Air Pollution Modeling and its Application XXIII

The most recent Global Burden of Disease study (Lim SS et al, Lancet 380(9859):2224–2260, 2012), for example, finds that combined exposure to ambient and indoor air pollution is one of the top five risks worldwide. Of particular concern is particulate matter (PM). Health researchers are now trying to assess how this mixture of air pollutants links to various health outcomes and how to tie the mixture components and health outcomes back to sources. This process involves the use of air quality models. As part of an EPA Clean Air Research Center, the Southeastern Center for Air Pollution and Epidemiology (SCAPE), a variety of air quality models are being developed and applied to provide enhanced temporal and spatial resolution of pollutant concentrations for use in epidemiologic analysis. Air quality models that are being further developed and used as part of the center include Bayesian-based ensemble methods and hybrid chemical transport-chemical mass balance modeling. The hybrid method uses knowledge of the emissions, modeling and measurement uncertainties, and can provide spatially and temporally complete pollutant fields.

An integrated model system based on the impact-pathway chain, EVA (Economic Valuation of Air pollution), has been developed, to assess the health-related economic externalities of air pollution resulting from specific emission sources or sectors. In this study, we apply the EVA system to Europe, and perform a detailed assessment of past, present, and future health-cost externalities of the total air pollution levels in Europe, represented by the years 2000, 2007, 2011, and 2020. Furthermore we perform a detailed analysis of health-related external costs from the ten major emission sectors and their relative contributions.

The paper contains a short description of the EVA system and the results from the assessment of the main contributors. The overall conclusion from the analysis of the ten major emission sectors in Europe is that the main contributors to health-related external costs are major power production, agriculture, road traffic, and non-industrial domestic combustion, including wood combustion. We conclude that when regulating the emissions of ammonia from the agricultural sector, both the impacts on nature and on human health should be taken into account. This study confirms that air pollution constitutes a serious problem to human health and that the related external costs are considerable.

Population-based epidemiological studies of air pollution have traditionally relied upon imperfect surrogates of personal exposures, such as area-wide ambient air pollution levels based on readily available concentrations from central monitoring sites. U.S. EPA in collaboration with University of Michigan is developing and evaluating several types or tiers of exposure metrics for traffic-related and regional pollutants that differ in their modeling approaches for addressing the spatial and temporal heterogeneity of pollutant concentrations. We hypothesize that using more refined exposure estimates will provide greater power to detect associations with health outcomes, particularly for traffic-related pollutants that can vary considerably over short distances and time scales. The Near-road Exposures to Urban air pollutant Study (NEXUS) design is focused on determining if children in Detroit, MI with asthma living in close proximity to major roadways have greater health impacts associated with air pollutants than those living farther away, particularly for children living near roadways with high diesel traffic. One tier for estimating exposures to traffic-generated pollutants uses local-scale dispersion modeling. Temporally and spatially-resolved pollutant concentrations, associated with local variations of emissions and meteorology, were estimated using a combination of the AERMOD and RLINE dispersion models, local emission source information from the National Emissions Inventory, detailed road network locations and traffic activity, and meteorological data from the Detroit City Airport. Hourly pollutant concentrations for CO, NO

x

, PM

2.5

and its components (EC and OC) were predicted at each study participant location (n = 160). The exposure metrics were evaluated in their ability to characterize the spatial and temporal variations of multiple ambient air pollutants across the study area. This research will be used for improving exposure assessments in future air pollution epidemiology studies, and for informing future multipollutant exposure analyses.

The growing availability of spatially resolved health data sets (i.e., resident and county level patient records) requires spatially resolved exposure or air quality metrics to investigate the impact of air pollution on health outcomes. While daily air quality data are essential in time-series epidemiologic analysis, the spatial distribution of the observations is limited. Air pollution modeling (i.e., chemical transport modeling (CTM)) addresses this by producing spatially resolved air quality predictions using terrain, emissions and meteorology inputs. However, predicted concentrations may be biased. This work incorporates unique data fusion approaches to combine air quality observations from regulatory monitoring networks (OBS) with the output from a CTM (CMAQ) to generate spatially and temporally resolved gaseous and PM species concentrations. Species concentrations alone cannot directly identify emission sources or characterize pollutant mixtures, therefore source apportionment (SA) models are required to estimate source impacts. The focus of this work is a comparison of SA results for three U.S. regions with differing air pollution sources, St. Louis, Missouri; Atlanta, Georgia; and Dallas-Fort Worth, Texas.

Recent increases in anthropogenic inputs of nitrogen to air, land and water media pose a growing threat to human health and ecosystems. Modeling of air-surface N flux is one area in need of improvement. Implementation of a linked air quality and cropland management system is described, and a tool to facilitate estimation of fertilizer input information required by the air quality model is presented. Preliminary evaluation of the coupled system against observations indicates improved wet deposition estimates.

Despite the significant NO

x

reduction in the past decade, ozone concentrations in the eastern US are in violation of the National Ambient Air Quality Standard (NAAQS). This is because the location- and time-specific effects of NO

x

emissions on ozone formation have not been taken into consideration under cap-and-trade programs where polluters trade their emission quotas on a one-to-one basis. To account for such effects, a cap-and-trade program can be reformed by inclusion of exchange rates set by the regulator on an hourly basis. We examine the performance of such a reformed cap-and-trade program using a case study of US power plants. Our results indicate that shifting emissions from high-damage hours to low-damage hours can significantly improve the performance of the system.

We use adjoint sensitivity analysis to examine the responses of a regulatory ozone attainment metric and adverse ozone health effects to location-specific reductions in anthropogenic NO

x

emissions. We find spatially heterogeneous responses among both attainment- and exposure-based metrics. Emission reductions in the western U.S. exert the most influence on attainment probability-weighted extreme concentrations (PWCs), while emission reductions in proximity to population centres in the eastern U.S. are most beneficial for reducing short-term ozone-related mortality.

Gaseous and particulate pollutants are deposited to the environment through dry, wet, and occult atmospheric processes. The US Environmental Protection Agency (USEPA), in conjunction with the National Park Service (NPS), the Bureau of Land Management (BLM), and other partners, has established the Clean Air Status and Trends Network (CASTNET) to provide estimates of the dry deposition component of total deposition of sulfur and inorganic nitrogen across the United States. CASTNET began operation in 1991 and currently features 90 active sites with many partners including multiple federal agencies, tribal, state, and local entities, and educational institutions. Most CASTNET locations are rural and intended as long term monitoring sites. Previously, CASTNET used an inferential method to estimate dry deposition by combining measured pollutant concentrations and modeled deposition velocities. Until recently, deposition velocities were modeled using the NOAA/USEPA Multi-layer Model (MLM), which incorporated meteorological measurements and information on the vegetative cover within 1 km of each site. These values were combined with wet deposition values provided by the National Atmospheric Deposition Program’s National Trends Network (NADP/NTN) to obtain total deposition. Recent changes to the methodology have improved data completeness.

Changes in urban design and traffic policy, demography, climate and associated adaptation, mitigation measures and environmental policies are likely to modify both outdoor and indoor air quality and therefore public health. The project aims to improve our understanding of future exposure situations and their impact on health, from an interdisciplinary approach. This will be achieved by using various state-of-the-art atmospheric models, measurements, epidemiological studies and reviews. To assess population full exposure, an integrated view accounting both for indoor and outdoor air pollution as well as for population time activity data will be developed. New dose-response functions will be estimated between health outcome, air pollution and temperature in order to better estimate the effects on the foetus and young children. Ultimately, scenarios of future urban climate and air quality will be simulated, combining future exposure scenarios, population scenarios and exposure-response functions to describe the effects of different trends and relevant policies on relative risk and burden of illness attributed to urban pollutants and their interactions with extreme temperatures. Also the mitigation strategies that can be used to reduce urbanization and climate change effects on the local urban meteorology and air quality will be assessed. With applications in several large European cities, the project will study the impact of several alternative adaptation scenarios on urban air quality and human health to a mid-century horizon (2030–2060) accounting for the effects of a changing urban climate. Scenario-based health impact assessments will combine exposure information from climate models, emission scenarios, policy evaluation studies and concentration calculations with exposure-response functions from epidemiological studies of vulnerable groups within the project and previously published functions for mortality and hospital admissions. The effects of socioeconomic and demographic trends will be discussed, the predicted health impacts and benefits associated with different interventions and policies and other urban changes will be described.

East Asia provides an ideal testbed to study aerosol feedbacks into climate via direct and indirect effects because of high anthropogenic emissions and unique climatology. In this work, an online coupled meteorology-chemistry model, WRF/Chem, is applied to simulate air quality and climate interactions for multiple months in 2001, 2005, and 2008 to characterize long-term seasonal variations of pollutant concentrations and quantify the contributions of anthropogenic aerosols to aerosol direct and indirect effects. The results show a reasonably good performance for most meteorological variables and chemical species concentrations. Large biases in some variables may be caused by large uncertainties in emissions. Anthropogenic aerosols in East Asia can reduce the surface net solar radiation by up to 6 % and enhance cloud condensation nuclei and cloud droplet number concentrations by a factor of up to 3 on domain-average, with much greater impacts over urban areas. These results suggest that aerosol feedbacks are potentially important over polluted areas and should be taken into account in the development of emission control and climate mitigation policies for areas where the aerosol feedback signals are strong.

A comprehensive investigation of the processes regulating tropospheric aerosol distributions, their optical properties, and their radiative effects in conjunction with verification of their simulated radiative effects for past conditions relative to measurements is needed in order to build confidence in their estimates of the projected impacts on future climate. This study aims at addressing this issue through a systematic investigation of changes in anthropogenic emissions of SO

2

and NO

x

over the past two decades in the United States, the consequent changes in anthropogenic aerosol loading in the North American troposphere, and subsequent impact on regional radiation budgets.

Increased growth in aviation activity in the future is projected to show increased emissions from this sector, and hence approximately proportional increases in concentrations if other factors were unchanging. However, emissions from other anthropogenic sources are generally expected to decrease due to several projected emissions control measures, and changes in climate will also occur. In this study, we evaluated air quality changes due to growth in aviation activities from 2005 to 2025, focusing on 99 major U.S. airports with aircraft activity data during landing and takeoff (LTO) activity developed for a growth scenario in 2025. We also assessed changes in climate based upon IPCC RCP 4.5 projections scenario, and used dynamically downscaled meteorology from the Climate Earth System Model (CESM) to WRF over the continental U.S. We performed six annual simulations at 36-km resolution using the WRF-SMOKE-CMAQ modeling system for 2005 and 2025, with and without aircraft emissions, and with and without changes in future year climate from CESM/WRF. We focused on assessing the incremental changes in O

3

, NO

2

and PM

2.5

due to changes in emissions (due to aircraft and non-aviation sources) and meteorology. We see a net increase in annual average PM

2.5

due to aviation increase from a factor of 5.5 (2025 vs. 2005) without incorporating change in climate to 5.9 with change in climate. Similarly, the changes in summer season average of daily maximum 8-h O

3

due to aviation changes from a factor of 3.1–3.3 with change in climate. Both these changes translate to about a ∼7 % additional increase in the future year that we attribute as the “climate penalty” factor. Detailed analyses of the O

3

changes show that the effect of change in climate is more pronounced at higher end of concentrations, where the grid-cells with values exceeding the U.S. NAAQS of 75 ppb see a 60 % increase due to change in climate. The changes in 1-h NO

2

due to aircraft increase by a factor of ∼2 in 2025 vs. 2005, with increases around major airports being as high as a factor of 6.

This study revolves around the use of a Non Negative Matrix Factorization method under constraints for the identification sources profiles as well as their respective contributions in three sites in northern France. Using PM

2.5

chemical analysis data, the model identified eight background and four local industrial sources profiles. In addition, the contributions of these profiles showed that secondary aerosols and combustion sources are the major constituents of the analyzed PM

2.5

, whereas industrial contributions were found majorly responsible for the elemental enrichments.

We use the WRF/Chem model to interpret observations of the aerosol concentration and its chemical composition both at surface level and along vertical profiles performed during an intensive campaign in July 2007 in Milan urban area. The model is added with a new diagnostic for aerosol budget analysis, building on that available for gas species, in order to study the contribution of upper levels processes on the aerosol formation at ground level. The analysis illustrates a quite variegated evolution of budget terms, which we found to depend strongly on the hour of the day, the vertical level, the aerosol compound, and the aerosol size. Primary components are generally emitted near the ground and rapidly transported by turbulent motions to the upper levels, where they gradually disperse and age. For some secondary components, such as nitrate, we calculate a net chemical destruction in the bottom layers, as opposed to a net chemical production higher in the boundary layer, which supply new material to ground level aerosol through turbulent mixing.

WRF/Chem has been updated in order to simulate the aerosol indirect effects using a new parameterization for production of secondary organic aerosol. The model has been evaluated over North Sea among the ATR-42 aircraft measurements of aerosol and cloud issued in frame European Integrated project on Aerosol Cloud Climate and Air Quality Interactions (EUCAARI). WRF/Chem tends to overpredict the number of condensation nuclei. Simulated liquid water content shows a bias of +15 %. Predicted cloud droplet number concentration is overestimated and radius effective droplet is underestimated.

As a pre-study for the 3-D-simulation of the hill cap cloud experiment HCCT 2010 the chemistry transport model COSMO-MUSCAT was tested in a 2-D-domain with a bell-shaped hill. The sensitivity of several target species (oxidants, pH, sulphate, organic mass, dicarboxylic acids, O/C-ratio) on the detail of the aqueous phase mechanism was investigated in a polluted urban and a clean continental rural air mass. For the treatment of aqueous phase chemistry the complex mechanism CAPRAM 3.0i in reduced version and the simple inorganic mechanism INORG was used. Differences between both mechanisms occurred for the oxidants which are not treated in INORG (HO

2

, OH, NO

3

) and the pH especially in the urban environment where much more organic reaction partners are also present in the aqueous phase. For O

3

, H

2

O

2

and the overall sulphate mass both mechanisms agree mostly within 5 %. Additionally, the composition of the organic mass has been analysed.

Atmospheric organic aerosol (OA) is highly complex and detailed mechanistic descriptions include hundreds or thousands of compounds and are impractical for use in photochemical grid models (PGMs). Therefore, PGMs adopt simplified OA modules where organic compounds with similar properties and/or origin are lumped together. The first generation volatility basis set (VBS) module grouped OA compounds by volatility and provided a unified framework for gas-aerosol partitioning of both primary and secondary OA and their chemical aging. However, a VBS approach with one dimension of variation (volatility) is unable to describe observed variations in OA oxidation state (i.e., O:C ratio) at a fixed volatility level. A two-dimensional VBS approach was introduced that tracks degree of oxidation in addition to volatility but further study is needed to fully parameterize 2-D VBS modules.

We developed a new OA module based on the VBS approach and implemented it in two widely-used PGMs. Our scheme uses four basis sets to describe oxidation state: two basis sets for oxygenated OA (anthropogenic and biogenic) and two for freshly emitted OA (from anthropogenic sources and biomass burning). Each basis set has five volatility bins including a zero-volatility bin for essentially non-volatile compounds. The scheme adjusts both carbon number and oxidation state in response to chemical aging by simplifying the 2-D VBS scheme. The new OA module is implemented in both the CAMx and CMAQ PGMs and evaluated for summer and winter 2005 episodes over the eastern US.

The online approach consists of the coupled treatment of chemical parameters, simultaneously with the meteorological parameters, in a single integrated modeling system that is referred to as chemical weather modeling. This approach offers the possibility to simulate the links and feedbacks between atmospheric processes that are traditionally neglected in air quality models. Both meteorological and chemical components are expected to benefit from this approach. Both approaches have advantages and disadvantages that make their use appropriate for different applications. This study discusses and evaluates the performance of the online Integrated Community Limited Area Modeling System (RAMS/ICLAMS) and the offline model, Comprehensive Air Quality Model with Extensions (CAMx), for a month long retrospective summertime text period, over Europe and the Greater Mediterranean Area (GMR). The implementation of the same chemical mechanisms, meteorological fields, emissions, initial and boundary conditions makes it easier to compare the results from the two models. However, there are some differences in the physical parameterizations utilized in the two models that are expected to result in differences between them. The feedback mechanisms are not activated in order to evaluate the performance of the two models regarding the advantages that the online approach offers (same projection, no interpolation in time and space, explicit calculation of the meteorological components, availability of the meteorological fields at each time step etc.). Results showed that the online approach gave better results regarding ground 1 h ozone and 24 h sulfate aerosol concentrations improving the main statistical parameters by roughly 20 % and increased correlation from 0.51 to 0.70. The differences may be mainly the outcome of the utilization of different methods for the calculation of the photolysis rates and the interpolation of the meteorological data for use in the offline model.

Air pollution reduction measures in Europe have led to marked decreases in the levels of primary pollutants such as SO

2

and NO

x

since the 1980s while for secondary pollutants like ozone and secondary inorganic aerosols (SIA) the effects of these measures are much less obvious. A recent extensive comparison of the composition of particulate matter smaller than 10 μm (PM10) in Switzerland between the years 1998/1999 and 2008/2009, for example, revealed no changes in the levels of nitrates despite large NO

x

emission reductions over this period. Similarly, aerosol sulphates decreased much less strongly than the precursor gas SO

2

.

In this study we used the online coupled regional chemistry-transport model COSMO-ART to investigate these non-linear responses of SIA components to changes in their precursor gases NO

x

, SO

2

and NH

3

focussing on wintertime conditions when PM concentrations are highest. We found that the reduced concentrations of NO

x

and NMVOC led to strong increases in the oxidation capacity of the European wintertime boundary layer as evidenced by pronounced enhancements in ozone, OH and N

2

O

5

levels by sometimes more than 100 %. These changes resulted in accelerated formation of nitrates and sulphates and correspondingly smaller relative reductions in their concentrations as compared to the precursor gases, particularly within the densely populated and industrialized regions in Europe.

This work aims to quantify the contribution of biogenic emissions to air quality levels of secondary pollutants in distinct areas of a study region taking into account the relevance of different processes, such as transport, deposition, and chemistry. In this sense the air quality model CAMx, with its source apportionment and process analysis tools, has been applied to Portugal, aiming to assess ozone (O

3

) and secondary organic aerosol (SOA) levels during a summer month for three biogenic emission scenarios, which were defined considering the variability of emission factors. Results for secondary gases and particulate pollutants have been explored in terms of source and process contributions. The quantitative analysis showed that biogenic emissions contribute up to 90 % of SOA predictions. The uncertainty in isoprene emission factors may affect the concentration of SOA by ±30 % in average; however, because the O

3

production is NOx limited at the studied areas biogenic emissions scenarios did not affect O

3

concentration levels, which varied only ±1 % in average.

A fully coupled on-line air quality forecast model, GEM-MACH, was used to study a case of cloud processing in an urban-industrial plume and aerosol-cloud-meteorology interaction. Preliminary results have shown a significant impact on modelled clouds and particulate sulfate concentrations due to the inclusion of the feedback from on-line aerosols to microphysics. Further tests and detailed comparison with in-situ measurements of this case are underway.

Cloud chemistry models include more and more explicit multiphase mechanisms based on laboratory experiments that provide kinetic constants, stability constants of complexes, hydration constant,

etc.

However, they are still subject to many uncertainties, related to the aqueous chemical mechanism they used which most of the time has never been confronted against laboratory experimental data. To fill this gap, we propose to adapt the M2C2 model to simulate irradiation experiments on synthetic aqueous solutions under controlled conditions (pH, pressure, temperature, light intensity,…) considering various chemical compounds that are supposed to contribute to the oxidative budget in cloud water (iron, oxidants such as H

2

O

2

). As target species, organic compounds (oxalic, formic, acetic acids) are taken into account since they are oxidized and are also good indicators of the oxidative capacity and potential iron complexant. Range of concentrations for all the studied chemical compounds are representative of

in situ

measurements. Numerical outputs are confronted to experimental data that consist in time evolution of the concentration of target species (Long Y, Charbouillot T, Brigante M, Mailhot G, Delort A-M, Chaumerliac N, Deguillaume L, Evaluation of modeled cloud chemistry mechanism against laboratory irradiation experiments: the HxOy/iron/carboxylic acid chemical system. Atmos Environ 77:686–695; 2013).

An experimental campaign has been conducted in Nov. 2007 in Duck, NC. Several sets of probes were used in various conditions, in order to measure either the marine aerosol vertical gradient and the flux emitted by the surf zone or the horizontal transport offshore. It was possible to find a new equation for the surf zone aerosol flux depending on the Wave Energy Dissipation (WED). The aerosol transport was simulated for offshore wind condition and compared to experimental data measured aboard a boat travelling along the offshore wind vector.

Current approaches to simulate secondary organic aerosols (SOA) in regional and global numerical models are based on parameterizations of the oxidation of precursor gases in the gas-phase and subsequent partitioning into particles. Recent findings suggest however that formation in the aqueous-phase of aerosols might contribute substantially to ambient SOA load. In this work we investigate the contribution of glyoxal to SOA through chemical processes associated with aerosols. Both a very simple and a more explicit mechanism of SOA formation from glyoxal was included in the regional chemistry transport model WRF/Chem. We simulated the first 2 weeks of June 2010 over the domain of California to make use of the extensive dataset collected during the CARES/CalNex field campaigns to evaluate our simulations. Contributions to total SOA mass were found to range from 1 to 15 % in the LA basin, and <1 to 9 % in the isoprene-rich eastern slopes of the Central Valley. We find that the simple approach previously used in box as well as global modeling studies gives the highest contributions. A combination of reversible partitioning and volume pathways can provide comparable amounts only if partitioning of glyoxal into the aerosol liquid-phase is instantaneous.

In the atmosphere, a significant fraction of organic aerosol (OA) makes up the particulate matter. Since OA consists of thousands of complex molecules, uncertainties remain concerning their sources and atmospheric processes leading to their transformation (Hallquist M et al. Atmos Chem Phys 9:5155–5236, 2009). OA are, in principle, divided into two categories: primary organic aerosol (POA), which is emitted directly into the atmosphere, and secondary organic aerosol (SOA), which is formed by the photo-chemical oxidation of volatile organic compounds (VOC). VOC emissions and consequently OA formation are strongly linked to the season of the year. To that purpose, simulations were performed in autumn (September 2008) and summer (June 2010) with the WRF-CHEM model including more precisely RACM (Stockwell WR et al. J Geophys Res 102:25847–25879, 1997) chemical mechanism for gas phase associated with MADE for aerosols (Ackermann IJ et al. Atmos Environ 32(17):2981–2999, 1998) and VBS (Ahmadov et al. J Geophys Res 117:D06301, 2012) for SOA. Moreover, model outputs are compared with measurements performed during each season (Freney EJ et al. Atmos Chem Phys 11:13047–13059, 2011) at the puy de Dôme monitoring site which is located in central France (45°46 N 2°57 E) and is a part of the ACTRIS (Aerosols, Clouds, and Trace gases Research InfraStructure) and GAW (Global Atmospheric Watch) network.

The main aim is to estimate the BaP concentrations in the atmosphere and know the main influences on BaP ground level concentrations. We have utilized CMAQ 4.7 version as core for our system. BaP degradation and adsorption mechanism have been integrated into the chemical-transport model. Emissions have been processed the EMIMO-UPM emission model with a top-down approach including detailed GIS data. The air quality model has been driven by the WRF meteorological model. Urban particulate material generally contains a significant amount of amorphous organic carbon. Absorptive mechanism plays the dominant role in the urban air and also in air affected by urban sources. Equilibrium partitioning calculations are based on absorptive partitioning model. Gas/particle partitioning of BaP is simulated as an absorption process into the organic mass on the aerosol particle assuming the formation of a quasi-ideal solution. BaP reaction with ozone is an important degradation pathway of the particulate BaP in the atmosphere. The degradation rate constant depends on the ozone gas to surface equilibrium constant and the maximum rate coefficient that would be observed at high ozone concentrations. A validation process of the BaP and other pollutant results has been conducted in the urban area of Zaragoza (Spain) during 10 weeks in 2010. The agreement is generally satisfactory. The simulated concentrations depend on the meteorological and emission inputs.

The POAEMM project consists of developing a tool for predicting the spatial and temporal variation of aerosols particles in a marine coastal environment. This is achieved by coupling the MEDEX parametric model (MIO) and the fine mesh meteorological model AROME developed by Meteo France. The first step has been to do the semi-automation of the MIO experimental station. Based on semi-empirical formulations, MEDEX is then revised in order to take better account of oceanographic and meteorological conditions.

Saudi Aramco is the oil industry of the Kingdom of Saudi Arabia with several activities related to the environment. In order to optimize daily operations and minimize environmental risks a forecasting system has been employed and setup in operations. The objectives of the system include prevention and mitigation of environmental problems, as well as early warning of local conditions associated with extreme weather events. The management and operations part is related to early warning of weather and dust storms that affect operations of various facilities, whereas the environmental part is mainly focused on air quality and desert dust levels in the atmosphere.

The FRAME (Fine Resolution Atmospheric Multi-pollutant Exchange) model is a Lagrangian atmospheric transport model with a horizontal grid resolution of 5 × 5 km

2

. This work uses FRAME to calculate annual average concentrations of three secondary inorganic aerosol species (NH

4

+

, NO

3

−

, SO

4

2−

) and gas phase SO

2

over the United Kingdom for the years 2000–2010. Modelled concentrations are compared to measurements from 12 monitoring sites which are operated with the UK Eutrophying and Acidifying Air Pollutants (UKEAP) Acid Gases and Aerosols Network (AGANet). The results showed good spatial correlation between measurements of gas and aerosol concentrations and modelled values (R

2

> 0.8). However FRAME underestimated decreases in sulphate and ammonium aerosol concentrations over the 11-year period and overestimated changes in SO

2

concentrations relative to measurements. Ongoing work to investigate temporal changes in atmospheric oxidation rates should clarify this difference. FRAME showed a more modest reduction in nitrate concentrations (compared to ammonium and sulphate), which is in agreement with measurements.

FRAME was used to estimate air concentrations for two future scenarios based on emissions predictions for the years 2020 and 2030 in which significant reductions in UK emissions in the UK are forecast (40 % for SO

2

and 47 % for NO

x

between 2010 and 2030). The relationships between long term changes in precursor gas emissions and the formation of secondary inorganic aerosol are discussed in terms of the modelling results.

Air pollution is associated with adverse effects on human health through exposure to PM and ozone. To design effective air pollution mitigation strategies it is essential to quantify the effect of pollutant emission reductions on their concentrations as well as health and ecosystem impacts. Within integrated assessment modeling source-receptor relationships (SRRs) are used that are based on chemistry transport modelling. Currently, these SRRs are made using invariant emission time profiles. The LOTOS-EUROS model equipped with a source attribution module was used to test this assumptions for renewable energy scenarios. Renewable energy availability and thereby fossil fuel back up are strongly dependent on meteorological conditions. We have used the spatially and temporally explicit energy model REMix to derive time profiles for backup power generation. These time profiles were used in LOTOS-EUROS to investigate the effect of emission timing on air pollutant concentrations and SRRs. It is found that the effectiveness of emission reduction in the power sector is significantly lower when accounting for the shift in the way emissions are divided over the year and the correlation of emissions with synoptic situations. The source receptor relationships also changed significantly. This effect was found for both primary and secondary pollutants. Our results indicate that emission timing deserves explicit attention when assessing the impacts of system changes on air quality.

In the atmosphere mercury exists in three forms: Gaseous Elemental Mercury (GEM), Gaseous Oxidized Mercury (GOM), and Particle Bound Mercury (PBM). GOM and PBM make up only 1 % of the total. But deposition, which is the only sink for atmospheric mercury, is dominated by these two species. Therefore, oxidation processes are key to understand the behaviour of mercury in the atmosphere. However, in the scientific community a consensus on the importance of oxidizing reactants, namely ozone, hydroxy radicals, and halogens, has not been reached.

This model study about the influence of chemical reactants on the regional transport of mercury is part of the European Union FP7 Research Project GMOS (Global Mercury Observation System). GMOS focuses on the improvement and validation of mercury models to assist establishing a global monitoring network and to support political decisions.

In the course of this study the Community Multiscale Air Quality (CMAQ) model was used to simulate the transport and deposition of mercury. CMAQ was run on a regional domain over Europe using different chemical mechanisms for the oxidation of mercury. The model results were compared to newly available long term observations of speciated mercury. Based on this unique dataset, the relevance of different chemical reactions for the oxidation and reduction of mercury were investigated. The main finding was that the emissions of GOM as well as the production of GOM by oxidation processes is vastly overestimated. Moreover, there are indicators that the fraction of PBM, produced by reaction of GEM with ozone and OH is overestimated.

Aviation is a unique anthropogenic source with 4-dimensional varying emissions, emitting 90 % of their emissions in upper troposphere at cruise altitudes (9–12 km). Aircraft emissions budgets in upper troposphere lower stratosphere (UTLS) region and their potential impacts on upper troposphere and surface air quality are not well understood. The key objective of this study is to characterize the aircraft emissions during full-flight activity in regional and hemispheric modeling scales, and assess their impacts on upper tropospheric chemistry and surface air quality. Using detailed spatio-temporal characterization of aircraft emissions along with other background emissions in the modeling domains, we studied incremental impacts of aircraft emissions focusing mainly on O

3

, NO

x

and PM

2.5

species. Comparison of modeling results with aircraft measurements showed improvement of model performance due to enhanced modeling platform and consideration of cruise altitude aviation emissions in the upper troposphere.

For the simulation of photochemically created pollutants like ozone it is essential to correctly consider reaction rates induced by short-wave radiation. In atmospheric chemistry transport models this is achieved by the use of either off- or online calculated photolysis frequencies. In this study the effect of different input parameters of a radiation model on the calculated photolysis frequencies have been investigated. In the second step an atmospheric chemistry transport model was used to assess the impact of changed photolysis frequencies on the simulation of ozone concentrations. The impact of changed radiation model input parameters on the calculated photolysis frequencies vary not only with regard to the changed parameter but also with regard to the to the species to be dissociated. Furthermore the impact of different sets of photolysis rates employed in a chemical transport simulation on the modelled concentrations is differed and likely to be less important than other aspects of the simulation like the resolution of the grid and the emissions used. Apart from major surface albedo changes (grass to snow) and extreme changes in total ozone column content for J

O3

clouds are the dominating factor in modifying the photolysis frequencies especially as they feature a highly temporal and special variation. The results show that simulated maximum ozone concentrations in areas with clouds are reduced.

In January 2013 exceptionally high levels of particulate matter (PM) concentrations were reported for the area around Beijing (40 N, 116 E) with maximum concentrations exceeding 500 μg/m

3

. Observations of the aerosol optical depth (AOD) within the Aeronet sun-photometer network showed an AOD of more than 2 on several days. In order to analyze this high pollution episode PM concentrations in China were simulated with the Community Multiscale Air Quality (CMAQ) model for the period from 10 December 2012 to 31 January 2013. Emissions were taken from the EDGAR data base. The most recent emission rates from 2008 were further increased considering the growth of China’s gross domestic product. The results were compared to ground based PM2.5 measurements taken at the US embassy in Beijing and to Aeronet sun-photometer observations. The model was generally able to reproduce the high PM levels measured in situ close to ground, however the largest peak on 12 January was not captured, because of an exceptionally strong temperature inversion close to ground that was not reproduced in the meteorological fields. The meteorological model significantly underestimated the relative humidity in the lowest layer, leading to some underestimations of the aerosol optical depth. An analysis of the most important source sectors showed that residential heating was the main emission source sector on those days with the highest aerosol concentrations.

This paper shows the effects of emission reductions from 2005 to 2020 levels on ozone and PM2.5 concentrations in Ontario based on modelling results. To assess the effect of these changes on Ontario’s air quality, the modelling system WRF – SMOKE – CMAQ was run with 2010 meteorology for the entire year. Emission estimates were based on US EPA Cross-State Air Pollution Rule Final CAP – BAFM 2005-Based Platform, version 4.2, Toxic rule 2016 with updates projected to 2020 for mobile emissions and Environment Canada 2005 and 2020 emission inventories.

Emission changes in Ontario and in the neighbouring US states (Ohio, Michigan, Indiana, etc.) from 2005 levels to projected values of 2020 are discussed. Modelling results were summarized to show spatial and temporal variability in ozone and PM2.5 concentration over Ontario due to emission reduction in the whole domain. Analysis was performed on annual values and on values obtained on days with high concentrations only.

A relatively simple Lagrangian atmospheric transport model (FRAME) was adapted to simulate the concentration and deposition of nine heavy metals (As, Cd, Cr, Cu, Pb, Ni, Se, V and Zn) in the UK. The modelled data was compared with annually averaged measured wet deposition of metals and concentrations in air. The model obtained good correlation with measurements of metal concentrations in air but with very large underestimates (normalised mean biases in the range −0.64 to −0.93), indicating a major under-estimate in total atmospheric emissions. Wet deposition was less closely correlated to measurements. Inclusion of estimates of spatial re-suspension of wind-driven dust for the UK in the model simulation led to an improvement in agreement with measured concentrations. However the amount of re-suspended material was considered to be highly uncertain due to the limited availability of measurement data of the heavy metal content of surface soil and dust.

Environment Canada’s “Global Environmental Multiscale – Modelling Air-quality and Chemistry” (GEM-MACH) is the Canadian operational air-quality model, used to provide forecasts of ozone, PM

2.5

and air-quality health metrics to the Canadian public. The operational GEM-MACH is an on-line model, but is not fully coupled, in that the chemical variables are not used to modify the weather. The model was converted to fully coupled status as part of Environment Canada’s participation in phase 2 of the Air-Quality Model Evaluation International Initiative, with three classes of modifications: (1) Additions required in order to allow feedbacks to take place between weather and chemistry; (2) Model improvements necessary to ensure feedback accuracy; (3) Model improvements to allow the use of AQMEII-2 prescribed inputs and diagnostic outputs.

The revised model is being used to generate four annual simulations of air-quality over North America, for “feedback” and “base-case” simulations for the years 2006 and 2010. Here, the initial test simulation results for 2006 of surface O

3

and PM

2.5

are compared using a simple statistical package, as a means of identifying cases wherein feedbacks have the greatest influence on the model’s chemical output. These instances will be put forward for further study under the multi-model framework of AQMEII-2.

Two new postprocessing methods based on analogs are proposed to reduce the systematic and random errors of air quality prediction. The analog of a forecast for a given location and time is defined as a past prediction that matches selected features of the current forecast. The first method is the weighted average of the observations that verified when the best analogs were valid (AN). The second method consists in applying a postprocessing algorithm inspired by the Kalman filter (KF) to AN (KFAN). The AN and KFAN are tested for ground level ozone and PM2.5 0–48 h predictions from the Community Multiscale Air Quality (CMAQ) model, with observations from 1602 surface stations from the EPA AirNow network over the continental United States for a 1-year period. Preliminary results of the new methods include a large reduction of the systematic and random errors of the direct model output, with an increase of the correlation between observations and predictions at all forecast lead times.

We studied and compared different operational modeling techniques that are used to generate regional scale concentration maps for PM

10

, PM

2,5

, NO

2

and O

3

over Belgium. The various techniques and resulting maps were analyzed, validated and compared aiming at identifying the best possible regional scale concentration map for each pollutant. A distinction was made between a temporal and a spatial validation. The temporal analysis revealed that an intelligent interpolation technique based on land use characteristics in general performs best in capturing the temporal aspects of air quality in Belgium for the investigated pollutants. For PM

10

and PM

2.5

this technique also performs best in generating the spatial pattern of the observed annually averaged concentrations. A deterministic model combined with a corrective ‘Optimal Interpolation’ data assimilation technique performs best in reproducing the spatial pattern of O

3

. For NO

2

the interpolation technique manages best in explaining the spatial pattern of the observed annually averaged concentrations in Belgium, but when restricted to the region of Flanders, it competes with a thoroughly calibrated Lagrangian type of modeling.

We incorporate the Regional Atmospheric Chemistry Mechanism (RACM2) into the Community Multiscale Air Quality (CMAQ) hemispheric model and compare model predictions to those obtained using the existing Carbon Bond chemical mechanism with the updated toluene chemistry (CB05TU). The RACM2 enhances monthly mean ozone by 2–10 ppbv in polluted areas compared to the CB05TU while reducing mean ozone by 2–6 ppbv in remote areas. We develop an effective halogen reaction that can consume ozone over the gulfs and oceans. The current CMAQ model uses substantially lower ozone deposition velocity over water compared to observed data. We modify the CMAQ deposition velocity to account for the enhanced deposition due to chemical interactions between ozone and oceanic iodide. The effective halogen reaction and enhanced deposition velocity reduce monthly mean ozone by 2–8 ppbv over water. The majority of the reduction occurs via the halogen reaction. A comparison of model predictions with available observed profile reveals that the RACM2 over-predicts surface ozone in polluted areas while improving the comparison in remote areas. Model predictions with the halogen chemistry and enhanced deposition velocity compare better with the observed data.

The current study intends to evaluate the fire emission estimates obtained from IS4FIRES v1.5. The system provides spatially and temporally resolved emission fluxes originated from wild-land fires. The emissions were obtained by utilising remote-sensing products of MODIS and SEVIRI instruments: TA and FRP. The primary scaling is based on emission factors for PM2.5 determined for seven land-use types: grass, crop residue, shrub, tropical, temperate and boreal forest, and peat. The PM2.5 emission fluxes can be converted to total PM and gaseous species using literature-reported scaling factors.

To evaluate the system, the fire emission fluxes were used as input to the SILAM model, which evaluated the dispersion and transformation of the released smoke. The observational datasets included AOD observations from MODIS. To facilitate the comparison and estimate the contribution from fires to AOD, SILAM inorganic chemistry calculated formation of secondary inorganic aerosol. Primary PM emissions from anthropogenic and natural sources were also included.

The model-measurement comparison showed that spatial and temporal distributions of the fire smoke are well reproduced. Nevertheless, the smoke from fires occurring in central Africa and South America are overestimated, and fires occurring in areas where peat and crop are dominant are underestimated. The optimization of the system, in general, results on a reduction of the emission coefficients, with exception of peat and crop, as expected; it reduces emission substantially especially for the areas where tropical and grass are dominating and fires tend to be very intense (Africa). Nevertheless, in some cases reduction seems to be counterproductive, emissions are heavily reduced.

While emissions in Europe tend to show negative trends and are relatively well known, for some other regions in the world largely positive emission trends are visible and accurate information on these emissions is lacking. Over these regions satellite observations can play a large role, especially when there are few ground based observations available. Satellite observations have been successfully combined with regional chemistry transport models to provide best possible air quality analyses and forecasts and to derive important information on emission sources.

LOTOS-EUROS is a regional chemistry transport model with an active data assimilation system aimed at modeling the European air quality. To be able to answer the air quality and emission information needs over regions outside Europe a new LOTOS-EUROS version has been developed which can in principle be applied over any region of the world. First calculations have been performed over China. Resulting NO

2

and PM fields have been compared to satellite observations from the OMI instrument and ground based observations in Beijing and Shanghai. The results look promising as the model is able to reproduce the spatial and temporal variability within the NO

2

and PM concentrations. Absolute differences between model and observations can be attributed to missing sources in the model, scaling issues and observational errors.

In this paper we will present the steps that have been taken to develop the new LOTOS-EUROS version, show results over China and discuss encountered issues and further development plans.

EPA’s 2008 National Emission Inventory has been incorporated into version 2 of the Hemispheric Transport of Air Pollutants Inventory. This work involves the creation of a detailed mapping of EPA Source Classification Codes (SCC) to the International Nomenclature for Reporting System (NFR). The mapping of SCC codes to the NFR system allows for comparison of USA emission inventories with other national inventories on a consistent basis. We summarize the emission estimates for 2008 and 2010 and provide a useful reference for linking USA inventories to Global inventories for use in regional and global chemical transport models.

This paper presents the main results of two specific studies related to the sensitivity of spatial resolution on chemistry transport modeling. The horizontal scale is addressed by simulating the air quality in Europe for the full year 2009 at 50 and 8 km resolutions on a very high performance computing infrastructure. An additional 2 km resolution was performed on a smaller period to assess the added value of a very high horizontal resolution during the air pollution outbreak of January 2009 when particulate and nitrogen dioxide levels reached alert levels. We reveal small scale air pollution patterns that highlight the contribution of city plumes to background air pollution levels. The influence of the vertical resolution on air pollutant concentrations near the surface is studied. The results are discussed in terms of differences on surface concentrations between the reference case and an improved resolution. Adding a point close to the surface appears to be important mainly for high nocturnal concentrations in very stable boundary layers.

We simulated the ozone and particulate matter (PM) concentrations in Europe with the regional air quality model CAMx for 1990 (reference year for the Gothenburg Protocol), 2005 (reference year for the revised Gothenburg protocol), 2006 (for model validation) and 2020 (target year for the revised Gothenburg protocol) to investigate the effects of emission reductions on air quality. The three emission scenarios used for 2020 were Baseline, Mid and MTFR, prepared by IIASA using the GAINS model. Comparisons between 1990 and 2005 suggested a large decrease (up to 50 %) in PM2.5 concentrations as shown also by observations. PM2.5 levels were predicted to decrease further by about 35 % (Baseline) and 45 % (MTFR) until 2020. Although ozone damage indicators AOT40 and SOMO35 were modeled reasonably well for the reference case 2005, the relative change between 1990 and 2005 did not match some of the observations. The significant decrease in indicators predicted for 2020 should be further investigated, because background ozone concentrations might have a strong influence on correct prediction of these indicators with thresholds. Our results suggested that the deposition of nitrogen compounds decreased by 20–50 % between 1990 and 2005 and the main reduction was due to the oxidized nitrogen species. On the other hand, the modeled deposition of the reduced nitrogen species in 2005 was higher than in 1990. The deposition of the oxidized nitrogen species was predicted to decrease further by about 40 % until 2020 while deposition of reduced nitrogen species would be higher in the future.

In Canada, air toxic pollutants are regulated under Schedule 1 of the Canadian Environmental Protection Act-1999. As part of the Canadian government’s Clean Air Regulatory Agenda (CARA), Environment Canada is working with Health Canada to develop air quality modelling capabilities to support assessments of health and environmental impacts from mobile source emissions. In current phase, there are 6 compounds of interest: benzene, 1-3 butadiene, formaldehyde, acetaldehyde, acrolein and 1-2-4 trimethylbenzene. Environment Canada’s AURAMS air quality model was modified to explicitly simulate the chemistry of these compounds with the SAPRC07-toxics mechanism. National emissions inventories were processed for air toxics species with source-specific VOC profiles, while on-road mobile sources were explicitly modeled. AURAMS was applied on a 45-km grid spaced domain covering Canada and the US, and over a summer and a winter period. Long-term analysis of >5 years of surface measurements from National Air Pollutant Surveillance (NAPS) Network for 10–54 stations were used to compare with this first model analysis. This paper will focus on the analysis of long-term ambient concentrations of air toxics in Canada, as well as initial results from the AURAMS model with air toxics capability for a summer and a winter period.

The choice of the parameterization scheme, the input parameters and the spatial resolution are options that can provide significantly different modelling results for the processes of desert dust production and transport. This work will discuss the level of coherence between a regional and a global modelling system with regards to the simulation of desert dust production and transport in N-Africa. The limited-area model SKIRON/Dust and the atmospheric chemistry general circulation model EMAC (ECHAM5/MESSy2 Atmospheric Chemistry) have been applied using a common physically-based dust emission scheme. The differences between the two modelling systems and the comparison with the observations will be analysed and discussed, as well as the plans for future work in their offline coupling.

By means of model simulations with the chemistry transport model CMAQ the influence of ship emissions in the North Sea on concentrations and depositions of sulfur and nitrogen oxides over Europe was investigated. Ship emissions for the North Sea of the base year 2008 were provided by the Dutch research institute MARIN. Based on this, emission scenarios were developed that comply to ECA regulations at different levels. Finally, the emissions were fed into the CMAQ model that simulates the fate of pollutants in the atmosphere in order to estimate concentrations and depositions of the pollutants of interest for each scenario. A comparison of the simulation results yielded a quantification of the changes of air pollution levels over the North Sea riparian states and, hence, provided information for estimating the benefit of installing ECAs in the North Sea. Concentration differences can reach up to 50 % close to shipping lines and still 25 % ashore. The ECA regulations to lower nitrogen and sulfur exhaust take effect at different time scales and are counteracted by the expected increase of fuel use due to increased ship traffic.

Health data geo-coded with residential coordinates are being used to investigate the relationship between ambient air quality and pediatric emergency department visits in the State of Georgia over the time period 2000–2010. Two types of ambient air quality data – observed concentrations from ambient monitors and predicted concentrations from a chemical transport model (CMAQ) – are being fused to provide spatially resolved daily metrics of five air pollutant gases (CO, NO

2

, NO

x

, SO

2

and O

3

) and seven airborne particulate matter measures (PM

10

, PM

2.5

, and PM

2.5

constituents SO

4

2−

, NO

3

−

, NH

4

+

, EC, OC). The observational data provide reliable temporal trends at and near monitors, but limited spatial information due to the sparse monitoring network; CMAQ data, on the other hand, provide rich spatial information but less reliable temporal information. Four data fusion techniques were applied to provide daily spatial fields of ambient air pollutant concentrations, with data withholding used to evaluate model performance. Two of the data fusion methods were combined to provide results that minimized bias and maximized correlation over time and space with withheld data. Results vary widely across pollutants. These results provide health researchers with complete temporal and spatial air pollutant fields, as well as with temporal and spatial error estimate fields that can be incorporated into health risk models. Future work will apply these methods to five cities for use in ongoing air pollution health studies and to examine strategies for incorporating land use regression variables for spatial downscaling of data fusion results.

The main aim of this work is the investigation of the contribution of different emission sources to the concentrations of the fine Particulate Matter (PM2.5) in the city-port of Patras in Greece using a meteorological and a photochemical model. Emphasis is given on the contribution of the maritime transport and the activities within the Patras harbor area for which the existing environmental information is very limited.

Over 98 % of ammonia (NH

3

) emission in Poland comes from agricultural activity. Compared to oxidised sulphur and nitrogen, which show significant downward trends after the year 1990, national total NH

3

emission is fairly stable at around 320 Gg. In this work, the Fine Resolution Atmospheric Multi-pollutant Exchange (FRAME) model is used to assess the long-term information on annual NH

3

and NH

4

+

air concentrations.

The main findings of our work show that the FRAME model is capable of reproducing air concentrations of NH

3

and NH

4

+

well for all selected years. There is no tendency for under or overestimation of the observed air concentrations. The FRAME model results are also in general agreement with the EMEP-Unified model results, though the differences in spatial resolutions of both models are important (FRAME 5 km grid and EMEP-Unified 50 km grid). There no significant trend in NH

3

annual air concentration in Poland if country average values are compared.

This study examines the contribution of pollutants (NOx and HNO3) from Highveld on pollution in Cape Town. For the study, we analyzed observation data (2001–2008) from the City of Cape Town air quality network and regional climate model (RegCM4) simulations (2001–2004) over southern Africa. The model accounts for the influence of complex topography, atmospheric condition, and chemical reactions among the atmospheric gases in simulating the emission and transport of the pollutants over southern Africa. The model results show that north-easterly flows can transport pollutants (especially HNO3) at low level (surface – 850 hPa) from Highveld to Cape Town. And in April, a tongue of high concentration of NOx and HNO3 extends from Highveld to Cape Town along southern coast. The study shows two dominant paths through which pollutants from Mpumalanga Highveld are transported to Cape Town. The first path follows the borders of South Africa, Botswana and Namibia, and along Western Cape to Cape Town while the second path is moves southward over South Africa to the south coast and then eastward along south coast to Cape Town.

Precursors of particulate matter, as well as particulate matter (PM) itself can be transported in the atmosphere over long distances, thus the problem of air pollution should be considered in the cross-border context. We used the Fine Resolution Atmospheric Multi-pollutant Exchange model to calculate the import and export of sulphur and nitrogen compounds and primary particulate matter (PPM

2.5

and PPM

10

) for the year 2007, separately for a central European country (Poland) and an island nation (the UK). For PPM the analysis was done for two cases – including and excluding sea salt aerosol (SSA). The absolute mass of import is higher for Poland than the UK in the case of pollutants for which SSA is not the main contributor (e. g. NO

3

−

, NH

4

+

). Excluding SSA the import of PPM

2.5

and PPM

10

is higher for Poland than the UK. Including SSA, for both countries the mass of imported PPM

10

and PPM

2.5

is higher than the national total emission. For most of the considered pollutants the highest contribution of import was calculated for the north-west and west part of the UK and the west and south-west part of Poland.

This paper traces the evolution of my “simple urban dispersion model”, beginning with the 1971 Atmospheric Turbulence and Diffusion Laboratory (ATDL) urban dispersion model, which was developed for application to urban area sources of radon. The Gaussian plume model is the basis, with use of Gifford’s novel derivation involving integration over upwind sources. The development continued with the addition of an urban meteorological preprocessor in the Hybrid Plume Dispersion Model (HPDM), accounting for the effects of strong mechanical mixing and increases in turbulence intensity in urban areas. Since 2000, the simple urban model was modified for US Department of Defense applications regarding toxic gas releases in built-up city centers, and tested with extensive field data. One model option is the Gaussian formula and another is an even simpler dimensionless relation.

Interactions between different spatial and temporal scales play a major role in determining the flow structure over the urban canopy in densely built agglomerations. The intense surface in-homogeneities result in the generation of additional terms in the turbulent transport within the urban atmospheric boundary layer (ABL), which in effect generate equally intense temporal in-homogeneities. Aiming to address the limitations which arise as a result of physical disparities between the scales involved, a novel two-way scheme has been developed within the frame of the EU MEGAPOLI project to couple the mesoscale model MEMO and the microscale model MIMO, utilising a collection of interpolating metamodels. The methodology developed was employed for assessing the effects of multiscale interactions on turbulent transport in the city of Paris using the MARS-aero model, and was validated based on comparisons with monitoring measurements. Simulation results reveal an improvement of the prediction skill over non-coupled models in accurately resolving scale interactions on the atmospheric flow within the urban ABL.

The identification and quantification of population exposure of children and elderly people to PAHs in urban areas are the major goals of the EXPAH LIFE+ Project (

www.ispesl.it/expah

). To reach these objectives an integrated approach, based on measurements and modeling techniques, has been set up to preliminarily reconstruct PAHs levels in the Rome metropolitan area. Field campaigns of particulate PAHs and PM

2.5

have been performed in different sites and microenvironments from December 2011 to July 2012. These data were essential to evaluate and integrate results of the Flexible Air quality Regional Model (FARM) that has been run from June 2011 to May 2012. PAHs modeled concentrations are presented for the city of Rome as well as a comparison with observations. Capabilities and limits in modeling PAHs in urban areas are then discussed.

The United Nations Environment Programme stated that the human exposure to air pollutants primarily emitted by road traffic is associated to nearly 0.8 million premature deaths annually, especially in sensitive groups such as children. This paper evaluates the individual exposure of students in a Portuguese town accounting for their walk to school and the time spent in the classroom. The analysis was carried out for 8 hypothetical walking routes that were tracked and profiled with a GPS. Ambient air pollutant concentrations of CO and PM10 were simulated with a Computational Fluid Dynamics (CFD) model, while indoor air pollutant levels were estimated for different classrooms using the simulated outdoor concentrations and a mass transfer approach. Results show that the individual exposure of pedestrians in a city is extremely spatially dependent, as a consequence of air pollutant dispersion patterns, leading to significant disparities between the children’s exposure. A tendency between the orientation of classroom’s façade and resulting exposure was not found, but in the case of the outdoor exposure the children coming from N and NE have the highest values. The variability of the estimated exposure values shows the magnitude of the error that can be committed when using a single value of air quality as a surrogate of air pollution exposure. This problem can be overcome by using building-resolving CFD models that provide an accurate and detailed understanding of how human behaviour and habits can affect the exposure of urban citizens and ultimately their health.

The effect of varying complexity of idealized urban canopy on street traffic generated air pollution is simulated using large eddy simulation. The geometry is modeled as an approximation to typical city centers in Central Europe and is also used for wind tunnel measurements. Four configurations have with increasing complexity have been considered. In this paper we examine average scalar concentrations and the scalar fluxes decomposed to the advective and turbulent parts.

In this work we simulated the Macdona (USA) chlorine railway accident with the Lagrangian dispersion model MicroSpray, twice: firstly by using the standard version and then by using a recently developed new two-phase module. MicroSpray was coupled with the diagnostic MicroSwift model, which provides the 3-D wind field in presence of obstacles and orography, and was used to test the two options. The results obtained in the two simulations, with and without the new module, are presented and the differences are discussed.

In 2050 the global fraction of urban population will increase to over 69 %, which means that around 6.3 billion people are expected to live in urban areas (UN, Urban population, development and environment.

http://www.un.org/esa/population/publications/2011UrbanPopDevEnv_Chart/urban_wallchart_2011-web-smaller.pdf

, 9 May 2013, 2011). Cities are the predominant places for human beings to settle down, thus becoming more vulnerable to extreme weather events aggravating phenomena like heat stress and decreasing air quality aroused by inner city pollution. Affecting human health and having the tendency to occur more likely in the future (Solomon et al. Contribution of working group I to the fourth assessment report of the intergovernmental panel on climate change, 2007. Cambridge University Press, Cambridge/New York, 2007), finding mitigation strategies to counteract future air quality related problems and ways to sustain development is of great importance. In this study, the mesoscale numerical model WRF/WRF-Chem is used on regional scale to simulate the effect of urban planning scenarios on dynamical processes impressing urban climate and air quality.

A new high-resolution dynamical downscaling strategy to examine how rural and urban areas respond to change in future climate, is presented. The regional climate simulations have been performed with a new version of the limited-area model of the ARPEGE-IFS system running at 4-km resolution coupled with the Town Energy Balance scheme (TEB). In order to downscale further the regional climate projections to a urban scale, at 1 km resolution, a stand-alone surface scheme is employed in offline mode. We performed downscaling simulations according to three model set-ups: (i) reference run, where TEB is not activated neither in 4-km simulations nor in 1-km urban simulation, (ii) offline run, where TEB is activated only for 1-km urban simulation and (iii) inline run, where TEB is activated both for regional and urban simulations. The applicability of this method is demonstrated for Brussels Capital Region, Belgium. For present climate conditions, another set of simulations were performed using European Center for Medium-Range Weather Forecasts global reanalysis ERA40 data. Results from our simulations indicate that the reference and offline runs have comparable values of daytime and nocturnal urban heat island (UHI) and lower values than the inline run. The inline values are closer to observations. In the future climate, under and A1B emission scenario, the three downscaling methods project a decrease of daytime UHI between −0.24 and −0.20 °C, however, their responses are different for nocturnal UHI: (i) reference run values remains unaltered, (ii) for the offline runs, the frequency of present climate weak nocturnal UHI decreases to the benefit of negative UHIs leading to a significant decrease in the nocturnal UHI over the city, (iii) for the inline run, nocturnal UHIs stays always positive but the frequency of the strong UHI decreases significantly in the future by 1 °C.

Further integration of different spatial scales in concentration modeling is important for assessing the European limit values for NO

2

. The local NO

2

-concentrations are influenced by the regional background, the local emissions and the street canyon effects. Therefore, it is important to combine all these contributions in the model chain that one wants to use for such an assessment. In this paper, we present the results of a coupled model chain, using an intelligent measurement interpolation tool, a bi-gaussian plume model and a street canyon model to simulate the concentrations over the city of Antwerp, Belgium. The results of this model campaign are validated against weekly averaged NO

2

measurements at 49 locations in the city of Antwerp, during both a winter and a summer week. It is shown that the model performed well, with an R

2

between 0.73 and 0.90, RMSE around 5 μg/m

3

and small biases. Next to this validation, the performance of different model parts is shown, in order to provide information on the importance of the different components.

In this paper, we investigate the influence of the changing NOx-split in traffic emissions on the ambient NO2-concentrations. First of all, we show that the NOx-split is indeed changing, by looking at measurements in tunnels and at urban locations. Secondly, we analyze on a local scale the effect of this changing split. It is shown that the a large influence can be found. Finally, we discuss some of the consequences of these changes, for instance, on the effectiveness of the EU-legislation.

The first results of a model validation test case, carried out in the research context of the COST ES1006 Action, are outlined and discussed. The validation exercise was established with an application-oriented approach, devoted to the investigation of the modeling performances in the emergency-response framework.

Air pollution can severely affect the health of citizens living in urban environment. Monitoring and assessment of air pollution by applying appropriate modeling tools are essential to provide information on air quality to the public. Several air quality forecast and information systems exist in Europe calculating present and predicted concentrations and potential exceedances using a combination of weather forecasting and chemical composition simulations. This paper presents an air quality prediction model system for the area of Budapest. The forecasting modeling system recently consists of a version of the WRF meteorological model and the CHIMERE chemistry transport model. The model system calculates the PM

10

, NO

2

, SO

2

and O

3

concentration values for the city area with 1-h temporal resolution. In this work we present the results of the validation of the prediction system. In the evaluation work, the PM

10

, NO

2

and O

3

data detected by the air quality monitoring network of Budapest, as well as the forecasted air pollutant concentration values of the air quality prediction model system are used.

The new residential district ‘CityLife’ is under construction inside the city of Milan, northern Italy, replacing the old structure of the trade fair. It consists of relatively insulated blocks, surrounded by many gardens and commercial buildings and no car traffic at the surface. The objective of this study is to simulate and compare the pollution level inside this new area with a more common residential district located not so far away. High resolution simulations have been performed inside two domains – having linear dimension of approximately 1 km – using a microscale modeling system, taking directly into account the effects of buildings and street canyons to the atmospheric mean flow and turbulence. Emissions coming from car traffic (around the new district and inside the old one), underground traffic (emerging from ground-level openings in the new district) and heating systems (during winter) have been taken into account. Background values have been estimated from local measurements using a box model and added to the simulated concentration fields in order to produce a complete view of the local pollution levels. Both average levels of pollution at district scale and local behaviours into sub-zones of the two domains where similar activities are supposed to take place are compared showing the potential benefits inside the new district both in absolute terms and in percentage, setting in evidence which portion of the pollution can be reduced by local interventions.

The turbulent velocity field of a dense fluid release has been measured in a water tank experiment carried out in the TURLAB laboratory in Turin (Italy). A vertical density driven current was created releasing a saline solution in a water tank with no mean flow. The experiment reproduces in physical similarity, based on the density Froud number, the release of a dense gas in the atmosphere and the PIV technique has been used to analyse the buoyancy generated velocity field. The high temporal and spatial resolution of the measurements gives a deep insight to the problems of the bouncing of dense gases and of the creation of the outflow velocity at the ground. The experimental findings are used to test and improve an original method for the dispersion of a positively and negatively buoyant plume. The method is based on the idea of Alessandrini and Ferrero (Phys. A 388:1375–1387, 2009) for the treatment of a background substance entrainment into the plume and it consists on the introduction of two fictitious scalars inside the Lagrangian Model SPRAY. The fictitious scalars represent the density and momentum difference between the plume and the environment air that naturally takes into account the interaction between the plume and the environment. As a consequence, no more particles than those inside the plume are released to simulate the entrainment of the background air temperature. In this way the entrainment is properly simulated and the plume sink is calculated from the local property of the flow. The approach is wholly Lagrangian.

The Gaussian plume model AEROPOL 5 is applied to estimate the yearly average NO

2

concentrations in Tartu, the second largest town of Estonia with about 100,000 inhabitants, for RHINE study. We apply the porosity concept by Genikhovich E, Gracheva I, Filatova E (Modelling of urban air pollution: principles and problems. In: Borrego C, Schayes G (eds) Air pollution modelling and its application, XV. Kluwer, New York, pp 275–283, 2002) in post-processing of modelled ground-level concentrations: the area under buildings is excluded from dispersion volume in each grid cell, thus the concentration is divided to the fraction of porosity, i.e. non-built-up area. It appears that porosity correction substantially enhances the site-wise correlations between model-estimated and measured concentrations, bringing the underestimated levels in particular monitoring sites closer to reality. Moreover, correlations are even higher, when dividing the “raw” modelled concentrations by squared porosities. We suppose that reason of non-linearity is in slowing down the wind between the buildings.

In 1999, a pilot monitoring program was conducted by PDVSA in order to assess air quality changes in Caracas as part of elimination of leaded gasoline in Venezuela, which started in the year 2000 with the implementation of premium unleaded gasoline and automobiles provided of catalytic converters. Elimination of leaded gasoline was completed in 2005 when its use was banned. This study presents an evaluation of an air quality model performance estimating carbon monoxide concentrations in Caracas from 2000 to 2010 in two scenarios: (i) fixed mixing height and (ii) variable mixing height, which consisted in a quantitative comparison of the box model output with observed data of carbon monoxide concentrations. The box model required information of seasonal behavior of meteorology, specifically wind speed data. In order to estimate carbon monoxide emissions, a traffic model which consider growth of vehicles fleet and employs emission factors associated with fuel quality changes during the study period was used. Results show that the model can estimate carbon monoxide concentrations with an agreement index of 0.52 and 0.57 that are acceptable. Some difference were found between observed and estimated values, mainly during the years 2005 and 2010, related to changes in vertical diffusion and another carbon monoxide source not considered in this model. The box model allows predicting concentrations of carbon monoxide with an NMSE of 0.11 and 0.12 when comparing with observed values for two evaluated scenarios.

Configurations of avenue-trees and a central park in an idealized urban neighborhood and their implications on traffic pollutant concentrations at the pedestrian level were investigated with Computational Fluid Dynamics (CFD). Steady state simulations were performed using a Reynolds Stress Model (RSM) extended with additional terms to represent the effects of vegetation on air flow. The results show that the type of configuration of avenue-trees and/or park has a clearly noticeable effect on the overall pollutant distribution and on the maximum concentration. The central park was found to lead to a general reduction of concentrations in its immediate vicinity and at locations downwind.

Regional air quality models are frequently used for regulatory applications to predict changes in air quality due to changes in emissions or changes in meteorology. Dynamic model evaluation is thus an important step in establishing credibility in the model predicted pollutant response. Beginning in 2003, the U.S. Environmental Protection Agency required substantial reductions in NO

x

emissions from power plants in the eastern U.S. which resulted in a decrease in ozone concentrations at monitoring sites in this region. This observable change in air quality has been used previously as a case study for dynamic evaluation of the Community Multiscale Air Quality (CMAQ) modeling system. Evaluation studies of previous CMAQ versions have shown that the model predicted decrease in ozone from 2002 to 2005 is less than the observed ozone change in this region. In this study, summertime simulations of 2002 and 2005 were conducted using the CMAQ version 5.0 modeling system released in 2012 that included several model updates aimed at improving the model predicted response, including updates to important model inputs. “Cross” simulations were conducted to separate the modeled impact of the emissions changes on ozone concentrations from the changes attributable to differences in meteorology across these years. Results show the change in the upper end of the ozone distribution explained by emission reductions is similar in magnitude to the change in ozone due to changes in meteorology across these years. Overall, the observed ozone decrease in the eastern U.S. continues to be underestimated by the model at both urban and rural monitoring sites.

An ensemble Kalman filter data assimilation system has been coupled to the Danish Eulerian Hemispheric Model (DEHM), a hemispheric-scale, offline, chemistry-transport model. We present an evaluation of the performance of the assimilation system in the framework of an Observing System Simulation Experiment (OSSE). This involves assimilating “pseudo-observations” derived from a reference simulation, and then comparing the assimilation run with the reference run. We focus on nitrogen dioxide and the pseudo-observations are generated to mimic the spatial/temporal pattern of retrievals from the Ozone Monitoring Instrument.

In regional model simulations, ozone introduced by boundary conditions (BCs) is important to model performance, model sensitivity and for policy reasons such as transport assessments. Quantitatively tracking the contribution of ozone BCs throughout a regional model is highly desirable. A simple and efficient solution is to add a tracer (O3BC) to the regional model simulation with a source term equal to the ozone BC. However, it is also necessary to model the removal of O3BC by deposition and chemistry. Modeling tracer deposition is relatively straightforward but modeling the photochemical removal of O3BC is complex. We developed a chemical mechanism for the photochemical removal of O3BC that is compact, represents all major removal pathways (i.e., photolysis, reactions with HOx radicals), accounts for important dependencies on NO, and can be implemented as an extension of the main gas-phase chemical mechanism. The method provides great flexibility by permitting O3BC to be factored into components that are then tracked independently. Example components of potential interest are different lateral boundaries segments (north vs. south boundary, etc.), altitude segments (boundary layer vs. free troposphere vs. stratosphere) and geographic origin (e.g., Asia vs. North America vs. Europe). Information about the geographic origin of O3BC can be derived from global model sensitivity simulations. The method is implemented in the CAMx regional model by using an existing model option to add reactive tracers (RTRAC) with complex chemistry. The method is applied and evaluated in simulations for continental North America.

In this study we present a novel approach for improving the air quality predictions using an ensemble of air quality models generated in the context of AQMEII (Air Quality Model Evaluation International Initiative). The development of the forecasting method makes use of modeled and observed time series (either spatially aggregated or relative to single monitoring stations) of ozone concentrations over different areas of Europe and North America. The technique considers the underlying forcing mechanisms on ozone by means of spectrally decomposed previsions. With the use of diverse applications we demonstrate how the approach screens the ensemble members, extracts the best components and generates bias-free forecasts with improved accuracy over the candidate models.

We propose a statistical air quality model that simultaneously performs two major tasks: (1) provides a computationally inexpensive means of modelling and forecasting complex space-time air pollution processes, (2) enables an informative and statistically defensible approach for evaluation of air quality models. Rather than working with raw data, we analyze the spatio-temporal variability of a pollution process by extracting data features using Empirical Orthogonal Function decomposition. Our modelling approach thus avoids the complications of point-to-point comparison, and the statistical model’s flexible structure allows for extensive fine-tuning. We develop and demonstrate our approach on observations and CMAQ model output of ozone episodes in the Lower Fraser Valley (LFV) of British Columbia, Canada. We believe the principle and methodology of our statistical analysis is applicable to a wide range of air pollution problems.

In this study, we provide an overview of the second phase of the Air Quality Model Evaluation International Initiative (AQMEII). Activities in this phase are focused on the application and evaluation of coupled meteorology-chemistry models. Participating modeling systems are being applied for annual simulations over both North America and Europe using common emissions and boundary conditions. We present an overview of these common input datasets. In addition, we discuss some of the planned analysis approaches using WRF-CMAQ simulations performed over North America for both 2006 and 2010.

AQUM is an online air quality modelling system which is used to provide the operational Met Office air quality forecast for the UK. The standard model configuration runs at a resolution of 12 km and covers a domain including the UK and part of Western Europe. The model is routinely verified against near-real-time surface pollutant measurements from the UK Automatic Urban and Rural Network (AURN) to provide a continuous evaluation of model performance. We have developed a new configuration of AQUM to run on the AQMEII (Air Quality Modelling Evaluation International Initiative) Phase 2 European domain at a resolution of 22 km, using the prescribed AQMEII emission datasets. This latter dataset contains wildfire emissions which are not included in the standard AQUM emissions. An initial analysis is conducted to compare the emissions over the UK used as input to both models. Model simulations for 2010 from the AQUM standard and AQMEII configurations are compared to AURN surface observations and an analysis of the effect of the differing model domain, resolution and emissions is made. In 2010 Russian wildfires constituted a significant additional source of pollutants in Eastern Europe; we evaluate the impact of these wildfire emissions on UK air quality.

The Air Quality Model Evaluation International Initiative (AQMEII) Phase-2 aims to inter-compare on-line couple regional-scale models over North America and Europe. Common chemical boundary conditions, emissions, and domain configurations are pre-defined.

In this study, we present the 2010 annual modelling results from two on-line integrated models, NMMB/BSC-CTM and Enviro-HIRLAM, as a contribution to the European domain simulations of AQMEII-Phase2 initiative and COST Action ES1004 EuMetChem. Both models have been configured according to requirements of AQMEII-Phase2 initiative. Model results will be evaluated against ground-based monitoring stations and vertical profiles. The performance and capability of both systems to reproduce surface and vertical structure of relevant reactive gases and aerosols will be discussed.

In this study we investigate the ability of the Chemistry Transport Model (CTM) LOTOS-EUROS to explain the observed decrease in secondary inorganic aerosol (SIA) and its precursors between 1990 and 2009 over Europe. The model explicitly accounts for cloud chemistry and aerosol thermodynamics. The results have shown that the model largely captures the observed trends in SIA and its precursors’ concentrations while it underestimates the interannual variability. Using a source-apportionment module the amount of SIA formed per unit emission was traced for a number of regions. The results show 20–50 % more efficient SO

4

2−

formation in 2009 compared to 1990, whereas the change in NO

3

−

formation per unit NO

x

emission is lower (−10 % to +20 %) for the same time period.

The 2-way coupled WRF-CMAQ meteorology and air quality modeling system is evaluated for high-resolution applications by comparing to a regional air quality field study (DISCOVER-AQ). The model was modified to better account for the effects of urban environments. High-resolution spatial data for fractions of impervious surfaces and tree canopy from the National Land-Cover Database (NLCD) were processed for each grid and used to scale ground heat capacity and to constrain vegetation coverage and characteristics. These simple algorithms along with modified albedo and roughness length in urban areas result in improved simulation of urban heat island and urban boundary layers. The reduced nocturnal stability and enhanced vertical mixing lead to reduced temperature and humidity biases and reduced under-predictions of ozone concentrations in urban areas.

At the Royal Meteorological Institute of Belgium the chemical transport model (CTM) CHIMERE runs at a spatial resolution of 7 × 7 km on a domain, covering Belgium and The Netherlands (680 × 680 km). The CTM is first run on a coarse resolution of 50 km covering Western Europe, using the meteorological fields of ECMWF. The emission database is provided by EMEP. The output of this model run is used as input for the boundary conditions for the coupled high resolution chemical transport model run.

From previous studies, it has been shown that this one-way nested coupling with the limited area NWP models ALADIN and ALARO (7 × 7 km) already improved 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 emission database used as input for this high resolution model is the TNO/GEMS emission database. In this study, we will also investigate the influence on the modelling of particulate matter when the CTM is coupled to these different high resolution NWP-configurations. We will investigate the air quality model forecast performance for a three day lead time on a daily basis, covering a one year time period (2008), implementing inline a new land surface scheme (SURFEX) – with and without taking into account The Town Energy Balance (TEB) – on the modelled ozone- and particulate matter concentrations.

In the frame of the European Consortium for Modeling of Air Pollution and Climate Strategies (EC4MACS) the CHIMERE chemistry transport model has been run over Europe for the entire year 2009 with a spatial resolution of 7 km with the aim of assessing the urban impact on daily exceedances of PM and NO2 in European cities. In order to better capture these urban impacts, improvements on urban scale meteorology, vertical resolution and emissions have been implemented. In the current work an evaluation of the model results against the AIRBASE European monitoring network measurements is done using model performance indicators (MPC) based on observation uncertainty.

The MPC used in this approach, constructed on the hypothesis that model results are allowed the same margin of uncertainty as measurements, are developed for four statistical indicators (Root Mean Square Error, Normalized Mean Bias, Normalized Mean Standard Deviation and temporal correlation) to summarize the model-observation errors in terms of phase, amplitude and bias. The utility of this approach is to provide a performance scale to inform the user on the expected value an indicator should reach for a particular modeling application. These indicators are then used to identify the strengths and weaknesses of the model application in terms of geographical areas, cities, pollutants and/or period of the year.

We explicitly address the fundamental issue of member diversity in multi-model ensembles. Common biases and redundancy are the two issues directly deriving from lack of independence, undermining the significance of a multi-model ensemble, and are the subject of this study. These two issues are analysed in detail using the AQMEII ensemble of AQ model results for four air pollutants in two European regions. We show that models share large portions of bias and variance, extending well beyond those induced by common inputs. We propose and discuss methods of member selection and rate the ensemble performance they produce. We conclude that, although independence of outputs may not always guarantee enhancement of scores (but this depends upon the skill being investigated) we discourage selecting the members of the ensemble simply on the basis of scores, that is, independence and skills need to be considered in separation.

The U. S. National Air Quality Forecasting Capability (NAQFC) provides air quality forecast for the nation by disseminating numerical model predicted surface concentration of O3 and PM2.5 to the public. However, the fidelity of NAQFC is dependent on the accuracies of the emission projection factors employed to estimate the various emissions. This study focuses on comparing variability of surface NOx and O3 concentrations for two emission modeling scenarios for July of 2011: the Base Case and a New Emission Case. The Base Case used the U.S. EPA 2005 National Emission Inventory (NEI2005), its projection procedures adhered to a standard practice used by NAQFC since its inception in 2003. The New Emission Case adopted a scaling procedure based on more recent EPA data demonstrated a significant reduction of the mobile source’s share of NOx emission among the major contributors. It reduced from 33.6 % in the Base Case to 25.2 % in the New Emission Case. This is even more significant if one takes into account the large increase in vehicle miles traveled since 2005. The NOx SIP Call had achieved significant reduction of NOx emission from power plants, but still lagged behind that achieved by the reduction in the on-road vehicular (mobile) exhausts. Geographically population change trends in the last decade do not necessarily translates into proportional changes in NOx emission.

Olesen et al. (Bound Layer Meteorol 131(1):73–83, 2009) have shown that current building downwash parameterizations, which are essential in the vicinity of build up areas, perform poorly at very short distances. Therefore, a new building downwash model has been developed based on the Thompson-wind tunnel dataset. The model consists of the Gaussian transport and diffusion equation, in which the variables xR (distance source-receptor) and Hs (stack height) have been replaced with functions that define a receptor dependent virtual origin. The virtual source functions account for the impact of building downwash (1) by changing the height of the plume to account for the vertical displacement of the plume and the extra turbulence around the building, (2) by using an upwind displacement of the plume origin to simulate the impact of the wind-upward displacement of the plume and (3) by accounting for the entrainment of clean air into the plume by increasing the distance between the virtual source and the receptor point. The model results are compared to the several measurement time series in Flanders, Belgium, at industrial locations where building downwash plays an important role. We have shown an increased modeling ability in all cases: for instance, biases ranging from −4 to −33 ng/m

3

(−46 to −77 %) improve to values ranging from 1.3 to −7.5 ng/m

3

(−29 to +14 %) and R

2

values improve dramatically (up to 0.83 from 0.65 in cases with a strong building downwash effect).

Knowledge on the forcing mechanisms (meteorological and chemical) that come into play in the Arctic environment is highly uncertain. We analyse data from measurements and mesoscale meteorological modelling for periods in summer 2011 and winter 2011/2012 to elucidate the boundary-layer features at Station Nord situated in Northern Greenland (81.6 N, 16.7 W). A major challenge for modelling is to connect local-scale observations with larger scales modelling of the atmosphere. In particular, in summer, bare soil in the vicinity of the station may not be present in the global boundary and surface conditions. Thus, in summer the deviations of modeled from measured values of temperature and humidity near the surface are larger compared to winter. We found that the underestimation of temperature near the ground is larger at clear sky compared to cloudy conditions; and the underestimation reached up to height 1-1, 5 km at clear sky and up to the first 100 m for the cloudy days. The measured wind speed profiles showed high variability, while the modeled were smoothed. During summer the modeled wind speed was close to or larger than the measured without clear indication for the role of clouds. In winter, the over-estimation of wind speed was more pronounced.

Sofia is 1.5 million citizens city situated in a mountain valley and thus characterised with complex structure of the planetary boundary layer (PBL) caused by both terrain and urban features. Stagnant stable conditions lead to fogs and smog in winter, and dry climate leads to high PM concentrations in summer albeit the deep PBL. In this study we evaluated the dynamics given by two configurations of the Weather Research Forecasting (WRF) model against data from consecutive high vertical resolution radiosoundings during 5 days in Sofia under early autumn convective conditions in PBL (Sofia Experiment 2003). Statistical comparisons between modelled and measured parameters showed that WRF simulated relative humidity, temperature and wind direction better in the transition than in the afternoon hours, while the wind speed was insignificantly better simulated in the afternoons. We found that WRF with Mellor-Yamada-Janjic PBL scheme simulated relatively well the vertical profiles of all analysed parameters, except for wind direction which was poorly resolved in the lowest 1,000 m above ground.

An application of the CALPUFF modeling system in the near-field in an area of very-complex topography is presented. Different configurations for both the meteorological and the dispersion simulations were tested. The RAMS high-resolution (1 km) output as well as the data from a network of surface stations and the high-resolution data from a wind profiler radar (WPR) located in the area were available for assimilation and validation purposes. The 250 m resolution three-dimensional CALMET meteorological fields, after being validated against surface and upper air data, were fed into CALPUFF to simulate the potential impacts of the emissions of a coke factory over a nearby urban area. Benzene was used as a tracer of the emitted pollutants. The results of the dispersion simulations were compared against the available hourly benzene records. Our results showed that, no matter the selected synoptic forcing for RAMS (NCEP/ECMWF), the assimilation of meteorological data into CALMET is necessary for a correct representation of the wind flows in the area, and as a consequence, for a correct representation of the actual benzene impacts.

Shortwave radiation is a key meteorological component influencing formation and destruction of many atmospheric pollutants, most of all tropospheric ozone. It is the main driver of photochemical reactions, affects surface temperature and enhances biogenic VOC production. Because uncertainties in meteorological fields provided by meteorological models highly affect chemical transport model simulations, accurate information on spatial and temporal variability of shortwave radiation is needed for reliable air quality modeling. The main aim of this study is to assess the meteorological model performance in representing observational data. In this study, the Weather Research and Forecasting (WRF) model is applied to the area of Lower Silesia, Poland, in order to test three shortwave radiation schemes: Goddard, RRTMG and GFDL scheme. The test period is a high ozone episode of 17.06-04.07.2008. Simulations were run with different shortwave radiation options, while all other physics parameterizations remained constant. The results were then evaluated based on radiation measurements conducted in the Observatory of the Department of Climatology and Atmosphere Protection in Wroclaw, Poland. There are some discrepancies between the employed parameterizations both in terms of quality of the results and computational costs but in general, all schemes applied show reasonable consistency with observations.

A sensitivity test has been carried out for five PBL schemes using the WRF meteorological model over Italy for February 2008. The model has been applied over the polluted area of the Po Valley with a grid resolution of 5 km. Model results have been compared to surface measurements as well as vertical profiles by balloon soundings, meteorological balloons and Lidar measurements. Results show a good agreement between simulations and observations for mixing ratio, while model reveals higher bias in reconstructing temperature, wind speed and planetary boundary layer height. The best performances in reconstructing PBL height are highlighted for YSU scheme, while other ground meteorological fields are better reproduced by ACM2 parameterization.

Synoptic weather pattern plays an important role in regional air quality. It influences strongly the daily maximum surface level ozone concentrations. We utilize Spatial Synoptic Classification methodology, surface ozone simulations from the NOAA National Air Quality Forecast Capability and observations from the US EPA Air Quality System, to investigate the behavior of surface ozone and air quality model’s performance in different synoptic weather patterns. This study was conducted for the following six air quality air-shed regions: California (CA), Michigan Lake (MI), Ohio River Valley (OV), Northeastern (NE), Southeastern (SE), and Texas (TX), during the ozone season (May to September) in years 2009–2011. Based on air mass’ thermal characteristics and origins, the SSC methodology provides six types of air mass classifications: Dry Polar (DP), Dry Moderate (DM), Dry Tropical (DT), Moisture Polar (MP), Moisture Moderate (MM), and Moisture Tropical (MT). Results show distinct behaviors of ozone’s magnitude and model’s performance among the various air mass types, implying the impact of synoptic weather patterns to regional air quality, and NAQFC’s success in reproducing these patterns. The MT and DT air mass types are the most important types conducive to high surface ozone episodes throughout theiroccurrence frequency and high ozone production efficiency, respectively. The MT air mass also shows highest in both mean bias and root mean square error, implying the high uncertainty in model performance due to cloudy or unstable atmospheric conditions.

A comparison between two WRF-Chem simulations has been performed in the framework of the AQMEII modeling initiative for the calendar year 2010. Model configurations shared the same physics options, but different chemical parameterizations were selected. The first run adopted RADM2-MADE/SORGAM chemical mechanisms, whereas the second one employed the CBMZ-MOSAIC chemical option. Comparisons between two simulations show that CMBZ is higher than RADM2 for gas concentrations. On the contrary, MADE-SORGAM reproduces higher aerosol values than MOSAIC, especially over land. The same behaviour is observed in comparison against ground – based measurements. Both simulations underestimate observed values of NOx and PM2.5.

Satellite data presents an unprecedented opportunity to improve emission inventories at a near-real-time pace. Here we demonstrate how to utilize satellite oceanic and atmospheric products to improve emission forecasting. First, we present the development and validation of a global high resolution marine isoprene emission product. Isoprene emission is calculated from NOAA global weather forecasting data and Chlorophyll-a and light attenuation rate at 490 nm (K490) data derived from the Moderate Resolution Imaging Spectrometer (MODIS) aboard Aqua. The emission product is validated with isoprene measurements from field campaigns. In the second case, nitrogen dioxide (NO

2

) data from the EPA Air Quality System (AQS) and the Ozone Monitoring Instrument (OMI) are used to examine the long-term trends in nitrogen oxides (NO

x

) emissions for the NOAA National Air Quality Forecasting Capability (NAQFC). Comparing of summertime NO

x

data from OMI, NAQFC and AQS over New York between 2005 and 2011 shows a similar reduction level from all datasets (33 % reduction from 2005 to 2011), but OMI and AQS agree better while NAQFC emission inventories fail to catch the gradual progression of emission reduction. These case studies, in addressing various aspects of emission uncertainty, collectively demonstrate that satellite remote sensing can play an important role in improving emission forecasting and, hopefully, air quality predictions.

In a study, we examine impact of assimilation in-situ measurements on the prediction of fine aerosol concentrations over North America in the summer of 2012. We use an online meteorology-chemistry model WRF-Chem and an assimilation system which includes the Gridpoint Statistical Intepolation (GSI) and an Ensemble Kalman Filter. We note large initial benefit of data assimilation and relatively quick deterioration of forecast verification scores with time. Causes for such deterioration include deficiencies of the initial state resulting from the lack of observations of the individual aerosol species and their vertical distribution, weaknesses of chemical and physical parameterizations and errors in emission sources.

Most regional scale models that are used for air quality forecasts and ozone source attribution do not adequately capture the distribution of ozone in the mid- and upper troposphere, but it is unclear how this shortcoming relates to their ability to simulate surface ozone. We combine ozone profile data from the NASA Earth Observing System (EOS) Tropospheric Emission Spectrometer (TES) and a new joint product from TES and the Ozone Monitoring Instrument along with ozonesonde measurements and EPA AirNow ground station ozone data to examine air quality events during August 2006 in the Community Multi-Scale Air Quality (CMAQ) and National Air Quality Forecast Capability (NAQFC) models. We present both aggregated statistics and case-study analyses with the goal of assessing the relationship between the models’ ability to reproduce surface air quality events and their ability to capture the vertical distribution of ozone. We find that the models lack the mid-tropospheric ozone variability seen in TES and the ozonesonde data, and discuss future work to determine the conditions under which this variability appears to be important for surface air quality.

This study is a first step towards building an atmospheric chemistry reanalysis modeling system. We aim to provide the air quality science community with three-dimensional (3D) reanalysis atmospheric chemical fields over the conterminous U.S. (CONUS). This initial 3D gridded reanalysis product is available at 12 km horizontal grid spacing with 22 uneven vertical levels extending from surface to 100 hPa. The principal components of the modeling system are the Weather Research and Forecasting meteorological model, a chemical data assimilation model based on an optimal interpolation scheme, and the U.S. EPA Community Multi-scale Air Quality modeling system (CMAQ). Only the Moderate Resolution Imaging Spectro-radiometer (MODIS) Aerosol Optical Depth observations are assimilated as we focus on the aerosol reanalysis at this early stage. CMAQ predictions before and after the assimilation are evaluated against the AIRNow surface PM2.5 (Particulate Matter smaller than 2.5

μ

m in diameter) measurements. Based on the preliminary results, the future directions to improve the chemistry reanalysis modeling system are discussed.

The U.S. National Aeronautics and Space Administration (NASA) has started a series of regional scale multiple platform field measurement intensives between 2011 and 2014. These measurement campaigns are collectively called the Deriving Information on Surface Conditions from Column and Vertically Resolved Observations Relevant to Air Quality (DISCOVER-AQ) project. The emphasis of spatial collocation of multiple space-, air- and ground-based measurements facilitate the characterization of the vertical distribution of various air pollutants. During the January 16 to February 8, 2013 DISCOVER-AQ campaign over the San Joaquin Valley (SJV), California, one of the frontier science questions addressed concerned the frequent wintertime Particulate Matter (PM) standard violation over SJV and its vicinities. This study represents our contribution of real-time air quality forecasting to support flight-planning during the campaign as well as post analysis and model evaluation for some of the air pollutants. Two sets of real-time forecasts based on a coupled National Center for Environmental Prediction (NCEP) North American Model (NAM) and the U.S. EPA Community Multi-scale Air Quality Model (CMAQ) are included in the study: (1) a 12 km horizontal resolution domain over the Conterminous U.S., and (2) a 4 km resolutiondomain over the SJV and its adjacent areas nested-in from (1) as its parent. During the campaign there were several episodes of high surface PM concentration in the lowest hundreds of meters over SJV. We performed post-analyses of meteorological parameters such as the planetary boundary height and of chemical characteristics such as verification of modeled concentration of PM components at surface levels. Comparisons of the performance of the two forecasts will shed insight on horizontal resolution requirement to capture the meteorology and chemistry characteristics in SJV air quality modeling. These analyses will aid policy and regulatory relevant decisions as well as assist in improving the model for future applications.

The NASA GEOS-5 model was used to extend the MERRA reanalysis with five atmospheric aerosol components (sulfates, organic carbon, black carbon, desert dust, and sea-salt). The obtained eight-year (2002–2009) MERRA-driven aerosol reanalysis (MERRAero) dataset was applied to the study of aerosol optical thickness (AOT) trends over the Ganges basin and north-west Bay of Bengal (BoB) in the early post-monsoon season. In October, in the absence of aerosol sources in north-west Bay of Bengal (BoB), MERRAero showed increasing AOT trends over north-west BoB exceeding those over the east of the Ganges basin. Various aerosol components showed strong increasing AOT trends over north-west BoB. Our analysis showed that the AOT trends over north-west BoB were reproduced by GEOS-5, not because of MODIS AOT assimilation, but because of the model capability of reproducing meteorological factors contributing to AOT trends. The following factors contributed to the increasing AOT trend over the area in question in October: an increasing number of days when prevailing winds blew from land to sea, resulting in a drier environment and an increase in air pollution over north-west BoB; wind convergence was observed over north-west BoB causing the accumulation of aerosol particles over that region, when prevailing winds blew from land to sea.

In this contribution we will show the impact on air pollution concentration of a Fire developed in the Surrounding area of Murcia (Spain). The impact on air pollution concentrations has been done using the WRF/Chem model developed by NCAR (US) and the Fire model implemented into WRF in on-line mode. The fire spread has been considered in on-line mode to change the use of the different land use grid cells as the fire spreads. This process changes the land use and affects substantially to the Fuel Moisture Content (FMC) modifying the surface turbulent energy balance and the pollution dispersion consequently. A new fuel moisture content model has been developed. The new module allows each time step to calculate the fuel moisture content of the dead fuels (1 h, 10 h, and 100 h) and live fuels. Two simulations have been performed over the Murcia area on September, 7th, 2010 with 9 h of fire starting at 19:09 over an area of 7 km × 7 km: (a) with a fire of 9 h simulation we run the WRF/Chem with 200 m spatial resolution over the fire domain in on-line mode with FIRE model. Emissions from fire have been accounted on; and (b) the same simulation than in case (a) but without fire (no emissions and no changes in the land use).

PM10 forecasts for Europe and the Netherlands are part of the daily LOTOS-EUROS products, next to ozone forecasts. So far, ozone ground observations and OMI NO

2

column observations were assimilated in the forecasting system, using an Ensemble Kalman filter. Previous studies have shown that assimilation of daily PM10 ground observations improve the PM forecasts. Ideally, ozone, NO

2

, AOD and PM10 can be assimilated within one analysis. If we want to combine ozone and PM assimilation in one system, assimilation of hourly PM observations is needed and the model system was extended by assimilation of PM10 and PM2.5 ground observations.

We have investigated optimal settings of the system, taking several aspects into account. First of all, a complication with regard to ozone and NO

2

is that LOTOS-EUROS underestimates the PM concentration. This was previously accounted for by a bias correction, but can more elegantly be solved by introducing a new species of unknown composition which is allowed to vary independently. Secondly, hourly PM observations can have relatively large errors and are sometimes influenced by very local conditions. Therefore, sensitivity experiments were done to investigate the sensitivity to and optimal settings for the correlation length between stations, error constraints on observations, noise factors on the emissions and correlation timescales. We present the lessons learnt and the current performance of the system.

Exposure assessment and development of control strategies are limited by the air pollutants measured and the spatial and temporal resolution of the observations. Air quality modeling can provide more comprehensive estimates of the temporal and spatial variation of pollutant concentrations, however with significant uncertainties. Source apportionment, which can be conducted as part of the air quality modeling, provides estimates of the impacts of sources on the mixtures of pollutants and contains surrogate estimates for pollutants that are not measured. This study details results using a novel spatiotemporal hybrid source apportionment method employed with interpolation techniques to quantify the impact of 33 PM

2.5

source categories. The hybrid model, which aims to reduce estimating uncertainties, adjusts original source impact estimates from a chemical transport model at monitoring sites to closely reflect observed ambient concentrations of measured PM

2.5

species. Daily source impacts are calculated for the contiguous U.S. Two interpolation methods are used to generate the data needed for spatiotemporal hybrid source apportionment. Hybrid adjustment factors are spatially interpolated using kriging, and daily observations are calculated by temporally interpolating available monitoring data. Methods are evaluated by comparing daily simulated concentrations—generated by reconstruction of source impact results—to observed species concentrations from monitors independent of model development. Results also elucidate U.S. regions with relatively higher impacts from specific sources. Monitoring data in this study originated from the Chemical Speciation Network (CSN), EPA-funded supersites, and the Southeastern Aerosol Research Characterization (SEARCH) Network. Results are to be used in health impact assessments.

An operational air quality forecasting model based on the Weather Research and Forecasting (WRF) model and the Community Multiscalar Air Quality (CMAQ) model is used to produce a three day forecast for O

3

, NO

2

, SO

2

, PM

10

and PM

2.5

and weather in the UK. In 2012 PM

2.5

was added to the UK air quality forecast. When compared to observations the forecast tends to underestimate the PM2.5 concentration. This study reviews the potential to improve the forecast by assimilation of measurement data for O

3

, NO

2

and particulate matter from the UK’s Automated Urban and Rural Monitoring Network (AURN) into the CMAQ daily forecast using a bias correction factors estimated by means of the Kalman filter (KF) approach.

Moderate particulate pollution events in mid-February and a high PM

2.5

pollution event between 5th and 8th March 2013 provided a good test case for evaluating the basic Kalman filter (KF) methodology. This improved the forecasts of daily mean and daily maximum 1-h concentrations of particulate matter PM

2.5

and ozone. It effectively eliminates the mean bias between the predictions and the measured concentrations and in most cases reduces the standard deviation of the bias. Using the basic algorithm on the more demanding task of predicting hourly average concentrations reduces the mean bias between the predictions and the measurement, but in many cases in this study it increased the scatter. More tuning to the Kalman gain factor is required before the hourly correction is recommended for application to UK forecasts.

In February 2013, in collaboration with the Air Quality Research Division, the Canadian Meteorological Centre (CMC) implemented into operations a new surface analysis for air quality species (ozone and PM2.5). The Regional Deterministic Air Quality Analysis (RDAQA) generates analyses every hour using the operational GEM-MACH Regional Air Quality Deterministic Prediction System (48 h forecasts on a domain with 10-km horizontal grid spacing and 80-vertical levels) to provide the trial fields. Surface observations are from Canadian regional data providers and the US EPA/AIRNow Program. An optimal interpolation scheme adapted to air quality is used to blend model and observations. A verification of the RDAQA shows major reductions in the error variance and bias of the analysis with respect to the model forecasts as compared to observations.

Environment Canada produces twice-daily, 48-h operational air quality (AQ) forecasts for a domain covering North America. At the core of the forecast system is the GEM-MACH model, an on-line coupled meteorology and chemistry model that includes a representation of gas-phase, aqueous-phase, and heterogeneous chemistry and a number of particulate matter (PM) processes. In this paper, a brief description of the recent changes to the Canadian National AQ Forecasting System is given, followed by a discussion of future development plans. The objective for the next version of the system is to deliver improved AQ forecasts by improving initial and boundary conditions and representations of emissions and processes.