations at Hanford since 1945. The U.S. Centers for Disease Control (CDC) has responsibility to insure that all tools used in environmental radiation dose reconstruction performed under its direction be technically sound and defensible. In order to address concerns about the validity of atmospheric transport and deposition models, the CDC coordinated a through technical review of the RATCHET model by an expert panel of air-pollution scientists. It has been suggested that a statistical (i.e., a sector-averaged Gaussian plume) model might be better suited for calculating monthly-averaged air concentrations than the RATCHET trajectory-puff model. In order to check this suggestion, we implemented a version of the widely used AIRDOSE-EPA model, and compared this model with RATCHET for the month of January 1945. Hourly meteorological input data for AIRDOS for January 1945 were taken from the RATCHET input meteorological file. We used the winds from the 200 ft level of the Hanford tower because these correspond closely to the winds at the estimated plume height and at the source location. Figure 1 shows the contours of monthly-averaged concentration (Ci m-3) calculated by the RATCHET model, and Figure 2 shows the results from the AIRDOSE model. The highest concentration contours extend to the southeast for both models; however, beyond about 100 km from the source, RATCHET predicts that the isopleths turn to the northeast, but no similar swing is predicted by the AIRDOSE model. This comparison helps to highlight the critical nature of dose reconstruction over long averaging times. Over the complex terrain of eastern Washington, a single meteorological station cannot represent the effects of varying topography and changing meteorological conditions. It is concluded that sector-averaged models such as AIRDOSE are not suitable for dose reconstruction calculations over regions of complex terrain.

C. J. Nappo, W. R. Pendergrass, R. M. Eckman

### Linear Advection Scheme for Air Pollution Transport Modelling from Individual Sources

Presented new technique of advection numerical modelling may be attributed to methods simulating continuous flow by the motion of ensemble of discrete elements. To these methods belong “particles-in-cell” techniques,1 Egan-Mahoney’s,2 “large particles” by Belotserckovsky3. The scheme considered differs from the, indicated methods in the following: 1.Pollution is imitated by discrete extended mass elements, which widths are 2 times higher than grid interval. In the course of the motion elements are imposed on each other.2.The ensemble of these elements unambiguously corresponds to continuous distribution of grid elements two times narrower which result from conditions of first three moments conservation. (for one - dimensional variant) of the elements and defines current mass distribution with the grid.3.“Summing” of mobile elements which mass centers hit one cell is a transformation of these elements distribution again with conservation of the first three moments. In case when elements number is not growing (instantaneous source), when computer resources provide operations with full array of elements the “summing” procedure is omitted and in this case the algorithm is particularly simple, effective and its precision is increased.

Margarita Pekar

### Neural Networks Predict Pollution

Air pollution is a very big problem in Slovenia. The greatest pollutants are two big thermal power plants (TPP-s) which are placed in the valleys near the coal mines. The coal has a very big percentage of the sulphur (up to 2%) and the TPP-s do not have wet desulphurisation. The results of this are the episodes of very high air pollution in the vicinity of the TPP-s. The local government intends to omit this episodes by forcing the TPP-s to reduce the power significantly (from 30% up to 70%). Such quick reduction of the TPP power is certainly an economical and technological problem.

Primož Mlakar, Marija Božnar, Martin Lesjak

### Results of Nested Wind Flow Simulations for the Athens Basin Using the Non-Hydrostatic Model MEMO

In the frame of the APSIS project a nested version of the nonhydrostatic mesoscale model MEMO was used to simulate the wind flow in Athens. The nested model version was applied as former applications of the non-nested version of MEMO for the Greater Athens Area proved lateral boundary conditions to falsify the wind field close to the boundaries [Moussiopoulos et al. 1993].

Rainer Kunz, Nicolas Moussiopoulos

### Influence of the Sea Breeze on the Air Pollution over the Attica Peninsula

In the frame of the APSIS project simulations of the dispersion of air pollutants over the Attica Peninsula are carried out for May 25, 1990. This day is characterized by clear sky conditions and a general flow from Northwest. Such conditions support the development of a sea breeze around the Peninsula, which is especially pronounced in the basin of Athens and the neighbouring coastal zone. The flow and turbulence fields are simulated with the the non hydrostatic model KAMM (Adrian and Fiedler, 1991). Based on these fields the dispersion of the air pollutants are calculated with the DRAIS model (Nester and Fiedler, 1992). The simpler chemical mechanism RADM1 was selected because only measurements of NO, NO2, SO2,CO and O3 are available, which are well predicted by the RADM1 mechanism.

K. Nester

### Numerical Simulation of the Flow Regime in Athens Area

The flow regime in Greater Athens Area (GAA) is numerically investigated by a mesoscale, higher-order turbulence closure model. The model is three-dimensional, hydrostatic, with a terrain following coordinate system and it solves the mean quantities as well as the turbulent energy equation prognostically. This enables the explicit treatment of horizontal inhomogeneity and unsteadiness, which are key features in the area under study. It has been developed at the Department of Meteorology in Uppsala (MIUU) (Enger, 1986).

D. Melas, I. Ziomas, C. Zerefos

### Prediction of Wind Flow and Ozone Formation in Athens for the APSIS B2 Exercise Using the EUMAC Zooming Model

The APSIS activity, which was initiated in October 1991 in the frame of the EUROTRAC subproject EUMAC, aims to intercompare model simulations of photosmog formation in Athens. In addition to more experience with regard to model evaluation, obvious benefits from such an intercomparison will be the excessive testing of mesoscale models and the determination of advantages and weaknesses of individual model concepts.

Nicolas Moussiopoulos, Peter Sahm

### Intercomparison on the Flow Field over the Attic Peninsula with Two Models

The circulation of air masses for the control of air pollution is particularly difficult to estimate in areas of complex terrain, due to the presence of mountains and hills or to the effect of large water basins. In the Athens area, the local circulation is mainly driven by the sea-land breezes, which are strongly influenced by a mountain ridge culminating at 1460m (Parnitha mountain fairly close to the sea-shore). Actual terrain slopes can be locally very large, which renders the simulation particularly challenging for mesoscale models. In this paper two different models participating in the APSIS intercomparison exercice are presented and initialized in order to describe the typical summer sea-breeze in the region of Athens.

G. Schayes, H. Gallée, G. Graziani, P. Thunis

### Studies with the Three-Dimensional Eulerian Photochemical Dispersion Model Mars for the APSIS B Case: Viability of the Inclusion of Horizontal Diffusion into the Implicit Solver of the Model

The Eulerian photochemical dispersion model MARS (Graf and Moussiopoulos, 1991) solves the three-dimensional dispersion equation for chemically reactive species by coupling the vertical diffusion to the chemical reaction mechanism, using a predictor-corrector integration scheme similar to the one proposed by Gear, while the horizontal diffusion and the advection terms are computed explicitly.

D. Berlowitz, N. Moussiopoulos

### A Numerical Study of Air Flow in Thessaloniki Area

A three-dimensional, higher order turbulence closure, mesoscale model for complex terrain is applied to simulate the flow dynamics in Thessaloniki area which is located in the north-east coast of Mediterranean. The present version of the model is hydrostatic and has a terrain following coordinate system. A more detailed description can be found in other literature references (Enger, 1986, Melas and Enger, 1993).

D. Melas, I. Ziomas, C. Zerefos

### Backmatter

Weitere Informationen