Air Pollution Modeling and its Application XXII

This work conducts a multi-physics ensemble of air quality projections in order to elucidate the spreads and uncertainties behind the election of the physical parameterizations in the regional climate models. Results indicate that the studied parameterized processes and air pollutants transport and dispersion are closely tied together, and hence the projected changes are strongly affected by the atmospheric variables on the projections for air quality.

Projecting climate change scenarios to local scales is important for understanding, mitigating, and adapting to the effects of climate change on society and the environment. Many of the global climate models (GCMs) that are participating in the Intergovernmental Panel on Climate Change (IPCC) Fifth Assessment Report (AR5) do not fully resolve regional-scale processes and therefore cannot capture regional-scale changes in temperatures and precipitation. We use a regional climate model (RCM) to dynamically downscale the GCM’s large-scale signal to investigate the changes in regional and local extremes of temperature and precipitation that may result from a changing climate. In this paper, we show preliminary results from downscaling the NASA/GISS ModelE IPCC AR5 Representative Concentration Pathway (RCP) 6.0 scenario. We use the Weather Research and Forecasting (WRF) model as the RCM to downscale decadal time slices (1995–2005 and 2025–2035) and illustrate potential changes in regional climate for the continental U.S. that are projected by ModelE and WRF under RCP6.0. The regional climate change scenario is further processed using the Community Multiscale Air Quality modeling system to explore influences of regional climate change on air quality.

A multi-scale dynamical downscaling system has been set up to investigate future air quality trends in Melbourne, Australia due to climate and/or emission changes. The system consists of a comprehensive air emissions inventory; an ensemble of climate trends, a regional climate model for downscaling from synoptic to regional scale and a meteorological-chemical transport model for downscaling from regional to urban scale. Air quality projections for 2030 and 2070 suggest that, in the absence of emission controls, ozone concentrations will increase, leading to a 20–25 % increase in population exposure. The outcomes for PM2.5 show mixed results depending on season. The air quality trends with three different emission scenarios for 2030 were also modelled under the same climate projection. Some impact measures, such as average ozone concentration, are insensitive to the choice of emission scenarios, while others such as exposure to nitrogen dioxide show significant variations for different scenarios.

The differences between air quality in urban and rural regions has been investigated for the current situation using both observations and modeling and also for future climate conditions

The clouds that develop in maritime or polluted environments have significant differences in their properties. A number of modeling sensitivity tests have been performed to describe the physical processes related to aerosol – cloud interactions at various stages of cloud development. Precipitation amounts and cloud structure were found to be very sensitive to changes in the size distribution and number concentrations of the aerosols. Certain combinations of CCN/IN properties and atmospheric properties may lead to significant enhancement of convection and precipitation. These interactions are not linear and it is the synergetic effects between meteorology and atmospheric chemistry that are responsible for the variation of precipitation.

Nitrogen oxides (NO

x

) are the main ozone precursors, and NO

x

control programs in the US have led to substantial reductions in emissions. However, it is unknown whether these programs have optimally reduced ozone concentrations. Current control programs do not account for spatial and temporal specificities of NO

x

emissions. In this paper, this shortcoming in traditional trading systems is addressed and a methodology for identifying optimal NO

x

emission control strategies is developed. The proposed method combines an optimization platform with an adjoint (or backward) sensitivity model. Using the proposed method, a 2007 case study of 218 US power plants is examined. The results indicate that differentiating between emissions can significantly enhance environmental performance.

Adjoint sensitivity analysis of numerical models provides a platform for directly linking public health effects with air quality for evaluating emission control policies in a more straightforward manner. We link epidemiological and valuation statistics to the adjoint of CMAQ and calculate sensitivities of short-term mortality-related benefits in Canada, the U.S. and Europe to anthropogenic NO

x

and VOC emissions across two continental domains. Our results show significant spatial variability in impacts of NO

x

and VOC emissions reduction on short-term mortality. We estimate that sensitivities of mortality-related benefits to 10 % NO

x

emissions reductions in major cities reach monetary values in excess of $635K/day in Europe and $355K/day in North America. We find that when the cumulative effects of anthropogenic emissions on O

3

and NO

2

population exposure are considered, NO

x

emissions reductions generally yield higher mortality-related benefits than the same relative reductions in VOC emissions.

This study is aimed to assess the health impacts of outdoor PM

2.5

concentrations to entire population of Estonia. As the air quality monitoring network in Estonia is rather sparse (six urban, three rural and a few industrial sites), the exposure assessment was based on long-term modelling, controlled by monitoring data from existing stations. The model runs were performed with IairViro modelling system, including the Eulerian MATCH model for country-wide run with 5 km grid resolution and AirViro urban model with 200 m grid resolution for five major cities. The database of pollution sources includes industrial, transport and domestic heating emissions. The average annual PM

2.5

concentrations were found 7–9 μg/m

3

at most of rural and in 9–13 μg/m

3

typically at urban areas, up to 30 μg/m

3

in some parts of capital city Tallinn. To estimate the health risks, the base-line national health statistics and exposure-response coefficients from previous epidemiological studies (ACS, APHEIS), were applied. It was found that on average 600 (CI 95 % 155–1,061) premature deaths per year are caused by PM

2.5

pollution in Estonia, which has population nearly 1.3 million. On average 5 months of life are lost, maximum 14 months in some parts of Tallinn. About 900 additional hospitalizations due to pulmonary and cardiovascular diseases occur per year.

The results from epidemiology time series models that relate air quality to human health are often used in determining the need for emission controls in the United States. These epidemiology models, however, can be sensitive to collinearity among co-variates, potentially magnifying biases in the parameter estimates caused by exposure misclassification error or other deficiencies in the time series models by orders of magnitude. As a result, we examined collinearity among several covariates typically used in air quality epidemiology time series studies (ozone, fine particulate matter and its species, and temperature). In addition, we examined the ability of a bias-correction technique applied to estimates simulated by the Community Multiscale Air Quality (CMAQ) model to “fill-in” for the spatial and temporal limitations of observations for purposes of reducing exposure misclassification. Specifically, we evaluated whether the bias-adjusted CMAQ estimates could replicate the correlation among variables seen in the observations. The results presented are for a domain east of the Rocky Mountains for the entire 2006 year and indicate that collinearity among covariates varies across space.

Benzo[a]pyrene (BaP) is a highly carcinogenic substance that is created as an unintentional byproduct of combustion processes.

In order to simulate the development of European BaP concentrations between 1980 and 2020 a consistent emission dataset for the time span 1980–2010 was created using the SMOKE-EU emission model. Moreover, a variety of different emission scenarios for the year 2020 was created to estimate the future development of BaP emissions. The emission datasets have been used as input for a modified version of the Chemistry Transport Model (CTM) CMAQ. This CMAQ version is to our knowledge the only regional CTM to include the heterogeneous reaction of particulate BaP with ozone which, besides wet deposition, is the main sink of atmospheric BaP.

It was found that BaP concentrations have been decreasing strongly between 1980 and 2000. This is due to the large reduction of primary BaP emissions from industrial processes, residential wood and coal burning, and vehicle exhausts. Depending on the emission scenario, simulated BaP concentrations over Europe between 2000 and 2020 are changing by −25 to +5 %. Because further reduction of industrial BaP emissions is supposed to be low, the future development depends mainly on the amount of wood used for heat production. Also changes of emissions of ozone precursors showed to impact the degradation of BaP.

Finally, several regions were identified in which the annual BaP target value of 1 ng/m

3

was exceeded. In 1980 this was the case for the Po-valley, the Paris metropolitan area, the Rhine-Ruhr area, Vienna, and Madrid. Predictions for 2020 lead to exceedances in the Po-valley, the Paris metropolitan area, and Vienna.

Epidemiologic studies of air pollution have traditionally relied upon surrogates of personal pollutant exposures, such as ambient concentration measurements from fixed-site pollutant monitors. This study evaluates the performance of alternative measured and modeled exposure metrics for multiple particulate and gaseous pollutants, in the context of different epidemiologic studies performed by EPA, Rutgers/Rochester/LBNL and Emory/Georgia Tech researchers. Alternative exposure estimation approaches used, included: central site or interpolated monitoring data, regional pollution levels based on measurements or models (CMAQ) and local scale (AERMOD) air quality models, hybrid models, statistically blended modeling and measurement data, concentrations adjusted by home infiltration rates based on LBNL algorithms, and population human exposure (SHEDS and APEX) model predictions. The Emory/Georgia Tech team examined the acute morbidity effects of ambient traffic-related pollutants (CO, NOx, PM

2.5

and PM

2.5

EC) and ozone using time series analyses of emergency department (ED) visits and case-crossover analysis of implantable cardioverter defibrillator (ICD) detected ventricular arrhythmias in Atlanta, GA. The Rutgers/Rochester/LBL team examined the associations between PM

2.5

mass and its species with myocardial infarction (case-crossover study) and adverse birth outcomes (cohort study) in New Jersey. Initially, the various exposure indicators/metrics were compared in terms of their ability to characterize the spatial and temporal variations of multiple ambient air pollutants across the different study areas. These metrics were then used to examine associations between ambient air pollution and adverse health effects. Next, pollutant–specific relative risks (RRs) obtained from epidemiologic analyses of the alternative exposure metrics were evaluated against those obtained from using a conventional approach (i.e., central site data alone). Pollutant and metric dependent exposure prediction differences were found in some cases, indicating a non-uniform exposure prediction error structure across pollutants. Results suggest the need for additional refinements to methods used to estimate exposures in support of different types of air pollution epidemiologic studies.

The ability of four (4) enhancements of gridded PM

2.5

concentrations derived from observations and air quality models to detect the relative risk of long-term exposure to PM

2.5

are evaluated with a simulation study. The four enhancements include nearest-neighbor (NN or central monitor), ordinary kriging (OK), FUSED (bias-adjusted model output), and direct use of model outputs (CMAQ). The methods are applied to the state of New York. After adjusting for PM estimation bias and range modification, FUSED PM and CMAQ had similar performance and were better than kriged and nearest neighbor PM.

One of the measures that is promoted in order to improve the air quality is the decrease of vehicle mileage. However, there are different reasons to assume that not every kilometer driven by car yields the same impact on air pollution nor on the exposure of people. For instance, some trips are driven at high speeds on highways, while others are driven at relatively low speeds in urban environments. This will have an impact on the resulting emissions. Furthermore, emissions exhausted during the night will have a larger impact on the ground-level concentrations than emissions exhausted during the day, due to the higher atmospheric stability. And another aspect is that emissions produced in cities will affect a larger population than emissions in sparsely populated areas. These and other aspects are examined in this chapter. Analysis of the results per trip purpose shows that driving 1 km to work increases the pollutant concentrations about 1.3–1.4 times more than driving 1 km to go shopping, mainly due, but not confined to, the time of the day at which the activity is performed.

This chapter assesses the use of an Emission Trading Scheme (ETS) for the year 2020 to reduce the emissions from international ships entering European Community ports. Additional costs and benefits of including the maritime sector into a land-based ETS for NO

x

and SO

2

are quantified.

The study presents modelled critical deposition load exceedances for over 4,700 representatively selected forest plots in 21 European countries. It is based on measured soil data, different deposition scenarios and an application of the Simple Mass Balance (SMB) model. Effects of climate change on critical loads and exceedances are presented for 108 intensive monitoring plots in 17 countries. Results suggest hardly any more exceedances of critical loads for acidity in the near future. In contrast, even a maximum feasible emission reduction scenario which will leave 10 % of the forest sites unprotected against nitrogen effects by the year 2020. Full implementation of existing clean air legislation will result in 20 % of unprotected forest sites. Forests are less sensitive compared to other ecosystems as for these areas with exceedances are up to 58 %. Under a climate change scenario, decreasing critical loads suggest increasing sensitivity towards nutrient nitrogen inputs. When comparing critical load exceedances over the period 2020–2100, the share of ‘safe’ sites is assumed to decrease from 60 % (constant climate) to 50 % (climate change).

The total mass of ammonia emitted in Poland has dropped in recent years. Despite this decrease, Poland is still one of the European countries with the highest NH

3

emissions. Ammonia and ammonium air concentrations and deposition of reduced nitrogen in Poland were modelled with the Fine Resolution Atmospheric Multi-pollutant Exchange (FRAME) model with an increased resolution of 1 × 1 km for the years 2005 and 2008. The model results were compared with air concentrations measurements made at the EMEP sites. The modelled NH

3

concentrations were further used to calculate the CLEs exceedances for the Natura 2000 sites in Poland.

Extensive studies have been undertaken to assess the influence of industrial and transport emissions on the contents and compositions of storage proteins in the seeds of tree species – acacia (

Robinia pseudoacacia

L.) and horse chestnut (

Aesculus hippocastanum

) – in the Dnepropetrovsk area. These investigations have shown that specificity and variability in the component compositions of proteins in the studied plant varieties as well as the alterations in the types of distribution of their electrophoretic spectra and frequencies of their occurrence depend on the levels and nature of pollution suffered by the urban environment.

Air pollution models are still associated with large uncertainties. The accuracy of air pollution models depends on uncertainty in the input parameter values and errors in model formulation. Meteorological parameters are influenced by random factors. Emission parameters are usually estimated rather roughly. These uncertainties have to be taken into account when we solve the decision making problems in air quality management systems. For this purpose the methods of optimization under uncertainty and fuzzy decision-making models are used. The model of multi-objective optimization is developed for fuzzy decision making in information systems for air quality management.

The impact of aerosol radiation absorption on the heat budget and the dynamics of the PBL has been investigated using a Large Eddy Simulation model.

In this work we use the atmospheric chemistry general circulation model EMAC (ECHAM5/MESSy2 Atmospheric Chemistry) with new physical parameterizations of desert dust emissions. We assess the impact of the different physical parameterization schemes, highlight the importance of several initialization fields, identify the key-input parameters and explore the benefits and/or disadvantages of using a-priori sources of atmospheric dust particles. This presentation will discuss primarily the processes that lead to dust emissions and explore the difficulties that arise from a global parameterization applied to areas with different and heterogeneous soils (e.g. North African versus Asian deserts). The performed case studies and sensitivity tests include comparisons with in-situ measurements of dust concentrations and satellite retrievals of aerosol optical depth. The conclusions from this work include recommendations on selecting the input fields depending on the application, in order to improve the modelling of dust globally.

Particle number (PN) concentrations over Europe were calculated by the CTM LOTOS-EUROS using aerosol dynamics from M7 and a new particle number emission database. Results were compared with PN observations (EUSAAR). In autumn and winter the correspondence was better than in spring and summer, possibly due to a poor modeling of nucleation and the incomplete contribution of secondary aerosols. In addition, the emission inventory was used under certain assumptions with regard to (amongst others) the hygroscopicity of particles, and the particles properties of wood combustion products. Sensitivity tests were performed to verify the impact of uncertainties in these assumptions.

A new scheme for vertical dispersion and dry deposition of atmospheric aerosols was recently developed by the authors (Kouznetsov and Sofiev, J Geophys Res 117:D01202, 1, 2012). The vertical transport scheme and deposition scheme for smooth and water surfaces are based on the exact solution of the steady-flux equation. The scheme for deposition onto rough surfaces uses a universal empirical function and two parameters describing the surface: aerodynamic roughness length and a “collection scale” that describes the surface elements. In present paper we briefly outline the scheme with some improvements of accounting for the deposition due to inertial impaction.

Aerosol activation is a key process in aerosol-AQ cloud interaction. Although it is widely studied within the climate modeling community it has not been attracting significant attention within the air quality modeling community. In this study an off-line, sectional, chemically-speciated regional air quality model, AURAMS, has been used to assess the impact of aerosol activation on the modelled regional particulate matter (PM) mass concentration and size distribution. Asimple activation scheme based on an empirical relationship between cloud droplet number density and aerosol number density is compared to a more physically-based activation scheme. Model simulations were compared to aircraft observations obtained during the 2004 ICARTT field campaign. Modelled aerosol light extinction and column aerosol optical depth (AOD) were computed in three different ways in the current study. Two of them based on Mie calculations and one empirical reconstructed mass extinction method. The magnitude of the modeled AOD varies significantly depending on the approach. The impact of different aerosol activation schemes on the modelled AOD in this case is generally in the range of 20–30 % for the two Mie methods. As the empirical reconstructed mass extinction method is not size dependent, it is less sensitive to aerosol activation.

Dust is one of the most commonly found pollutants for many areas around the globe. Recent field and satellite observations suggest the importance of convectively generated downdrafts on the mobilization processes of dust particles and on their long range transport. RAMS/ICLAMS simulations of convective storms for the area of NW Africa allowed the description of the cool pool generation and the associated dust production mechanisms. Evaporation of rain droplets as they fall through the warmer and unsaturated air is responsible for the formation of a fast propagating density current. The inner structure of this system is characterized by increased turbulence, small scale vortices and strong updrafts at the edge of the front. Mobilization of dust occurs mainly in two ways: First, due to suspension of loose soil particles as the system passes over bare soil and sandy areas and second by sweeping pre-existing dust and creating a characteristic dust wall that reaches up to 3 km height.

An updated version of Weather Research and Forecasting/Chemistry model (WRF/Chem) with a new parameterization for the secondary organic aerosol (SOA) production based on the Volatility Basis Set (VBS) is evaluated over Europe in May–June 2003. Results are compared against surface observations issued from the EC/OC 2002–2003 campaign of European Monitoring and Evaluation Program (EMEP). WRF/Chem underestimates the elemental carbon (EC) and the organic aerosol mass (OA) by −7 and −38 %, respectively. The analysis of OA: EC ratio reveals that the under prediction may be mainly explained by a misrepresentation of anthropogenic emissions of carbonaceous aerosols due to the coarse resolution of the inventory. The modeled concentration of OA constituent is nearly constant during the day. The predicted SOA/OA ratio has a value of ∼80 %. The biogenic SOA (BSOA) are 30 % of the total OA mass. Dry deposition velocity of Volatile Organic Compounds (VOC) oxidation products is a source of uncertain in the SOA budget.

Clouds droplets are directly dependent on aerosols that are present in the atmosphere and on their probabilities to act as cloud condensation nuclei (CCN). So, it is necessary to study aerosols distributions native to different air masses in order to understand cloud droplets distributions. Sensitivity tests, using WRF‐chem (Weather Research and Forecast), should allow an improvement in the knowledge of aerosol properties and their impacts on cloud droplets and precipitation. From this perspective, simulations for various air masses are simulated with WRF-chem using RADM2 (Stockwell et al., J. Geophys Res, 95:16343–16367, 1990) chemical mechanism associated with MADE/SORGAM aerosol module (Ackermann et al., Atmos Environ, 32(17):2981–2999, 1998) which dissociates aerosols into three modes: Aitken, accumulation and coarse modes. Furthermore, comparisons are presented between model outputs and observation data from the puy de Dôme site. This site, part of the ACTRIS network, has been equipped with many probes in order to characterize the physical, chemical and optical properties of aerosol particles, to quantify gases (O

3

, CO, CO

2

, NO, NO

2

, …), to measure radiation and to document typical meteorological parameters and in particular cloud parameters (cloud water content, cloud droplet concentrations and mean radius).

A method to use fences of activated carbon fiber (ACF) is presented to reduce high NO

2

, PM, and VOCs concentrations in urban atmosphere without excess energy use. Polluted air in road space moves with natural and car-induced winds to contact with and to flow through ACF fences, and the ACF layer removes these pollutants. That this “energy-free” equipment can reduce high NOx by 15–50 % at the roadside is numerically evaluated.

The vehicular emission is the main source of fine particles in Brazilian Cities. A comprehensive study was performed from 2007 to 2009 with 24 h daily sampling of fine particles in an experimental site in six Brazilian capitals: Sao Paulo, Rio de Janeiro, Curitiba, Porto Alegre, Recife and Belo Horizonte. The polycarbonate filters collected at each site with Harvard sampling, were submitted to gravimetrical analysis for identification of PM2.5 concentration, to reflectance for Black Carbon concentration, to X-ray fluorescence analysis for elemental composition and to ion chromatography for an ion sand cations composition and concentration. The average PM2.5 concentration were 28, 19, 17, 17, 16 and 11 μg/m

3

in São Paulo, Rio de Janeiro, Belo Horizonte, Curitiba, Porto Alegre and Recife, respectively. Black Carbon accounted for approximately 30 % of the PM2.5 mass concentration in the more air pollution impacted cities: Sao Paulo, Rio de Janeiro and Belo Horizonte. The Black Carbon was used as a tracer for diesel fuel emission and biomass burning. The elemental chemical composition of the PM2.5 was used to identify source-related fractions of fine particles, by means of Receptor Models. The results were used to examine the association of these fractions with daily mortality in each of the six cities. Principal Matrix Factorization (PMF) was applied to the elemental concentration data in order to identify the sources of fine particles, specifically the participation of the vehicular emission. These results were compared to the previous analysis performed with Absolute Principal Component Analysis (APCA). The participation of the vehicular fleet to the PM2.5 mass concentration was significant, explaining in the most urbanized area even 40 % of its mass. These results show the relative importance of the vehicular emission to health injury.

The understanding and the anticipation of the environmental fallout in case of severe nuclear accidents with radioactive releases is crucial for the environment. In this study we aim to improve our knowledge on the aerosol particles scavenging, in particular the washout by raindrops with a diameter larger than 1 mm.

We describe the implementation of a Kalman filter-based adjustment scheme for the correction of deterministic air quality forecast results. This scheme exploits the information on the mismatch between the deterministic forecast and observations of the prior period to calculate correction regression coefficients for the next forecast step. This method was applied to ground-level daily O

3

and PM

10

concentration fields simulated by the regional-scale deterministic air quality model AURORA for the year 2007, for a domain covering Belgium, and employing observations from the AirBase data archive. From a cross-validation analysis, it was found that the correction method improved the accuracy of daily mean PM

10

and daily maximum O

3

concentrations substantially.

Within the EU-FP7 project PASODOBLE five AIRSHED services have been developed to bridge the gap between the GMES core atmospheric service at coarse resolution and the PASODOBLE regional/local/urban downstream services at high resolution. For these Airsheds regional air quality forecasts and analysis are produced at a horizontal resolution of 5–10 km. Within this presentation we will present the PASODOBLE Airsheds focusing on the issues encountered while developing the systems.

The nesting within the European ensemble forecasts from the core atmospheric service poses a problem due to timing and lack of available information in the ensemble forecasts. Recommendations for the ensemble forecasts have been made to include more components and the timing issue will be solved by the delivery of 96 h instead of 72 h ensemble forecasts. The data assimilation of observations to improve the forecasts forms another issue due to time consuming calculations and lack of timely observations.

Lotos-Euros is a chemistry transport model developed in the Netherlands, and is used for air quality assessments and forecasts. Operational air quality forecasts for the Netherlands concerning ozone and PM10 are made available on the RIVM webpage (

http://www.lml.rivm.nl/verw.html

) and are used to warn the population in case of predicted exceedances of air quality standards. Lotos-Euros is also contributing to the model-ensemble based air quality forecasts for Europe (MACC project,

http://macc-raq.gmes-atmosphere.eu/index.php?op=get

). Currently, the system is expanded to assimilate routine surface observations from European networks as well as satellite observations such as OMI NO

2

. The Ozone Monitoring Instrument (OMI) is a Dutch-Finnish instrument on the NASA EOS-Aura mission, and has a capability to detect boundary-layer NO

2

with a unique resolution of about 20 km. The assimilation in Lotos-Euros is based on the ensemble Kalman filter technique in which model parameters such as the NO

x

emissions, VOC emissions, PM sources, ozone boundary conditions and/or deposition velocities are adjusted to improve the comparison with the observations. In our contribution we will discuss the experiences with the assimilation of NO

2

tropospheric columns from the OMI instrument. In particular we will focus on the model-OMI comparisons over Europe, the analysis improvements, the analysed emissions and the impact of OMI NO

2

data assimilation on other constituents such as ozone.

Air quality managers have traditionally relied on deterministic and statistical approaches for forecasting concentrations. In some parts of the world, concentration forecasts are used to trigger urgent and sometimes drastic preventative measures in the hope that severe pollution episodes are alleviated or avoided altogether. These preventative measures, however, are often applied rather indiscriminately to all emission sources without accounting for specific influence of each individual source on the forecast outcome. This blanket approach fails to improve air quality in a cost-effective manner. Adjoint sensitivity analysis approach provides a unique framework for efficient forecast of the impact of individual sources on short-term air quality. We demonstrate how this method can be used for short-term emission behavior modification resulting in amelioration of pollution episodes.

To study and forecast atmospheric tracer concentrations at ground level, an assimilation system is available around the LOTOS-EUROS model based on the Ensemble Kalman filter technique. For applications focusing on air-quality related to aerosols, the available observation data is usually limited to ground based observations of total PM2.5 or PM10, and model uncertainty is specified for the emissions. In this study, the key parameters of the assimilation system have been varied: the assumed temporal variation in the emission uncertainty, the amplitude of the representation error, the localization length of the analysis, the averaging period of the observations, and the number of ensemble members in the filter. Although in theory these parameters are all important, the most important parameters are those related to the representation error between simulations and observations.

Since 2003, the national PREV’AIR system (

www.prevair.org

) has been delivering daily air quality forecasts of atmospheric pollutants (O

3

, NO

2

, PM

10

, PM

2.5

) over Europe and France. Those products are based on chemistry-transport modeling and in particular on the outputs from CHIMERE model. Analysed air quality maps of the previous day are also produced by mixing observed and simulated data through a geostatistical approach.

More recently a methodology has been developed to improve the accuracy of CHIMERE forecasts. It consists of two main stages.

In a preliminary stage, statistical short-term forecasting models are built and validated for each French and European rural or (sub)urban background monitoring site. The response variables are the O

3

daily maximum, NO

2

daily maximum and PM

10

daily average concentrations of the current and the next 2 days. This part is based on the developments carried out and tested within CITEAIRII project (

www.citeair.eu

).

In the operational stage, the statistical models identified as reliable enough are applied with a daily frequency to predict concentrations at the corresponding monitoring sites. Locally forecast concentrations are then combined with CHIMERE forecast concentration fields according to the same geostatistical approach as aforementioned. This has been tested so far for O

3

and PM

10

. Validation against independent data shows a significant improvement of the forecasts compared with raw CHIMERE outputs.

In this study the benefits of coupling data assimilation with ensemble forecasting are demonstrated for the production of improved air quality forecasts and detection of threshold exceedances.

Environment Canada routinely issues twice-daily, 48-h public forecasts of (a) gridded surface and near-surface O

3

, PM

2.5

, and NO

2

concentration fields made by the GEM-MACH15 on-line chemical weather forecast model on a 15-km North American grid plus (b) point-specific forecasts for Canadian cities of the national Air Quality Health Index (AQHI) prepared by a statistical post-processing package called UMOS-AQ. The AQHI is a health-based, additive, no-threshold, hourly AQ index that ranges from 0 to 10+ and is based on a weighted sum of local O

3

, PM

2.5

, and NO

2

concentrations. An objective analysis scheme for surface O

3

, PM

2.5

, and NO

2

, which will provide model-measurement data fusion and model error diagnostics, is now being tested. These recent advances as well as plans for further improvements to the AQ forecasting system are described.

The ensemble Kalman filter is a commonly used framework for chemical data assimilation. Random perturbations are required for the ensemble initial conditions and to account for model error. For a chemical transport model, such perturbations should represent appropriate scales of variation and correlations in the horizontal, vertical and chemical dimensions. We present a sampling scheme to generate normally distributed perturbations with covariances based on a climatological background covariance matrix, estimated with a spectral decomposition, assuming horizontally homogeneous and isotropic error correlations. We tested the sampling scheme with an ensemble Kalman filter coupled to the Danish Eulerian Hemispheric Model, a three-dimensional chemical transport model. Observations of CO were assimilated, leading to substantially reduced bias at surface monitoring stations.

The data assimilation method Ensemble Kalman Filtering has been used with the Chimere-model for tropospheric ozone over Europe and the results are presented and discussed.

Updatable Model Output Statistics (UMOS) (Wilson JW, Vallée M, Weather Forecast, 17:206–222, 2001) methodology in air-quality forecasting (UMOS-AQ) has shown great ability to improve direct model output. The UMOS-AQ (Antonopoulos S, Bourgouin P, Montpetit J, Croteau G, Forecasting O

3

, PM

25

and NO

2

hourly spot concentrations using an updateable MOS methodology. In: Proceedings for the 31st NATO/SPS international technical meeting on air pollution modeling and its application, Torino, Italy, pp 309–314, 2010) system produces one equation for each station, predictand, model run, forecast hour and season. A limitation of the method however is the fact that we only obtain point forecasts. An optimal interpolation solution on the UMOS-AQ forecasts using Simple Krigging is presented in which the model’s output is used as a trial field. The parametrization chosen is such that the radius of influence is approximately two grid points with most of the weight coming from the UMOS-AQ forecasts. Preliminary results showed significant improvements over the model’s forecasts in regions with a high density of observation stations.

The NOx-emission trend has been evaluated over Europe using the LOTOS-EUROS model and the NO2 tropospheric columns from OMI.

Operational forecasting of allergenic pollen faces several challenges, which are not common in the air chemical composition forecasting. One of them is practical inapplicability of the standard data assimilation (DA) methods, which originates from basic features of the problem. The pollen observations are manual and become available with a delay of about a week. The bulk of the information has daily resolution, whereas the diurnal variability of the concentrations exceeds one order of magnitude. Finally, the season duration at each particular place is rarely more than 1–2 weeks, which makes the 1–2 days old information outdated. Adaptation of the DA methods for pollen forecasting has to target the emission source parameters, which would have a longer lasting effect. Some of these parameters also have long correlation distance, which increases the value of sparse observational network.

In this study we assess the impact of improved meteorological data over the Milan area for the forecasting of air quality.

In this study we address the real-time emission capturing of intra- and exo-domain wild fires in the US National Air Quality Forecasting Capability (NAQFC).

The EMEP MSC-W group have as their main aim the support of air pollution policy in Europe, primarily under the Convention on Long-range Transboundary Air Pollution, but also for the European Commission. Traditionally, the EMEP MSC-W model has covered all of Europe with a grid-size of about 50 km, and extending vertically from ground level to the tropopause (100 hPa). The model has undergone substantial development in recent years, and is now applied on scales ranging from local (ca. 5 km grid size) to global. The main scientific challenges include those associated with e.g. formation of organic aerosols, but all activities are limited by uncertain inputs, particularly emissions (e.g. BVOC), and too few measurements of key compounds. In a longer-term perspective EMEP needs to develop links covering a range of scales and issues, including the interaction with Earth-system models.

An attempt has been made by a combination of theory, observations, and especially modeling to determine the concentrations of trace gases and aerosols in the troposphere, without the influence of human beings.

During the last decade, the efforts in the field of atmospheric modeling have been directed towards a joint ‘online’ approach of the meteorological and chemical processes in order to treat, in an integrated way, their complex and interactive links. This approach allows for the use of all meteorological fields as input in the chemical submodel, at each model time step, as well as takes into consideration the concentration of the pollutants on the meteorological processes (radiation and cloud microphysics). In this work, the air pollution over East Mediterranean is assessed with the use of the fully coupled, two-way interactive modeling system ICLAMS. Simulations have been performed for several test cases enhancing on the feedback mechanisms that occur. The summer case (July 2005) is used as base evaluation case due the increased insolation and high photolytic activity. The ozone model results are compared with measurements from stations from the European Monitoring and Evaluation Programme (EMEP). The correlation coefficient over all model domain reaches 0.7 with a root mean square error 22.78 μg/m

3

and mean bias 0.74 μg/m

3

. The spring case (April 2004) highlights on the impact of mineral dust and sea salt on the photolysis rates that in turn determine the concentration of ozone in the area. The presence of dust over Finokalia station (Crete, Greece) leads to a decrease in the photolysis rates of NO2 and O3. Another winter case has also been discussed regarding the role of anthropogenic aerosols on clouds and precipitation.

We conducted a study of ozone formation in the Lower Fraser Valley (LFV), using WRF-SMOKE-CMAQ models, observations and emission inventories in order to understand relationships between the reduction in both amount and location of precursor emissions and spatio-temporal changes in episodic ambient ozone concentrations over the last 20 years. A dynamical model evaluation shows that the modeling framework is able to capture the changes in both magnitude and spatio-temporal structure of ozone concentrations over the 20-year period. We model ozone formation for four episodes, which both capture the observed changes in ozone reduction and the different meteorological regimes that occur during LFV ozone episodes. The SMOKE emission inventories are adjusted to account for temporal changes in amount, and location of emissions, based on population shifts. Model runs allow us to isolate the effects of emission changes from meteorological changes. Results show that the western LFV has been, and remains VOC-sensitive; the central LFV has changed from VOC-limited to NOx-limited; and the eastern LFV has been, and remains NOx-limited. Analysis shows that the ozone production efficiency as a function of NO has increased noticeably in the eastern LFV. This has likely offset some of the benefits resulting from local NOx emission reductions.

Recirculation of pollutants is often invoked as a cause of degraded air quality episodes. In this modelling study, we investigated recirculation in order to explore its behaviour in the Lower Fraser Valley (LFV), British Columbia. HYSPLIT was used to produce trajectories from WRF output for seven severe episodes, covering the four main circulation regimes conducive to ozone episodes. Both internal and external recirculations within our domain of interest were observable, but they do not have the same frequency for all regimes.

The chapter presents an assessment of the continental scale atmospheric dispersion of volcanic ash and sulphur dioxide from the recent eruptions of the Eyjafjallajökull (2010) and Grimsvötn (2011) volcanoes. The study is based on a combination of modelling, remote-sensing, and in-situ observations: while the release height is obtained from in-situ observations, the emitted mass flux of SO2 and particulate matter is calibrated using the satellite-based instruments. The analysed features include the split of the Grimsvötn plume to high-altitude SO2 and middle-troposphere ash clouds and the temporal variation of the Eyjafjallajökull emission composition.

The Community Multiscale Air Quality (CMAQ) modeling system was applied to a domain covering the northern hemisphere; meteorological information was derived from the Weather Research and Forecasting (WRF) model run on identical grid and projection configuration, while the emission inputs were derived from global inventories. The ability of the model to represent long-range transport of pollutants is analyzed through comparisons with aircraft measurements from the 2006 INTEX-B field campaign, ozonesonde profiles, and remotely sensed observations of aerosol optical depth. Time varying lateral boundary conditions from these hemispheric scale calculations were used to drive regional-scale air quality simulations over a finer resolution domain covering the continental United States. Comparison of model predictions with surface O

3

and PM

2.5

measurements indicate comparable or better performance relative to other approaches (e.g., other global models, static profiles). The successful expansion of CMAQ to the hemispheric scales now provides a conceptual framework to examine interactions between atmospheric processes occurring at various spatial and temporal scales in a consistent manner.

Modelling has been performed of the European nitrogen budget, and the effects of including the bi-directional surface-atmosphere exchange of ammonia have been investigated.

The NMMB/BSC Chemical Transport Model (NMMB/BSC-CTM) is a new air quality modeling system under development at the Earth Sciences Department of BSC in collaboration with several research institutions. It is an on-line model based on the NCEP new global/regional Nonhydrostatic Multiscale Model on the B grid (NMMB). NMMB is an evolution of the WRF-NMMe model extending from meso to global scales. Its unified nonhydrostatic dynamical core allows regional and global simulations and forecasts. NMMB/BSC-CTM incorporates an aerosol module that simulates the global life cycle of mineral dust, sea salt, black carbon and organic carbon, and sulfate. Additionally, a gas-phase chemistry module has been implemented.

Detailed knowledge of sea-salt aerosol (SSA) space-time variations is essential for a deeper understanding of the process of SSA loading in the atmospheric boundary layer. In order to reproduce variability of SSA concentrations over the Mediterranean Sea, the regional DREAM-Salt model has been running daily at Tel-Aviv University since February 2006 (

http://wind.tau.ac.il/salt-ina/salt.html

). The model performance in producing accurate SSA forecasts over the Mediterranean Sea was evaluated using a 5-year record (2006–2010) of daily SSA mass concentration measurements at the island of Lampedusa. Model-vs.-measurement comparisons showed a distinct dependence of model performance on wind direction. On average, for wind direction from 30° to 300°, the model performance was quite acceptable. It was characterized by a relatively high correlation of over 0.65 and a rather small mean bias. For north winds (0°–30°, and 300°–360°), some discrepancy between modeled and measured SSA concentrations was observed. This was characterized by the model underestimation of SSA measurements and a rather low correlation between model data and measurements. Probably, for north winds, SSA production in the surf zone, located in the vicinity of the monitoring site, contributed to observed SSA concentrations.

We present an overview of results from an enhanced sub-grid scale approach to characterize air quality impacts of aircraft emissions at the Hartsfield-Jackson Atlanta International airport (in the U.S.) for June and July 2002 using an adaptation of CMAQ called the Advanced Modeling System for Transport, Emissions, Reactions, and Deposition of Atmospheric Matter (AMSTERDAM). Aircraft emissions during the landing and takeoff cycle (LTO) and below 3,000 m were represented as plume-in-grid (PInG) emissions using AMSTERDAM’s PInG treatment. Initial results from CMAQ-AMSTERDAM focusing on impacts from aircraft emissions to inorganic PM

2.5

and total PM

2.5

indicated aircraft increased average total PM

2.5

concentrations by up to 235 ng m

−3

near the airport and by 1–7 ng m

−3

throughout the Atlanta metro area. However, aircraft reduced concentrations by 0.5–1 ng m

−3

downwind of the airport, attributable to reductions in sulfate aerosol. The subgrid-scale concentrations were an order of magnitude higher than the grid-based concentrations due to aircraft. In an earlier study when aircraft emissions were modeled by CMAQ as traditional point sources within the ATL airport grid cell, we showed that modeled secondary organic aerosol (SOA) concentrations increased by 2 % due to primary organic aerosol (POA) emissions from aircraft, which provided additional surface area for SOA to partition onto. We now present results from additional modeling work performed to a) enhance organic treatment in a 1-D aerosol microphysics model used to provide engine-specific emissions parameters for CMAQ, and b) examine aircraft’s impacts on secondary organic aerosol concentrations using the volatility basis set (VBS) within CMAQ-AMSTERDAM to represent the formation and aging of organic aerosols. Parameterization for the VBS components was determined using current and ongoing field study measurements, chamber studies, and box models specific to aircraft SOA formation, which showed that non-traditional precursors of SOA (NTSOA) were a much higher contributor to aircraft-specific SOA than previously understood.

A three-dimensional chemical transport model (CTM), PMCAMx-2008, was applied to Europe to study the influence of emissions changes on fine PM levels. Various emissions scenarios were studied, covering summer and winter periods to quantify also the seasonal variation. Reduction of NH

3

emissions seems to be the most effective control strategy for reducing PM

2.5

over Europe, in both seasons, mainly due to reduction of NH

4

NO

3

. A reduction of SO

2

emissions has a significant effect on PM

2.5

levels over the Balkans during summer, due to decrease of sulfate, while the reduction of anthropogenic OA emissions has a strong effect on total OA mainly in areas close to emissions sources. The NO

x

emissions control strategy seems to be problematic in both seasons. Our analysis in European Megacities, based on a scenario zeroing all anthropogenic emissions, showed that the contribution of local emissions on total PM

2.5

, depends on the chemical component, with local sources being especially important mainly for black carbon (BC).

The LIFE + Environmental Policy & Governance project ATMOSYS has the objective to implement an advanced and comprehensive air quality modeling system as a web-based service used by policy makers. The ATMOSYS system is based on advanced technology, including prognostic 3-D atmospheric computer models, data assimilation techniques, CFD modeling, and on results from recent and on-going national and European research projects. Its comprehensive character resides in the multiple scales and scale interactions covered by the system and in a coherent approach for forecasts, assessments, and scenario studies alike.

Model calculations have been performed of the emissions of radioactive material from the Fukushima Daiichi nuclear power plant over the northern hemisphere.

Eleven state-of-the-science regional air quality (AQ) models, exercised by 20 independent groups in Europe and North America, have been assembled for the Air Quality Model Evaluation International Initiative (AQMEII). The modelled ground-level ozone mixing ratios are collectively examined from the ensemble perspective and evaluated against observations from both continents. We aim at creating optimized ensembles in order to capture the data variability while keeping the error low. It is shown that the most commonly used ensemble approach, namely the average over all available members, can be outperformed by subsets of members optimally selected in terms of bias, error, and correlation, independent of the skill of the individual members. A clustering methodology is applied to discriminate among members and to build a skilful ensemble based on model association and data clustering.

The impact of aerosol-radiation feedback on meteorology and subsequent changes in pollutant distributions over Europe and the North Atlantic were investigated with WRF/Chem. The case studies for two summer months show that the direct effect of aerosol particles on solar radiation is dominated by semi-direct effects that are developing after a few days. Strong deviations from the baseline case, which does not include any feedback, were found for the global radiation over the North Atlantic and Northern Europe when the indirect aerosol radiative effect is considered.

We present a regional emission inventory constructed based on satellite observations of fire activity (MODIS) and the ORCHIDEE vegetation model, and its application to air quality forecasting. After a brief description of the variability of fire activity in the Euro-Mediterranean region during the past 8 years, a full evaluation of the emissions is performed for the case study of the summer of 2007, during the large Greek fires event. Therefore, regional simulations undertaken with the CHIMERE chemistry-transport model (CTM) are compared to surface and satellite observations of trace gases and aerosols.

At the Royal Meteorological Institute of Belgium the chemical transport model (CTM) CHIMERE runs at a spatial resolution of 7 km. The CTM is first run on a coarse resolution of 50 km covering Western Europe. As a NWP input, the meteorological fields of ECMWF are consulted. The emission database is provided by EMEP. This model run is used as input for the boundary conditions for the coupled high resolution (7 km) chemical transport model run.

This one-way nested coupling with the high resolution limited area NWP model ALARO (spatial resolution of 7 km) improved already significantly the modelling of ozone concentrations by improving the physical parameterization, i.e. cloud cover, surface temperature, surface sensible heat flux, surface latent heat flux and relative humidity fields.

The model domain of ALARO covers Belgium and the Netherlands (680 × 680 km

2

). The emission database used as input for this high-resolution model is the TNO/GEMS emission database (Visschedijk AJH, Zandveld PYJ, Denier van der Gon HACA, High resolution gridded European database for the EU integrate project GEMS, TNO-report 2007-A-R0233/B, 2007).

We will present the influence from the improved physical parameterizations of the NWP model on air quality model forecast performance for two different time periods. Also the influence on the implementation of a new surface scheme (SURFEX including The Town Energy Balance (TEB) scheme) in the NWP limited area model will be shown.

Ten state-of-the-science regional air quality (AQ) modeling systems have been applied to continental-scale domains in North America and Europe for full-year simulations of 2006 for the Air Quality Model Evaluation International Initiative (AQMEII), whose main goals are model inter-comparison and model evaluation. Model simulations are inter-compared and evaluated with a large set of observations for ground-level particulate matter (PM

10

and PM

2.5

) and its chemical components. Analyses of PM

10

time series show a large model underestimation throughout the year. Moreover, a large variability among models in predictions of emissions, deposition, and concentration of PM and its precursors has been found.

The present work aims to studying the local to regional atmospheric pollution transport and transformation processes over Bulgaria and to tracking and characterizing the main pathways and processes that lead to atmospheric composition formation in the region. The US EPA Models-3 system is chosen as a modelling tool. As the NCEP Global Analysis Data with 1° resolution is used as meteorological background, the MM5 and CMAQ nesting capabilities are applied for downscaling the simulations to a 9 km resolution over Balkans and 3 km over Bulgaria. The TNO emission inventory is used as emission input. Special pre-processing procedures are created for introducing temporal profiles and speciation of the emissions. The study is based on a large number of numerical simulations carried out to four emission scenarios – with all the emissions and with excluded biogenic emissions, emissions from energetics and road transport. Some results from the numerical simulations concerning the main features of the atmospheric composition in Bulgaria and the contribution of the different emission categories are demonstrated in the paper.

Source apportionment/labeling has been inserted in the LOTOS-EUROS model, and tests have been performed to show the impact of different source areas and categories.

The results of numerical simulation of environment pollution after accident at the ammonia pipe are presented in this paper. The problem was solved for two different scenarios. Firstly only the ammonia ejection into the atmosphere was considered. 3D equation of pollutant dispersion (k – gradient model) and model of potential flow were used to simulate the process of air pollution. At the second step the problems of river Dnepr pollution and evaporation of ammonia from the water surface were considered. The developed numerical models and the code were used to calculate the scale of the air and water pollution. The code was used to calculate the toxic gas penetration into the dwellings of the settlements which are situated near the ammonia pipe. It allowed obtaining the information about the possibility of safety people evacuation.

Chemical composition of the troposphere is controlled by several major emission sources, such as anthropogenic releases, biogenic gaseous and particulate emission, wild-land fires, wind-driven dust and sea salt emissions, lightning, and transport from the upper atmosphere. These sources contribute to a comparatively limited list of inorganic species and a vast variety of organic substances, which are transported and mixed in the troposphere but also take part in chemical and physical transformations. Due to inter-relation of the transformation processes and synergetic effects of the pollutants impacting human health and ecosystems, modern chemical transport models include a wide range of species, as well as their interactions. This paper presents a new generation of dispersion model SILAM, an open-code system developed at FMI in collaboration with VTT Energy (Finland), Main Geophysical Observatory (Russia), University of Tartu (Estonia), Medical University of Wien (Austria), Danish Meteorological Institute (Denmark).

The aim of this work is the assessment of changes in modelled regional air quality, by the comparison of the EMEP inventory vs. the application of a regional inventory based in a mixed top-down and bottom-up methodology over the NW of the Iberian Peninsula. Air quality simulations were carried out with the CHIMERE model for an ozone episode in 2008. Comparison against measurements using the Delta Tool show better scores for the regional inventory daily maxima 8 and 1 h ozone simulation results.

The aim of this paper was to calculate emission, air concentration and deposition of cadmium (Cd), and lead (Pb) for Poland with the Fine Resolution Atmospheric Multi-pollutant Exchange model (FRAME). Calculations of anthropogenic sources of heavy metals (HM) were based onPM

10

emissionsand coefficients of activity and emission, dependent on SNAP sector and type of fuel. Natural emissions were estimated using PM

10

wind blown dust from the Nat Air project and HM concentration in topsoil, taken from the Geochemical Atlas of Europe. Emission maps were used in the FRAME model, and annual average concentration and total deposition of individual HM were calculated at a 5 km × 5 km spatial resolution. It was found that the regions with the highest values of HM concentration and deposition were Upper Silesia (industrial region), as also legally protected areas of National Parks.

The study is dedicated to perform a quantitative assessment of the present status and future prospects of the Baltic Sea (BS) area for two possible emission control measures. The consideration is given to environmental impact, mainly happening via total oxidised nitrogen (TON) deposition.

As part of the AQMEII project, vertical profiles of a number of variables simulated by several state-of-the-science regional-scale air quality (AQ) models are evaluated against measurements collected by instrumented aircraft over Europe (EU) and North America (NA) continental-scale domains for the full year of 2006. Tropospheric profiles of ozone, carbon monoxide, wind speed, temperature and relative humidity simulated by the AQ models at 12 selected airport locations in NA and 3 in EU are considered here for model evaluation. Moreover, in this study, several model outputs are inter-compared to examine the models’ ability to reproduce the observed variability for ozone.

Numerical implementation of CTM should satisfy a set of contradictory requirements related to the quality of the model predictions, minimization of the volume of computations, applicability of the model to different meteorological situations etc. In this paper we present several numerical recipes that were proven to be efficient in constructing CTMs and dispersion models developed at the Voeikov Main Geophysical Observatory (St. Petersburg, Russia) and Finnish Meteorological Institute (Helsinki, Finland). Assuming splitting, we discuss approximations of the physical processes separately but they can be used jointly as well.

A meteorological database of high resolution (∼10 km) climate modeling results for three time slices – 1960–2000 (Control Run, CR), 2020–2050 (Near Future, NF) and 2070–2100 (Far Future, FF), produced by ALADIN-Climate model following the IPCC scenario A1B, future concentrations of many air pollutants are estimated. The dispersion calculations are made basically by US EPA Models-3 system (MM5, CMAQ, SMOKE). The mean concentrations of four main polluters (ozone ADM, SO2, NO2 and PM10) for the last 10 years of each time slice are calculated. The changes in pollution levels are presented and commented.

The eruption of Island volcano in April 2010 caused enormously big troubles for air transport over Europe for a long period of time. The losses and inconveniences for air companies, common business and usual passengers are difficult to be estimated but in any case are rather considerable. The insights from this extraordinary event are that serious efforts must be put in studding not only the volcanic events but in creating tools for reliable forecast of volcano products (mainly ash) distribution in case of eruption. There are few centers devoted to observation and forecast of such events. Some meteorological services lately created respective systems. The ENSEMBLE consortium leaded by European JRC in Ispra, Italy, which is aimed at elaborating ensemble forecast on the base of individual forecasts of almost all European Early Warning Systems (EWS) in case of nuclear accident decided to launch a series of exercises devoted to simulation of the first week air pollution dilution caused by Island volcano eruption. Bulgarian ERS (BERS) was upgraded as to be able to take part in these exercises and its results and comparisons with other model results are the object of this work.

Comparative study of two chemical transport models of CMAQ and CAMx with WRF (Weather Research and Forecasting) model under the identical conditions of meteorological field as well as emission inventory was carried out to evaluate their performances to predict atmospheric aerosols in East Asia focusing on the Seoul Metropolitan Area. Predicted concentrations of PM10, PMC (particulate matter with less than 10 μm and larger than 2.5 μm), PM2.5, sulfate, nitrate, organic carbon, elemental carbon and soil component were compared with those measured in Korea especially with intensive measurements in the SMA. CMAQ and CAMx generally depicted the variation of temporal trend and spatial distribution of the aerosols components but the PM10 and PMC were intrinsically underestimated by both models due to inherent uncertainty and lack of information for emission data of fugitive soil dusts. CAMx has a tendency to predict higher PM species concentrations than those of CMAQ. According to soccer plots of fractional bias vs. fractional gross error in annual PM chemical species, both CMAQ and CAMx, however, are acceptable in the range of between good performance and average performance. It is inferred from this study that the main reasons to induce differences between CMAQ and CAMx in predicting PM species are due to dry deposition mechanism and nesting method.

In this study, we investigate the effects of revised Gothenburg target emissions on ozone and particulate matter in Europe with a focus on Switzerland by means of the 3-dimensional air quality model CAMx and the meteorological model WRF. We use the GAINS emissions for the CORINAIR SNAP sectors and the European countries to prepare gridded emission scenarios. In this paper, we discuss the project goals and preliminary model results for various emission scenarios with respect to the reference case.

With the endorsement and support from the U.S. Environmental Protection Agency, European Commission, and Environment Canada, a project entitled Air Quality Model Evaluation International Initiative (AQMEII) was launched in 2009 by bringing together scientists from Europe and North America (Rao ST, Galmarini S, Puckett K, Bull Am Meteorol Soc 92:23–30, 2011). Several regional-scale numerical photochemical models were applied over the North American and European domains with 2006 emissions inventory. Several papers resulting from this international collaborative effort were accepted for publication in the AQMEII special issue of

Atmospheric Environment

. Also, a large 4-D database, assembled by EU Joint Research Centre for the AQMEII project, is now available to all scientists interested in developing innovative model evaluation techniques (Galmarini S, Rao ST, Atmos Environ 45(14):2464, 2011). Having successfully completed the first phase of AQMEII, Phase 2 of AQMEII was launched at the 2011 AQMEII workshop in Chapel Hill, NC, USA to focus on the interactions of air quality and climate change. In Phase 2, coupled meteorology-atmospheric chemistry models will be exercised over the two continents with a common emissions database to assess how well the current generation of coupled regional-scale air quality models can simulate the spatio-temporal variability in the optical and radiative characteristics of atmospheric aerosols and associated feedbacks among aerosols, radiations, clouds, and precipitation. The results from AQMEII Phase 2 would be useful to policy makers for developing effective policies to deal with air pollution and climate change.

Traditional monitoring networks measure only total elemental carbon (EC) and organic carbon (OC) routinely. Diagnosing model biases with such limited information is difficult. Measurements of organic tracer compounds have recently become available and allow for more detailed diagnostic evaluation of CMAQ modeling results, which allow for more explicit representation of secondary organic aerosols. An enhanced version of the model makes it possible to track contributions from various sources of primary organic aerosols and elemental carbon, providing more in-depth evaluation of model biases. An ambient PM

2.5

measurement campaign conducted in four Midwestern U.S. cities in March 2004 February 2005 allows for direct comparison of modeled and measured organic carbon concentrations by primary and secondary source category.

A variant of the operational forecast configuration of the Met Office’s newly developed Eulerian Air Quality Forecast Model was used to generate an air quality hindcast for 2006 as part of a DEFRA model intercomparison. Verification of predicted ozone concentrations was carried out by comparing against hourly observations from 15 rural and urban background sites spread over the UK. Models were primarily assessed statistically using standard metrics including bias, mean error, correlation, and fraction of predictions within a factor of 2 of observations for (a) all observations, and (b) periods of elevated ozone (>100 μg/m

3

). We will present results showing that the Met Office model is competitive with other models for hourly ozone, but is best in class at modelling episodes of elevated ozone. The results indicate that the availability of high quality met data and interactive treatment of chemistry and meteorology are both important in modelling ozone episodes.

In order to assess the effects of model coupling on air quality simulations, the on-line meteorology and chemistry model WRF/Chem has been compared to the well known chemistry and transport model CAMx. Models have been applied over the Italian domain with a grid resolution of 15 km for both a winter and summer episode of 2005. The comparison of winter PM10 and summer O3 shows an overall agreement between the two models results, although some systematic differences have been highlighted. PM10 is usually higher in WRF/Chem than CAMx, especially in the Po Valley area for both winter and summer times. Models have been also evaluated against some measurement stations. Winter PM10 values are underestimated by the two models, while summer O3 data are generally overestimated.

Air quality model simulations are performed and evaluated for Houston using the Community Multiscale Air Quality (CMAQ) model. The simulations use two different emissions estimates: the EPA 2005 National Emissions Inventory (NEI) and the Texas Commission on Environmental Quality (TCEQ) Emissions Inventory. A comparison of predictions with observed data from the 2006 TexAQS-II Radical and Aerosol Measurement Project (TRAMP) suggest that while the predicted oxides of nitrogen are greater than observations, predicted volatile organics (e.g., ethane, acetone) are substantially lower than the observations. Predicted hydroxyl radical predictions are in good agreement with the observations. Hydroperoxy radical predictions, however, are substantially lower than the observations.

PMCAMx-2008, a detailed three-dimensional chemical transport model, was applied for the first time in Europe to simulate fine organic aerosol concentrations during the months of May 2008, February 2009, July 2009 and January 2010. The model includes a state-of-the-art organic aerosol module which is based on the volatility basis set framework treating both primary and secondary organic components as semivolatile and photochemically reactive. The model predicts that fresh primary OA (POA) is a small contributor to organic PM concentrations in Europe during the summer and spring, and that oxygenated species (oxidized primary and biogenic secondary) dominate the ambient OA. During the winter model results suggest a significant underestimation of OA emissions from wood burning. Use of very high resolution emissions inventory does not result in a significant improvement of the predictions of the model over the Megacity of Paris.

In this study, we analyze the impacts of perturbations in meteorology and emissions and variations in grid resolution on air quality forecast simulations. The meteorological perturbations considered in this study introduce a typical variability of ∼1 °C, 250–500 m, 1 m/s, and 15–30° for temperature, PBL height, wind speed, and wind direction, respectively. The effects of grid resolution are typically smaller and more localized. Results of the air quality simulations show that the perturbations in meteorology tend to have a larger impact on pollutant concentrations than emission perturbations and grid resolution effects. Operational model evaluation results show that the meteorological and grid resolution ensembles impact a wider range of model performance metrics than emission perturbations. Probabilistic model performance was found to vary with exceedance thresholds. The results of this study suggest that meteorological perturbations introduced through ensemble weather forecasts are the most important factor in constructing a model-based O

3

and PM

2.5

ensemble forecasting system.

The current study proposes a novel approach for the multi-model ensemble to be applied in air pollution forecasting. The methodology is based on decomposition of air pollution time series on different components (short-term, daily fluctuations, synoptic scale, etc.) and calibration of the ensemble for each of these components independently taking into account the performance of individual predictors. Therefore, the same model may have a different contribution for the ensemble at high and low frequency fluctuations. The Kolmogorov-Zurbenko (KZ) low-pass filter is used for the time series decomposition. The Fourier analysis is implemented to determine the contribution of different frequencies to the data variance allowing better understanding of the model performance and to define the ensemble weights. The methodology was tested using a group of four different air quality models that were applied over mainland Portugal for the 2006 year, and for main pollutants like O

3

and PM10. The approach implemented in this work was compared with one of the most used ensemble technique showing clear advantages.

An evaluation has been made of a range of simple and complex atmospheric transport models, applied to estimate sulphur and nitrogen deposition in the UK in order to provide information to policy makers to support decisions on future model use. The models were evaluated by comparison with annually averaged measurements from the national monitoring networks. A number of statistical metrics were output to assess model performance and the models were compared graphically by plotting cross-country transects of concentrations in air.

Photochemical model evaluation has been performed by clustering all relevant parameters.

A modeling system has been developed to predict accurately the downwind air quality impacts of prescribed burns. The system has been evaluated in applications to monitored burns and a long-range smoke event detected by the regional PM

2.5

monitoring network in Southeastern USA. Uncertainties in the estimation of emissions have been identified and sensitivities of predicted PM

2.5

levels to smoke injection height versus PBL height, and wind speed and direction have been quantified. More accurate wind predictions, currently provided by WRF, would significantly improve the performance of the modeling system.

Sulphur dioxide (SO

2

) and nitrogen oxides (NO

x

) form two of the largest contributors to PM2.5 in Europe: ammonium sulphate, (NH

4

)

2

SO

4

, and ammonium nitrate, NH

4

NO

3

. Understanding the formation regimes for these components is important for the achievement of the reduction objectives established in the European legislation for PM2.5. The present work uses the CALIOPE air quality forecasting system to investigate the formation of secondary inorganic aerosols (SIAs) and their gaseous precursors over Europe. Results points out that the continental regions in Europe tend to be HNO

3

−

limited for nitrate formation. Regulatory strategies in such regions should focus on reductions in NO

x

rather than NH

3

to control NH

4

NO

3

.

The effect of using in dispersion modeling different parameterizations for the wind velocity fluctuations standard deviations is investigated for low-wind conditions.

In this work we explore the effectiveness of theoretical information approaches, based on uncorrelation and Mutual Information, for the reduction of data complexity in multi-model ensemble systems. A distance matrix, measuring the inter-dependence between data, is derived, with the scope of clustering correlated/dependent models together and selecting a few representative models from each cluster. We test the efficacy of these distance measures to detect dependence structures and select “representative models”. We apply this analysis in the context of atmospheric dispersion modeling, by using the ETEX-1 dataset and the data made available by the more recent AQMEII (Air Quality Model Evaluation International Initiative) efforts. We show how the selection of a small subset of models data, according to uncorrelation or mutual information distance criteria, usually suffices to achieve a statistical performance comparable to, or even better than, that achieved from the whole ensemble dataset, thus providing a simpler description of ensemble results without sacrificing accuracy

The chemical segregation of isoprene has been investigated over heterogeneous surface conditions, and first results are presented and analysed.

The Chemistry Transport Model (CTM) REM-Calgrid (RCG) has been applied to investigate the non-linear relationship between emission changes and modelled SIA (=SO

4

+ NO

3

+ NH

4

) concentrations for a high PM10 episode in spring 2009. Emissions were reduced for the model domain only and the model domain including Boundary Conditions. It was found that the effectiveness of emission reductions increases with increasing emission reduction area whereas results were least dependent on the size of emission reduction area for ammonia emission changes.

Simulated meteorological fields, ground values of aerosol concentrations (PM2.5) and aerosol concentration profiles, all calculated within the Air Quality Model Evaluation International Initiative (AQMEII) for the year 2006 for Europe are used to investigate the relation of the height of the planetary boundary layer (PBL) to the modelled aerosol concentrations. It was found that in many cases simulated PBL heights match the observed values at Lindenberg, Germany, quite well at noon and under convective conditions. On the other hand, in winter and at night time, when the atmosphere is more frequently stably stratified, the modelled values underestimate the observed PBL height quite significantly. A strong relationship between the simulated PBL height and the aerosol concentrations at ground was found. While those models which showed good agreement with the observed PBL height typically underestimate PM2.5, too low PBL heights were connected with higher aerosol concentrations close to ground and therefore lower bias. Model biases in PBL height and in PM2.5 concentration at ground were found to be highly anti-correlated.

Model evaluation studies are essential for determining model performance as well as assessing model deficiencies, and are the focus of the Air Quality Model Evaluation International Initiative (AQMEII). The chemistry-transport model system COSMO-MUSCAT participates in this initiative. In this paper the robustness and variability of the model results against changes in the model setup are analyzed. Special focus is given to the formation of secondary particulate matter and the ability to reproduce unusually high levels of PM10 in Central Europe caused by long-range transported smoke of widespread agricultural burning and forest fires in western Russia. Seven different model configurations are investigated in this study. The COSMO-MUSCAT results are evaluated in comparison with ground-base measurements in Central Europe. The analysis is performed for two selected periods in April/ May and October 2006 which are characterized by elevated concentrations of PM. The model sensitivity is studied against changes in the used grid resolution, the meteorological forcing and the applied aerosol module. Possible reasons for differences in model results will be discussed.

This study shows long-term ABL wind profile features by comparing long-range wind lidar measurements and the output from a mesoscale model. The study is based on 1-year pulsed lidar (Wind Cube 70) measurements of wind speed and direction from 100 to 600 m with vertical resolution of 50 m and time resolution of 10 min at a coastal site on the West coast of Denmark and WRF ARW (NCAR) simulations for the same period. The model evaluation is performed based on wind speed, wind direction, as well as statistical parameters of the Weibull distribution of the wind speed time series as function of height. It is found that (1) WRF is generally under predicting both the profiles of the measured wind speed, direction and power density, (2) the scatter of observations to model results of the wind speed does not change significantly with height between 100 and 600 m, and (3) the scale (A) and shape (k) parameters of the Weibull distribution above 100 m. The latter signifies that the model suggests a wider distribution in the wind speed compared to measurements.

A comparison of the results from two independent CMAQ modelling systems for the same UK 2003 scenario suggests the need for improvement to temporal profiles of the input emissions data used by atmospheric models.

In this study, we use measurements performed under the MOZAIC program to evaluate vertical profiles of meteorological parameters, CO, and ozone that were simulated for the year 2006 with several versions of the WRF/CMAQ modeling system. Model updates, including WRF nudging strategies, boundary conditions, lightning NO emissions, and vertical transport were found to affect the simulated profiles. MOZAIC descent/ascent pairs provide a unique dataset to investigate PBL structure and its representation in regional-scale models during afternoon and early evening hours. Other data sources (e.g. ozonesondes, field studies such as DISCOVER-AQ) can be used to complement MOZAIC measurements in evaluating upper air predictions by regional-scale models over various airsheds.

In some areas of Slovenia air pollution is still a major problem due to industrial sources, traffic, local domestic heating and complex topography. PM10 pollution is a severe problem as there are many contributors such as industry, domestic heating, traffic and other. PM10 air pollution is a problem in several Slovene cities and industrial regions. In the paper a national research project with the title “Prognostic and diagnostic integrated small scale air pollution modeling system« is described. Such modeling system can significantly contribute to proper understanding of the air pollution in smaller areas with very complex topography. As an example for system development and validation the industrial region Zasavje was chosen which is characterized by highly complex terrain that makes the problem very difficult.

Modelling air quality requires the description of a large number of processes interacting each other. In order to properly model concentrations of atmospheric pollutants it is crucial to have a realistic reproduction of meteorological parameters, which can be critical in areas presenting a complex orography like the Italian peninsula. This work shows an analysis of the results obtained with the national model MINNI at two different horizontal resolutions (20 and 4 km), for a whole year over Italy. Comparisons between modelled and observed temperature and pollutants concentrations are carried out. The prediction of temperature is improved with the increase of model spatial resolution, as it is for pollutants like NO

2

and CO, while the improvement is not always evident for O

3

concentrations. Results are discussed providing an interpretation of the observed features.

The SILAM model is running in Finnish Meteorological Institute (FMI) for European air quality forecasts (http://silam.fmi.fi). The 7 km resolution of both TNO-MACC emission grid and Nordic model domain is too coarse for urban and industrial areas. The Estonian modelling domain, operated jointly by the University of Tartu and Estonian Meteorological and Hydrological Institute (http://meteo.physic.ut.ee/silam, resolution 3.3 km) includes the Baltic countries and southern Finland. In this study the database of pollution sources on the territory of Estonia is refined and performance of SILAM with the new database is studied. The new database presents better spatial distribution, differences in yearly emission budgets should be looked into in further studies. Average concentrations of NO

x

, SO

2

and particulate matter are reproduced rather well, NO

x

underestimated slightly and particulate matter significantly. The new database shows an improvement in hourly and daily correlations, especially for urban stations.

The accurate simulation of PBL dynamics is the most critical aspect of meteorological physics for air quality modeling. In this work, the performance of WRF meteorological model in three different types of high-O

3

episodes (Saavedra S. Characterization of tropospheric ozone episodes in Galicia by applying different simulation techniques and interpretation of field data. Ph. Degree thesis (in Spanish), University of Santiago de Compostela, 2010) in the northwest region of the Iberian Peninsula (IP) is evaluated. Four different planetary boundary layer (PBL) schemes were tested. Surface temperature and wind (speed and direction) simulation results were compared to hourly measurements at 10-m meteorological towers. Following statistics proposed by Emery (Enhanced meteorological modeling and performance evaluation for two Texas ozone episodes. Prepared for the Texas natural resource conservation commission, by ENVIRON international corporation, 2001) for wind speed and temperature, the best model performance is achieved using the Yonsei University (YSU) scheme; Asymmetric Convective Model (ACM2) is more suitable for temperature, and Mellor-Yamada Nakanishi and Niino Level 2.5 PBL (MYNN) are better for wind speed.

The WRF model has been used to make forecasts for ozone, using an artificial neural network.

In this work we present a new model and simulations for more realistic scenarios incorporating in the diffusive model the dependence of the eddy diffusivity profile on the temporal variable. Numerical results and comparison with experimental data are presented.

The present study proposes a mathematical model for dispersion of contaminants in low winds that takes into account the along-wind diffusion. The solution of the advection-diffusion equation for these conditions is obtained applying the 3D-GILTT method. Numerical results and comparison with experimental data are presented.

In this study, we employed Taylor’s statistical diffusion theory, a model that describes observed turbulent spectra and an analytical dispersion model, to investigate turbulent statistical characteristics and diffusive properties associated with the north wind phenomenon in southern Brazil. Numerical results and comparison with the Prairie-Grass experiment are presented.

Physical and chemical characteristics within the two distinct meteorological regimes, i.e., convective and stable atmospheric conditions in a complex highly urbanized terrain of the California South Coast Air Basin (CSCAB; the Los Angeles area) were investigated. The Community Multi-scale Air Quality (CMAQ) model was used with a horizontal resolution of 5 km × 5 km to produce the 3D fields of pollutant concentrations. Input meteorological fields were obtained by the MM5 numerical weather prediction model while the input emissions were provided by the Californian Environmental Protection Agency. Modeled meteorological surface parameters and their vertical profiles as well as modeled planetary boundary layer heights (PBL) were compared to the corresponding measurements. The CMAQ simulations of ozone concentrations were compared against the relatively large number of measurements from the CSCAB area. The main goal of the research was to identify the governing atmospheric processes and sources in the coastal area that contributed to the high levels of pollution and to investigate the air quality model’s capabilities to simulate the air quality in the complex topography.

Despite the advent of high performance numerical tools, the spatial gap between meso and microscale models is still a challenge in atmospheric modeling. The way boundary conditions are prescribed is crucial for the correct linkage between scales, especially in inhomogeneous canopies, which induce complex airflow dynamics and pollutant concentration patterns. The purpose of this work is to evaluate the sensitivity of microscale modeling results to distinct methods of deriving inflow vertical wind profiles in urban areas. The methodology involves the comparison of three methods for prescribing inflow profiles: (1) Log wind profile based on meteorological observations, (2) the mesoscale model WRF running at 200 m resolution, and (3) the LES model ARPS. The different profiles obtained were fed into the CFD model VADIS. The comparison of results against measured SF

6

concentrations shows that ARPS is capable of accounting for upwind region characteristics, providing more reliable inflow conditions. Moreover, VADIS has provided an enhanced and detailed insight over the 3D wind flow behaviour, accounting for the effects of buildings and trees, and its effect over dispersion.

A Lagrangian particle dispersion model (LPDM) driven by velocity fields from large-eddy simulations (LESs) is used to determine the mean and variability of plume dispersion at local and regional scales in a highly convective boundary layer. Comparison of the LPDM-LES results with data from two field experiments shows that they match the observations reasonably well.

The paper evaluates the air pollution budget of Paris by means of multi-domain nested simulations with detailed import-export fluxes delivered by the SILAM model. It is demonstrated that as a grand-total, Paris is a net exporter of nitrogen compounds, several VOCs and primary PM emitted from anthropogenic sources or produced chemically on a short time scale (a few hours). The city is a net importer of NH

3

. The budgets of pollutants with long lifetime and high free-troposphere concentrations (CO, O

3

) were shown to depend on height.

The purpose of the study was to establish an efficient height for block buildings used as geo-data by simplified urban dispersion models when the buildings in the Real world could have gabled roofs or carry a lot of objects as chimneys, antennas and other things. Assuming that CFD with RANS models can mimic the Real world, comparisons are made on the flow around buildings. Efficient height is defined as the height of a flat roof building when the flow around it most resembles the flow around a building with another shape. The results indicate that the efficient height should be somewhere between just below rooftop height and rooftop height + 5 m. But most of the results indicate that the efficient height should be close to the roof top value.

Assessing local air quality can be a challenging task. Indeed, local air quality is strongly dependent on local factors but also regional and in some cases even global effects have to be taken into account when assessing local air pollution concentrations. Furthermore, large gradients in pollutant concentrations can be present in the urban environment. In order to assess the local air quality for the city of Antwerp, a combination of an Eulerian dispersion model, a measurement interpolation tool, a Gaussian plume model and a simplified version of the OSPM street canyon model have been coupled to each other, taking into account double counting effects of local emissions. The coupled model which combines the regional, urban and street canyon scale has been applied for the city centre of Antwerp and its harbour. This results in detailed maps with a resolution up to 30 m for four pollutants: PM

10

, PM

2.5

, EC (elementary carbon) and NO

2

. Furthermore, several abatement measures have been assessed in order to improve the urban air quality. It has been shown that local (traffic) measures only have a small effect on total mass PM

10

and PM

2.5

concentrations, but exhibit a larger effect on EC and NO

2

-concentrations.

The Town Energy Balance (TEB) single-layer scheme is implemented into a numerical weather prediction model running operationally at ∼4 km resolution. The primary question addressed is the potential use of TEB, at this relatively coarse resolution, in an operational configuration to improve sensible weather performance over Belgium. Simulations with and without TEB are first evaluated against 2 m observations and wind above the urban canopy for 2 months (January and July 2010). The 2 m temperature and 2 m relative humidity improve compared to measurements in urban areas. The comparison of wind speed and wind direction above the urban canopy indicate that the structure of the flow in urban areas is better reproduced with TEB. Results from a 36-h case study during a high heat day (8 July 2010) indicate that even at this coarse resolution, TEB is able to reproduce correctly the intensity of the observed UHI of Brussels.

The objective of this experimental study is to determine processes of pollution ventilation above the X-shaped street intersection in an idealised symmetric urban area for several approach flow directions. A unique experimental set-up for simultaneous measurement of the flow velocity and the tracer gas concentration in a high temporal resolution is assembled. Advective and turbulent vertical scalar fluxes are computed from the measured data in a horizontal plane above the street intersection. Vertical turbulent pollution transport was found to be a significant and positive contribution to the total vertical transport of pollutants form the intersection. Prevailing events in vertical scalar and momentum fluxes were determined using quadrant analysis.

Eulerian models, which are widely used in air quality modeling are based on a three dimensional grid in which emissions are released. Point sources are mixed in a cell that can be a few kilometers long and therefore in regional scale simulations, the approximation made by the model on the emissions can have a significant impact. Plume-in-grid (PinG) models combine an Eulerian and a plume or puff, typically Gaussian, model aiming at having a better representation of sources within grid cells. Although point sources are convenient to model dispersion from a chimney, line sources are better suited for roadway traffic pollution modeling. Here we present a new PinG model that uses Gaussian line source model embedded within the Eulerian model Polair3D. Emissions are treated as steady-state plumes released from line sources. After a description of the model we present here its application to a case study that includes multiple roadways.

The present paper describes the some results of the modelling system for operational response to accidental releases of harmful gases in the atmosphere (as a result of terrorist attack or industrial accident). The system is based on the following models: WRF, used as meteorological pre-processor; SMOKE – the emission pre-processor; CMAQ – the Chemical Transport Model (CTM) of the system. For the needs of the emergency response preparedness mode the risk is defined as probability the national regulatory threshold values for toxic gases to be exceeded. Maps of the risk around potential sources of emergency toxic gas releases are constructed and demonstrated in the present paper. Some examples of the system “operational mode” results are demonstrated as well.

The application of Lagrangian dispersion models, as CALPUFF, at local scales requires as input accurate and very high resolution meteorological fields. In these high resolution applications, the computational cost of numerical weather forecast models, as WRF, recommends the steady-state nesting of a diagnostic model, as CALMET, in order to properly consider the influence of land use and terrain topography over complex terrain domains.

In this work, CALMET diagnostic model is nested to different WRF model simulations over a complex terrain and coastal domain around As Pontes Power Plant during three different periods when some primary pollutants glc peaks were detected. Different horizontal resolutions for the WRF simulations were considered as data input to CALMET. For the meteorological models validation and intercomparison, temperature, wind speed and wind direction statistics were obtained by comparison against measurements from surface meteorological stations located in the domain.

This paper presents verification results of numerical simulations of meteorological conditions leading to an air pollution episode on 6–9 November 2010 in Istanbul. WRF-ARW ver 3.3 is run with GFS and ECMWF input. For verification, horizontal wind components and temperature data of nine meteorological stations are used for both simulations. Results indicate that the model is successful in simulating the meteorological conditions with both two different input data, especially with the GFS input, although ECMWF runs had a finer horizontal resolution of 1 km.

The new version of WRF Modeling System (ver. 3.3) is run on nested domains of resolution 27, 9, 3 and 1 km. The three innermost domains with resolution of 1 km overlap three very different topographical areas with the biggest Israeli cities Jerusalem, Tel-Aviv and Haifa. Namely, Jerusalem is located on the southern spur of the Judean Mountains with the elevation of 700–800 m. The Great Tel Aviv urban area includes Tel Aviv itself and several satellite towns. The urban area lies about 20 km along the Mediterranean shoreline at coastal plain and extends about 10 km inland. Haifa is situated on the slopes of the Carmel range, stretching from the Mediterranean shoreline towards the southeast. The object of study is the high resolution pattern of wind and temperature fields, vertical turbulent structure within and over the urban canopy and the urban heat island (UHI) effect. The simulated by the WRF-ARW model wind field and atmospheric turbulent structure are the base for air pollutants transport and diffusion within urban canopy and outside the city. A coupled WRF-Chem modeling system is applied to analyze the temporal and spatial dynamics of pollutant concentration fields, first of all nitrogen oxides from traffic exhausts. The general goal of the study is to develop an easy-to-use tool for urban air quality forecast, based on the operative short-term weather forecast.

The CERES platform developed by CEA to assess the atmospheric dispersion and the environmental impact has been coupled with the Model of Multiphase Cloud Chemistry (M2C2) developed by LaMP in order to take account of the chemical reactivity during dispersion. The method used consists to provide lookup tables of the concentration gradient issued by M2C2 to characterize the chemical reactivity in the atmosphere then correct each puff during atmospheric dispersion calculations. A case of NH

3

release is presented and shows that NH

3

concentration in the air is less important taking account of the chemical reactivity, NH

3

undergoing photolysis and oxidation by radicals. Several tests are ongoing to validate the method and improvements have to be done to reduce approximations.

This work presents results of numerical simulation of turbulent flow in a fully developed flow in a series of street canyons. The motivation for this numerical study was an experiment by Kellnerová et al. (Detailed analysis of POD method applied on turbulent flow. In: EPJ web of conferences, vol 25. http://dx.doi.org/10.1051/epjconf/20122501038, 2012). The flow, as well as scalar fluxes, is investigated using proper orthogonal decomposition and spectral analysis. The flow is compared for two shapes of buildings.

In order to account for health impact of radioactive, chemical or biological releases in the urban environment, the CERES platform used by CEA has been improved by the coupling with SIRANERIK which is a model dedicated to pollutant dispersion over urban areas in unsteady situations. In case of a crisis, CERES is able to calculate very quickly noxious species concentration and to evaluate health impact of the releases. Thanks to its graphical interface, cartographies are provided as support to decision process.

High winter-time PM

10

, sulfate, nitrate and ammonium levels in Istanbul were investigated using a high resolution WRF/CMAQ mesoscale model system. The results suggested that the system was capable of producing the magnitudes. PM

10

levels calculated by the model underestimated the observations with an average of 10 % at Bogazici University sampling station, whereas an overestimation of 12 % is calculated for all stations. Base case results together with the sensitivity studies pointed significant contribution of local sources.

A three dimensional tool coupling the thermal energy balance of buildings and the modeling of the atmospheric flow in urban areas has been developed. In this work the effects of complexity in real urban geometry on the interaction between airflows and radiation exchanges with different surfaces is tested. The model results are encouraging and give insight into local surface-atmosphere processes, but further and more rigorous testing has to be performed with other datasets.

This work investigates the impact of terrain heterogeneity and stability conditions (stable and unstable) on the local turbulence kinetic energy (TKE) around an industrial site using 36 months of data collected at the experimental site SIRTA (Instrumental Site of Atmospheric Research by Remote Sensin). These measurements show that the normalized TKE (TKE/U

2

: where U is the wind speed) is highly depended on the upstream complexity of the terrain (presence of forested areas or buildings) whatever the stability conditions (stable, unstable). The highest values of normalized TKE are obtained for unstable thermal stratification for all wind direction sector in the zones located behind the forested areas.

Tree basic parameterization schemes: bulk Richardson number method (Rb-method), PBL resistance laws (RL-method) and combined (Rb-RL)-method are modified and applied to areas with large roughness (urban canopy, forest, buildings, etc.), accounting for a wide range of turbulent PBL urban regimes.

Dispersion of liquid or solid aerosol pollutants via an explosive device is a multiple scale process. We model the blast wave propagation with a BKW equation of state using a modification to the hydrodynamic tool box OpenFOAM. (OPENFOAM® is a registered trade mark of OpenCFD Limited, the producer of the OpenFOAM software.) We proceed into a BKW modified Boussinesq phase using the standard k-ε model for turbulent dispersion. The agreement of the detonation product cloud top final height of the computation and field observation is to within 10 ÷ 20 %.

A street canyon model has been used to calculated concentrations of NO2 in a specific situation with detached houses and a specific traffic lane width.

An investigation has been performed on the increased dispersion of a plume from a tall stack in the vicinity of a wind turbine.

Transports of substances to or from Seveso classified facilities are always a dangerous link in today’s industrial societies. Despite the risk of serious accidents, it is common that these transports use routes that pass through densely populated urban areas. We present a computational concept employing a multi-model simulation platform that allows for simulation of diverse scenarios. This platform allows us to simulate severe incidents in a city including leakage from either a stationary or moving transport vehicle, dispersion of the released substance, and finally the injury panorama of the affected population. Different scenarios of leakage are studied in combination with various environmental conditions regarding parameters such as wind strength and direction. The dispersion model utilizes a synthetic 3D model of the city in which the wind driven convective fields spread the released substance. In this model, a large number of virtual habitants are allowed to walk on the streets following individual paths while obeying a set of predetermined rules of behaviour. Each individual accumulates an individual dose of the agent by inhaling the harmful substance and is subject to injury scaling from no harm at all to lethal injury. The total expected injury panorama of the population in the city, for the current scenario, is finally obtained from the statistics of the synthetic actors that represent the real population.