Air Pollution Modeling and its Application XXVI

In Eastern Germany winter episodes with PM10 exceedances of EU limit values are often connected to continental air masses, combining polluted air from Eastern Europe with air pollution from local urban sources. The EU air quality legislation requires the analysis of the contribution of such cross-boundary transport to exceedances for development of effective policy measures. To this end we have performed a source apportionment study to establish the main sources of particulate matter during high PM episodes in Eastern Germany. We have run the LOTOS-EUROS model with its labelling based source apportionment tool and found that the contribution from cross boundary transport becomes more dominant during the episodes. The results from the model are currently compared to a measurement based source attribution (PMF) and a more detailed evaluation of episodes with exceedances is being performed. The modelled concentrations have been evaluated against PM composition observations which revealed that especially the modelling of carbonaceous aerosol is challenging and lead to a large underestimation of modelled PM10 levels in winter. The use of an updated bottom-up inventory in combination with temperature dependent temporal variability for residential combustion emissions leads to an increase of carbonaceous aerosols and reduction of PM biases. Another improvement was realised through an update of the deposition routine over snow that leads to a strong reduction of the underestimation during cold PM episodes with snow conditions.

The eastern United States has seen dramatic air pollution emissions reductions since the turn of the century. These emissions reductions have in turn been linked to widespread reductions in ozone (O3)—between 2000 and 2016, the US EPA reports a reduction in 4th highest mean daily annual 8-hr O3 of 15% (from 82.3 to 69.6 ppb) across 206 sites nationwide. Reductions, however, have not been spatially uniform or linear with emissions reductions, and therefore motivate an investigation into spatial and source-specific O3 production efficiency (OPE). OPE is a measure of the number of O3 molecules produced per emitted NOX (NOX = NO + NO2) molecule. We assess OPE using both model-based and empirical approaches. We modelled July OPE in 2001 and 2011 using CMAQ-DDM version 5.0 with a 12 km resolution over the eastern US. CMAQ-modelled OPE is taken as a ratio of electricity generating unit and mobile source sensitivities, and controls for differences in O3 and NOZ deposition rates. Measurements were taken from the SEARCH network, which reports sub-daily observations of many gaseous and particulate species along with meteorological measurements at eight sites in the southeastern US. Using measurement data, we stratified days based on their emissions-independent photochemical state, and estimated OPE using a spline model to assess the relationship between O3 and NOX reaction products (denoted NOZ). Both approaches yield an increase in OPE with decreasing NOZ, indicating an increasing effectiveness at lowering O3 for subsequent NOX emissions reductions. Electricity generating unit OPEs are low near individual sources, but generally higher than on-road mobile source OPEs throughout the domain, suggesting that further utility NOX emissions reductions will reduce regional O3 concentrations more efficiently than mobile source NOX emissions reductions.

This study aims to quantify the contributions to surface ozone (O3) concentration in the Iberian Peninsula (IP) from the main NOx emission sectors in the region along with the imported O3 during a 10-day episode. The work applies an Integrated Source Apportionment Method within the CALIOPE air quality system at 4-km resolution. This study finds that the imported O3 is overall the larger contribution to its surface concentration. Contributions from local/regional sources are decisive in the O3 peaks downwind of main nitrogen oxides (NOx) hotspots in the IP under stagnant conditions.

This study uses a photochemical Air Quality Model applied across the continental US to identify source categories and chemical species (hydrocarbons and nitrogen oxides) that have the largest impact on concentrations of ambient formaldehyde. We contrast the sensitivities of formaldehyde to those of ozone. Although reactions of organic radicals with nitrogen oxide can produce high yields of formaldehyde, the concentrations are more sensitive to hydrocarbons. Biogenic sources of hydrocarbons contribute the most to formaldehyde sensitivity in July, with contributions from isoprene, other alkenes and direct emissions. These results indicate that different strategies may be needed to reduce ambient ozone and formaldehyde concentrations.

In this paper, we discuss the importance of biogenic volatile organic compound (BVOC) emissions used in air quality simulations and how the model results are affected by the choice of the BVOC emission model. The European air quality in 2011 was simulated using CAMx regional air quality model with two different BVOC emission models: PSI-model and MEGAN. Especially isoprene and monoterpene emissions calculated by the two models differed significantly both in amounts and their spatial distribution. In general, MEGAN produced much higher isoprene emissions while PSI-model generated more monoterpene emissions. The difference in emissions between the two models was shown to be as high as a factor of 3 in summer. The choice of the BVOC emission model had significant consequences especially on the formation of organic aerosols as well as on ozone and inorganic aerosols. Using MEGAN led to relatively higher ozone concentrations in summer while much more SOA (secondary organic aerosol) was formed when PSI-model was applied. Our results suggest that the amount and spatial distribution of BVOC emissions might affect the oxidant concentrations (OH and nitrate radicals, ozone) leading to significant differences in SOA, ozone, particulate nitrate and sulfate concentrations calculated by different BVOC emission models.

Researchers who perform air quality modeling studies usually do so on a regional scale. Typically, the boundary conditions are generated by another model which might have a different chemical mechanism, spatial resolution, and/or map projection. Hence, a necessary conversion/interpolation takes place which introduces additional error. In a broader sense, air pollution is a global issue, thus, limited area modeling on the regional scale is not well suited to represent long-range transport from key source regions in other parts of the world. We have developed a prototype system to link the Model for Prediction Across Scales—Atmosphere (MPAS-A) with the Community Multiscale Air Quality (CMAQ) model to address these shortcomings. Pollutant transport is conducted within MPAS-A, rather than in CMAQ, to provide consistency with the meteorological processes. A coupler has been constructed to facilitate data exchange between the two models. Initial test simulations show reasonable results when compared with observational data.

We model the changes in wet and dry deposition amounts of reactive nitrogen and sulfur over the 1990–2010 period using the WRF-CMAQ modeling system. WRF-CMAQ simulations for this 21-year period were conducted over a domain covering the northern hemisphere using a horizontal resolution of 108 km and a nested domain over the contiguous U.S. using a grid of 36 km resolution. The impacts of contrasting changes in emission patterns across the Northern Hemisphere, i.e., reductions in North America and Europe vs. increases across regions in Asia, on changing deposition amounts over terrestrial and aquatic ecosystems in these regions is analyzed.

A trend analysis is performed over the period 2006–2015 for concentration and deposition of nitrogen compounds over the Netherlands. The analysis is performed with high resolution (~1 km) model simulations with the Gaussian plume model OPS and the grid model EMEP4NL. Both models use the same MACC III emission distribution for countries outside of the Netherlands, and spatially more detailed emissions for the Netherlands itself. Emission totals per SNAP sector per country are used over the period 2006–2015, according to the latest CEIP expert estimates. The OPS model is driven with yearly specific meteorological fields provided by the Royal Netherlands Meteorological Institute (KNMI), while EMEP4NL is driven by meteorological output data from the open source WRF meteorological model. Results from the model calculations are first compared with measurements. Next, the focus of the analysis will be on the effect of atmospheric chemistry on trend analysis of nitrogen components, like ammonia and ammonium in the atmosphere, and the dry and wet nitrogen deposition to the surface: OPS strongly parameterizes the chemistry, whereas EMEP4NL uses a state-of-the-art chemistry scheme. The influence of atmospheric chemistry on modeled trends in concentration and deposition is determined by comparing the trend calculations of both models, and their connection with trends in emissions of precursor gases over the period 2006–2015.

Nitrogen and phosphorus are two main nutrients responsible for eutrophication of the Baltic Sea. Almost all phosphorus is entering the sea via rivers, whereas 20–30% of nitrogen is deposited from the air. Therefore, there is a need for monitoring atmospheric nitrogen deposition to the Baltic Sea. Time series of annual nitrogen depositions to the Baltic Sea have been calculated for the period 1995–2015 with the same version of the EMEP MSC-W model. They show significant inter annual fluctuations due to changes in meteorological conditions from one year to another. To reduce the influence of meteorological conditions on the results the so called “normalized” depositions have been also calculated. They indicate a clear decline of annual depositions of oxidized nitrogen to the Baltic Sea and only minor decline of reduced nitrogen depositions in the considered period. Emissions from Germany and Poland are the main sources contributing to deposition of oxidized nitrogen to the Baltic Sea basin followed by the ship traffic on the Baltic Sea and on the North Sea. Transportation and combustion are the main emission sectors contributing to oxidised nitro-gen deposition, whereas, agriculture is the dominating emission sector contributing to reduced nitrogen deposition.

The FRAME atmospheric chemistry transport model was applied to calculate the concentration of NH3 at a 1 km resolution over the UK. The results showed that the 1 μg m−3 critical level for NH3 was exceeded for 60% of the UK land area and the 3 μg m−3 critical level was exceeded for 3% of the land area. Model simulations using historical emissions suggested that average NH3 concentrations in the UK have increased by a factor of 2.5 between 1970 and 2010 due to both an increase in NH3 emissions and large reductions in SO2 emissions causing a reduction in the availability of H2SO4 to react with NH3.

The spatial distribution of Age of Air (AoA) in the stratosphere has been extensively used to evaluate and compare general circulation and chemical transport models. We performed multi-decade simulations of SF6 and several AoA tracers in the atmosphere with Silam chemistry-transport model driven with ERA-Interim reanalysis. The resulting distributions of AoA agree well with each other and with simulations made earlier, however disagree with AoA derived from observed SF6 concentrations in polar areas. The simulations of SF6 were made in two variants: fully passive passive, and with account for mesospheric depletion and for gravitational separation. The results indicate a good agreement with observations by MIPAS satellite instrument for non-passive SF6, and have substantial discrepancies in polar areas for passive SF6. Thus we conclude that the discrepancy between SF6 and modelled AoA originates from the violation of the assumption of the SF6 passiveness.

Air quality is primordially affected by anthropogenic emissions and has a tremendous impact on human health with more than 6 million premature deaths worldwide in 2015 (Landrigan et al. in The Lancet Commission on pollution and health, 2017) [3]. Biogenic emissions of aerosols such as pollen also impact the human wellbeing. The industrialized world suffers from a global increase in the burden of allergic respiratory diseases. Air pollution can influence both allergens and allergic subjects by increasing the immune reaction, and/or by an intensified biogenic emissions. In Europe, a quarter of the population suffers from pollinosis, whereas in some countries the prevalence is over 40%. To date, pollen of various trees and grasses in Belgium are monitored by the Belgian Scientific Institute for Public Health (Sciensano) at five stations on a daily basis. This sparse sampling cannot cover the spatial representativeness of the airborne pollen. Chemistry Transport Models (CTM’s) are therefore an interesting tool to both quantify and forecast its spatial and temporal distribution. Here we show the results of the spatio-temporal modelled birch pollen over Belgium using the CTM SILAM. This model is driven by 2008 ECMWF meteorological data and a MACC-III birch tree fraction map showing the spatial distribution of potential pollen sources. Pollen modelling is based on the temperature degree days approach.

A new modelling system, SMART, is under development for the simulation of accidental releases dispersion. The interfacing code ARAMIS was originally created to interface the non-hydrostatic atmospheric model MOLOCH and the Lagrangian stochastic dispersion model SPRAY. Here, a comparison between simulations with the new modelling suite SMART and the RMS modelling system, applied in several previous assessment studies, is presented for a case of a release in complex terrain in southern Italy. The new suite is planned to be adopted as a tool for emergency response purposes in any part of the Italian territory at any time.

Prince George British Columbia has among the highest levels of ambient particulate matter (PM) in western Canada. In order to effectively lower ambient PM levels, management agencies need to be able to attribute ambient levels to specific sources, which can then be targeted for reduction. Dispersion modelling can be used to attribute sources to ambient levels, but one issue is that emissions from many PM sources are poorly known and must be estimated. The Calpuff dispersion modelling system was applied to all known sources of PM10, PM2.5, NOx and SO2 affecting the Prince George airshed over the three-year period 2003–2005. The model results were evaluated by comparison with ambient levels as well as with results from a speciation study using Positive Matrix Factorization and Chemical Mass Balance techniques. This resulted in constraints on the emissions of some of the more poorly characterized sources. In addition, a web-based visualization and scenario tool was developed for air quality managers to enable them to make science-based decisions to improve air quality. The results of the modelling and source attribution will be discussed, and the web-based scenario tool demonstrated.

In late September and October 2017, Ru-103 and Ru-106 have been detected throughout the northern hemisphere by national environmental radioactivity monitoring networks and by the International Monitoring System that is being established to verify compliance with the Comprehensive Nuclear-Test-Ban Treaty. Ru-103 (half-life: 39.26 d) and Ru-106 (half-life: 373.6 d) are radioactive particulates that have no natural sources and for which there is no measurable global background.

The Comprehensive Air quality Model with extensions (CAMx) has been employing ISORROPIA for inorganic aerosol thermodynamic calculations. Recently, the Equilibrium Simplified Aerosol Model (EQSAM) was added to CAMx as an alternative to ISORROPIA for computing concentrations of the inorganic ions: sulfate, nitrate, ammonium, sodium and chloride. By design, EQSAM has more streamlined algorithms than ISORROPIA, which may lead to different model concentrations and faster run times. We apply CAMx using both thermodynamic equilibrium modules for simulations of the continental US with 12 km grid resolution for winter and summer months. Model predictions of inorganic ions by both algorithms are compared and model performance is evaluated against ambient data collected by speciated PM monitoring networks. The overall model run times are also compared. Our purpose is to introduce EQSAM as a newly available option in CAMx and provide information relevant to choosing between the two available aerosol thermodynamics modules in CAMx taking into consideration the scheme’s performance attributes as well as the requirements of each model application.

This study utilized an advanced sensitivity analysis, the higher order Decoupled Direct Method (HDDM-3D) as implemented in the Community Multiscale Air Quality Model (CMAQ) to quantify the impacts of aviation emissions during the landing and takeoff (LTO) cycle at nine individual airports; five located in regions of attainment of O3 and PM2.5 NAAQS: Boston Logan (BOS), Kansas City (MCI), Raleigh Durham (RDU), Seattle-Tacoma (SEA), and Tucson (TUS); and four located in regions of nonattainment: Chicago O’Hare (ORD), Hartsfield-Jackson Atlanta (ATL), John F. Kennedy (JFK), and Los Angeles (LAX). Fuel burn changes needed at the four nonattainment airports ranged from −14.9 (357,185 less tons) to −3.3 (55,715 less tons) times less fuel burned and from 1.6 (29,826 more tons) to 3.1 (79,584 more tons) times more fuel burned to reduce ambient PM2.5 by 0.1 μg/m3 and O3 by 1 ppb, respectively. Fuel burn changes needed at the five attainment airports ranged from 20.4 (39,516 more tons) to 48.0 (397,180 more tons) times more fuel burned and from −449.0 (−477,734 less tons) to −24.0 (46,648 less tons) times less fuel burned to increase ambient PM2.5 by 0.1 μg/m3 and O3 by 1 ppb, respectively. Using these estimates for a range of airports, we demonstrate an illustration of how HDDM-based sensitivity calculations can be used to develop source specific impacts on potential attainment designations for a region.

The GEM-MACH-Global model is a global online meteorology-chemistry system currently being developed at the Department of Environment and Climate Change Canada (ECCC). The model is an extension of the Department’s operational, regional GEM-MACH numerical weather and air quality prediction system. The objectives for its development are to improve our understanding of the long range transport and fate of criteria air contaminants, and to improve our forecasting system by providing chemical boundary conditions for the regional air quality forecast system, and background fields for global chemical data assimilation (O3 and NOy species). For this purpose, GEM-MACH-Global was recently updated with a comprehensive photolysis module (JVAL14-MESSy) and a detailed gas-phase chemistry module based on the SAPRC07C mechanism. Compared to its original ADOM2 chemistry mechanism, the revised gas-phase chemistry is more explicit, with new species and ~15 additional reactions important in the upper troposphere and lower stratosphere (UTLS) region. Furthermore, a lightning emission module was incorporated to represent NOx emissions aloft. These changes were evaluated with a 2010 annual simulation on a 400 × 200 global-grid. The simulation included inputs from 2010 HTAP global anthropogenic emissions, GFEDv3 biomass burning emissions and ECCC’s operational meteorological analyses. The presentation will describe the current state-of-science development of GEM-MACH-Global and show comparisons results of the annual simulation.

The objective of this study is to develop the Hamilton Airshed Modelling System (HAMS) providing a platform to better understand the processes and contributions to Hamilton’s air quality, informing future policy and human health impact decisions. Air quality in an urban airshed is influenced by local, regional and transboundary sources, geography and meteorology. HAMS must handle different emission sources, and the transportation and dispersion of emissions, to achieve realistic simulations of local impacts on air quality. HAMS relies on the development of two key data sets. The meteorology dataset impacts the transportation, transformation, dispersion, and deposition of pollutants over the challenging terrain in the Hamilton area. The emissions dataset represents the local and regional sources and contaminants influencing the air quality with contributions from industrial, commercial, residential, biogenic, and transportation sources. The Community Multi-scale Air Quality (CMAQ) model combines these datasets in a nested a one-way grid formation from regional (36 km) to local (1.33 km) scales and validates modeled output against observations. Recent local studies indicate that mobile and industrial sources are the primary emission sources. This study further examines the source contributions of mobile, industrial and background sources to local impacts on air quality.

To assess the impact of cities and urban surfaces on climate, the modeling approach is often used with inclusion of urban parameterization in land-surface interactions. This is especially important when going to higher resolution, which is common trend both in operational weather prediction and regional climate modelling. Inclusion of urban related effects can differ significantly across the land-surface models and urban canopy parameterizations. For adaptation and mitigation measures applied in big cities, especially in connection to climate change perspective, it is important to assess this uncertainty as well as to analyse the effects, which can affect air quality situation within the cities. These are main tasks of the new project URBI PRAGENSI. We performed experiments to assess urban effects on climate over central Europe for the decade 2001–2010, using two regional climate models (RegCM4 and WRF) in 10 km resolution driven by ERA-Interim reanalyses, three surface schemes (BATS and CLM4.5 for RegCM4 and Noah for WRF) and five urban canopy parameterizations available: one bulk urban scheme, three single layer and a multilayer urban scheme.

Megacities are now a common phenomenon in many regions around the world and present major challenges for the global environment. The concentrations of people and their activities have resulted in higher demand for energy and consumption of fossil fuels, leading to air pollution that affects public health and visibility, causes regional haze and acid deposition, and alters the earth’s climate. Recent advances in real-time pollutant measurement technologies and improved air quality models are allowing scientists to better understand the emission sources of pollutants and the complex atmospheric processes leading to severe air pollution, and providing policy makers the tools for designing cost effective mitigation strategies. This study addresses the effects of megacities and urban complexes on the Earth’s atmosphere, using Mexico City and Santiago as examples of cities that have been actively managing their air quality. Both cities demonstrate the types of environmental problems experienced by many urban centers and confront similar challenges to solving them. With appropriate planning, dedicated scientific research, robust emissions control policies, and effective access to advanced technologies and financial support, these urban centers also have the opportunity to manage the growing population sustainably while reducing atmospheric pollution and its impacts.

During the last 15 years, the official assessment maps in Flanders have improved from a map consisting of only measured data points to a high resolution assessment which covers the complete area and takes into account several types of sources and street canyons. In order to improve this level of detail, multiple steps were taken. First of all, a land use regression model was introduced at an hourly scale at 4 × 4 km2 resolution. Secondly, a Gaussian model was added for both point and line sources, correcting for emission double counting. Finally, a street canyon model was added to the chain, leading to improved resolution in these street canyons. In this work, we will discuss the problems encountered in these years such as how to account for double counting of emissions, how to correct the locations of the simplified road network and how to determine when street canyon calculations must be performed and how we solved them.

This work focus on the assessment of green infrastructures benefits on air quality in Porto urban area, applying the CFD model VADIS to a particular area within the city. Three scenarios have been considered: (i) the baseline refers to the current morphological characteristics of the area; (ii) a green scenario comprises the replacement of built-up areas by green areas; and (iii) a green scenario corresponding to the implementation of green roofs. The results of baseline simulations shows a good agreement with local measurements with a NMSE of 0.4, 0.6 and 2.1 for CO, NO$$_2$$ and PM10 concentrations, respectively. The benefits of green infrastructures on air quality are assessed for future medium-term climate scenarios (2041–2070), applying a cascade of numerical models, from global to urban scale, the WRF-VADIS modelling system. Future climate data point out a decrease in the number of days with moderate to strong wind speed, and an increase in the number of days recording low wind speed conditions. The assessment of green infrastructures effects on air quality under future climate focus on low wind speed conditions. The results clearly show the disturbances exerted by green infrastructures, which are positively or negatively affecting mainly the adapted areas and their close surroundings.

In the GEM-MACH URBAN project, Environment and Climate Change Canada (ECCC) high-resolution (2.5-km) Global Environment Multiscale-Modelling Air-quality and Chemistry (GEM-MACH) model and the Town Energy Balance Model (TEB) are being employed to examine the impact of the urban surface exchange scheme on the transport and diffusion of air pollutants in large cities such as Toronto, New York and Detroit. Simulation results show that while the TEB scheme causes O3 mixing ratios to increase, it leads to a decrease of CO and NOx mixing ratios and air quality health index (AQHI) values in the urban centers in both summer (July) and winter (January) months. The TEB scheme also has a big impact on the vertical diffusion coefficient, atmospheric boundary layer (ABL) height and air temperature. Comparisons against ECCC’s meteorological and air quality (AQ) observation networks suggest that the inclusion of TEB scheme improves the forecasts of both surface temperature and pollutant mixing ratios.

The dominant influence of shipping emissions on air quality of port cities and their surrounding urban areas has been demonstrated in a number of case studies, based on both assessment of monitoring data and the application of modelling tools. In the present work, chemical dispersion calculations are used for assessing the impact of marine traffic on air quality of Piraeus, Greece, and the neighbouring areas of the Athens urban agglomeration. A comprehensive marine emissions inventory is compiled using a bottom-up methodology on the basis of AIS traffic data, as well as activity and source parameter data obtained from other national databases, covering both merchant and passenger traffic in the area. A set of dispersion calculations are performed using the MEMO/MARS-aero chemical dispersion model for the Attica region, revealing the dominating contribution of the port area on the pollutant levels over the southern part of the city of Piraeus. Assessment of a hypothetical scenario, involving the implementation of cold ironing for ships at berth, indicates that emissions in the hoteling phase are significantly reduced, resulting to a notable reduction of concentrations in the port and the surrounding areas.

Large-Eddy Simulation (LES) models are so far barely applied to dispersion and chemical transformation of pollutants in urban air quality studies. Within the joint project MOSAIK (Modellbasierte Stadtplanung und Anwendung im Klimawandel/Model-based city planning and application in climate change) a new LES based state-of-the-art microscale urban climate model PALM-4U, has been developed. The new model includes both gas phase and aerosol chemistry. For practical applications, our approach is to go beyond the simulation of single street canyons to chemical transformation, advection and deposition of air pollutants in the larger urban canopy. First LES results of a test case for an urban quarter of Berlin (Germany) are presented.

Barcelona city (Spain) has a very high vehicle density (approx. 5500 vehicles km−2) being the majority diesel (64%). Barcelona traffic stations report chronic exceedances of nitrogen dioxide (NO2) European annual regulatory limits since the year 2000. In December 2017, a Low Emission Zone (LEZ) has been implemented in Barcelona to restrict access to the ring-road area to gasoline-powered passenger cars before 2000, diesel passenger cars before 2006 and vans before Euro 1 (registered before 1994) during air pollution episodes. This policy is planned to become permanent on December 2020. This work is an initial step towards evaluating the impact of Barcelona LEZ on air quality using CALIOPE-Urban. CALIOPE-Urban is a street-scale modelling system that couples CALIOPE air quality mesoscale modelling system, which provides air quality forecasts at 1 km horizontal resolution over Barcelona city, with R-LINE. Here we evaluate CALIOPE-Urban and assess its sensitivity to structural reductions of NOx emissions. We evaluate the coupled modeling system using observations from an experimental campaign in April to May 2013 in Barcelona city.

This study investigates the contributions of four major emission sources—industry, road traffic, shipping and residential heating—on air quality in the harbour city of Hamburg using a local-scale modelling system comprising meteorological, emissions and chemical transport models. Moreover, human exposure with regard to the overall air quality and the emissions sources under investigation was calculated. Based on detailed emission inventories and an evaluated CTM system, this study identifies road traffic as a major source of PM2.5 pollution and exposure during the entire year and in almost all populated areas in Hamburg. Overall, the highest contributor to PM2.5 concentrations is the industrial sector focussing on less populated areas.

Biogenic volatile organic compound (BVOC) emissions from terrestrial ecosystems drive distributions of several atmospheric constituents relevant to air quality and climate. BVOC emission rates can vary more than an order of magnitude over spatial scales of a few kilometers and time scales of less than a day which makes estimation of these emissions especially challenging. New improvements to the Model of Emission of Gases and Aerosols from Nature (MEGAN version 3) are described including (1) a transparent approach for assigning emission factors and other model parameters, (2) updated emission factors and algorithms based on recent measurements, and (3) treatments for previously unrepresented processes including stress induced emissions and canopy heterogeneity. The estimated emissions are compared to alternative model approaches and evaluated with aircraft measurements of concentrations and fluxes. Remaining gaps and priorities for future progress in biogenic organic emission modeling are also discussed.

Atmospheric chemistry transport models (CTMs) need spatially and temporally resolved emission data as input. Atmospheric concentrations of pollutants as well as their deposition depend not only on the emitted amount but also on place and time of the emissions used for the model calculations. Available emission inventories, both regional and global ones, typically provide annual emissions of specific substances on a predefined grid. Often, this grid is of coarser resolution than the model grid and the temporal resolution is not higher than monthly. In addition, many species like volatile organic compounds (VOCs) or particulate matter (PM) are only given as lumped sums and not split into their chemical components. This requires further processing of the emissions in order to produce sufficiently resolved data sets for follow-up CTM runs. As a consequence, emission models were developed for the purpose of creating “model-ready” emissions. They use methods that depend on the emission sector and the additional data available for the disaggregation of the inventory data, e.g. land use and population density data. Recently, new methods for specific sectors like agriculture, residential heating and traffic have been developed. The most commonly used global and regional emission inventories are summarized and an overview of currently applied methods to spatially and temporally disaggregate emission inventory data is given. Particular emphasis is laid on the temporal disaggregation by presenting methods that allow the creation of individual time profiles for each model grid cell.

Air quality modeling at high spatial resolution over large domains enables comprehensive health impact assessment. Spatially finely resolved emission inventories are a crucial component for reliable modeling. Spatialization of emissions from disperse emission sources (e.g. road transport) is performed using GIS-based spatial information, i.e. spatial proxies (e.g. road network and traffic volume data). For some important emission source sectors, however, it is challenging to define proxies that adequately represent the spatial distribution of emissions, and, for the lack of more representative information, population density is often used as a proxy. However, that is rarely a realistic representation and might distort the resulting assessments of their population exposure and health impacts. This study presents the impacts of the spatial allocation process and its improvements for machinery sector, by using an emission model at 250 m grid resolution in Finland. The corresponding influence on the modeled population exposure to PM2.5 is also presented. The improvements in the gridding procedures had a substantial impact on the modeled concentrations, especially in areas with denser population. For example, the emissions in Helsinki area from the machinery sector decreased by 41% due to the improvements. We conclude that it is necessary to use more realistic spatial proxies instead of the population density for evaluating the emissions originated from various emission categories.

Commercial shipping fleet and its emissions can be modeled in detail, but the emission from leisure craft are often invisible for activity based, bottom-up emission inventories. A new model (FMI-BEAM) describes the emissions from the leisure craft fleet in the Baltic Sea area, complementing the existing STEAM emission modeling suite. BEAM combines information from over 3000 boat marina locations, national small boat registries, Automatic Identification System data and boat survey results to derive leisure boat emissions for over 250,000 boats around the Baltic Sea coastline. The location of marinas and boat counts were determined from satellite images and other available data. With the BEAM leisure craft simulation model the spatial and temporal distribution of air emissions can be estimated. The presented results describe our first attempt to generate fuel consumption and emission inventory for small boats which have been underrepresented in current emission inventories. Small boat activity often occurs near the coastline in vicinity of populated areas and leisure craft emissions contribute to local air quality. The emissions of carbon monoxide and hydrocarbons are high compared to the emissions of commercial shipping, because very high emission levels are allowed for old small boat engines according to current legislation. The approach described in this paper can be applied in larger scale, for example to cover European coastlines in the future.

Assessment of air pollution health effects of road traffic should accurately characterize also the non-exhaust sources to support efficient emission mitigation. The use of the recently developed new-generation non-exhaust emission models opens new opportunities to extend and improve the traffic PM emission factors and the emission inventories. This study utilizes the non-exhaust emission model NOn-exhaust Road TRaffic Induced Particle emissions (NORTRIP) to evaluate the PM10 and PM2.5 emission factors separately for different climatic zones, road and street categories, and road maintenance practices in Finland. The non-exhaust emissions are influenced by climatic and weather factors, especially the road surface moisture, and have seasonal peaks particularly in winter and spring, due to enhanced formation of layers of street dust and their suspension to the air. The results demonstrated that changes in the selection of the types of winter tyres and their use, as well as road maintenance interventions could substantially reduce the non-exhaust emissions, especially in spring. Data gaps were identified in collecting the relevant inputs for the model. The methods presented in this study could be used at least in the whole Northern part of the world, i.e., the Nordic countries and Northern Central Europe, Northern Russia, North America, and parts of China and Japan.

Atmospheric carbon monoxide (CO) emissions estimated from inverse modeling analyses exhibit large uncertainties, due, in part, to discrepancies in the tropospheric chemistry in atmospheric models. We attempt to reduce the uncertainties in CO emission estimates by constraining the modeled abundance of ozone (O3), nitrogen dioxide (NO2), nitric acid (HNO3), and formaldehyde (HCHO), which are constituents that play a key role in tropospheric chemistry. Using the GEOS-Chem four-dimensional variational (4D-Var) data assimilation system, we estimate CO emissions by assimilating observations of CO from the Measurement of Pollution In the Troposphere (MOPITT) and the Infrared Atmospheric Sounding Interferometer (IASI), together with observations of O3 from the Optical Spectrograph and InfraRed Imager System (OSIRIS) and IASI, NO2 and HCHO from the Ozone Monitoring Instrument (OMI), and HNO3 from the Microwave Limb Sounder (MLS). Although our focus is on quantifying CO emission estimates, we also infer surface emissions of nitrogen oxides (NOx = NO + NO2) and isoprene. Our results reveal that this multiple species chemical data assimilation produces a chemical consistent state that effectively adjusts the CO–O3–OH coupling in the model. The O3-induced changes in OH are particularly large in the tropics. We show that the analysis results in a tropospheric chemical state that is better constrained. Our experiments also evaluate the inferred CO emission estimates from major anthropogenic, biomass burning and biogenic sources.

Experimental air-quality forecasts for the Canadian provinces of Alberta and Saskatchewan have been carried out since 2012, using a 10 km/2.5 km nested resolution version of Environment and Climate Change Canada’s Global Environmental Multiscale-Modelling Air-quality and Chemistry (GEM-MACH) on-line air-quality model. We describe here some of the main results of that work, and a major upgrade of this forecasting system (based on work carried out following a 2013 monitoring intensive field campaign in the Athabasca oil sands region of Canada). The new forecasting system has been designed in preparation for a follow-up field campaign, taking place during April and June of 2018.

The HiRes-2 system, operational since 2015, provides daily forecasts of potential prescribed fire impacts on air quality. The system uses the WRF and CMAQ models for meteorology and air quality computations. A decision tree model predicts prescribed fire activity based on the weather forecast by analyzing historical burning patterns under similar meteorological conditions. Prescribed fire emissions are estimated using land-based but satellite-enhanced fuel load maps, consumption forecasts, and emission factors derived from laboratory and field measurements. The DDM-3D feature of CMAQ is used to calculate the impacts of prescribed fire emissions on pollutant concentrations. The expanded new system, HiRes-X, has many extensions for dynamic air quality and human exposure management. Daily forecasts of air quality and prescribed fire impacts are disseminated through an online interactive tool, which uses webGIS technologies to display the forecasting products in real time and retrospectively, along with relevant earth observation datasets. These datasets can be overlapped with geographical and population information to visualize the air pollution impacts on sensitive places and vulnerable communities.

Since 2002, Environment and Climate Change Canada (ECCC) has been producing hourly surface analyses of pollutants covering North America which have been used in numerous health impact studies. The analyses are produced using an optimum interpolation scheme that combines the operational air quality forecast model GEM-MACH outputs (CHRONOS model outputs were used prior to 2009) with real-time hourly observations of O3, PM2.5, PM10, NO2, and SO2. We examine how passive observations, that are not used to create the analysis (i.e. cross-validation), can be used to evaluate the analysis error and to optimize the input error statistics.

Volcanic ash and other aerosols such as desert dust form significant hazards for aviation and can cause both direct safety threats and significant economic losses. However, forecasts of aviation hazards have generally been deterministic, although the available computational resources would easily allow for them to be ensemble-based. In principle, ensemble-based forecasts can enable more accurate error estimates and thus an improved risk management framework. Advanced data assimilation methods, such the Ensemble Kalman Filter, coupled with a meteorological forecast ensemble, provide increased accuracy and the possibility to estimate the source term by taking into account its correlation with the observed ash concentration.

There have been large advancement in modeling science and chemical constituent measurement of air pollutants harmful to the public health. The National Oceanic and Atmospheric Administration (NOAA) National Air Quality Forecasting Capability (NAQFC) is a vital service that NOAA provides to the general public to help safeguarding the public health as well as the environmental resilience through announcement of information-driver mitigation and adaptation action. NAQFC is poised to upgrade from using the Community Multiscale Air Quality Model (CMAQ) version 5.0.2 to version 5.2. This is noticeable a multiple sub-version number leaping forward corresponding to major upgrades in chemistry and emission sciences. The following lists the major science upgrade: (a) upgrade gas chemistry for the Carbon-Bond Mechanism version 5 (CB05) to version 5 Revision1 (CB05R1); (b) Inclusion of Halogen chemistry; (c) Employed more explicit speciation for isoprene and monoterpenes from biogenic sources; (d) Upgraded the aerosol module using a more sophisticated secondary aerosol production suite of multi-generational oxidation mechanism; and (e) Application of a fuller set of National Emission Inventory (NEI) from base year 2014. We tested the new system for a summer case retrospectively and compared its forecast performance with the real-time operational NAQFC. The U.S. Environmental Protection Agency (EPA) AIRNow monitoring network was used to verify the forecast accuracy. We noticed considerable discrepancies in the performance of the two realization of forecasting simulations.

Environment and Climate Change Canada’s ADAGIO project (Atmospheric Deposition Analysis Generated by Optimal Interpolation using Observations) produces maps of wet, dry and total annual deposition of oxidized and reduced nitrogen (N) and sulphur (S) and ozone in Canada and the United States by combining in an optimal way observed and modeled data. Optimal interpolation methods are used to provide the best objective analyses of seasonally-averaged surface concentrations of gaseous, particulate, and precipitation species predicted by Environment and Climate Change Canada’s in-line regional air quality model GEM-MACH (Global Environmental Multiscale model—Modelling Air quality and Chemistry) based on the difference between the modeled and measured values at network observation sites. The resulting objective analyses (OA) for gas and particulate species concentration fields are then combined with effective deposition velocities from GEM-MACH to calculate dry deposition. Concentrations of precipitation ions are combined with precipitation amounts from the Canadian Precipitation Analysis (CaPA), in which all available precipitation data sets are used to adjust precipitation amounts predicted by GEM, to calculate wet deposition. Results from the 2010 development year are compared with previously-generated wet deposition kriging maps, results from the USEPA’s Total Deposition (TDEP) method, and surface measurements not used in the analysis where available. It was found that the biggest sources of uncertainties are the dry deposition velocities and error statistics (weight matrix used to produce OA). Therefore, more work is needed to reduce these uncertainties.

North American air quality (AQ) forecasts made by the Environment and Climate Change Canada (ECCC) operational regional AQ prediction system since 2015 have used input emissions files based on Canadian, U.S., and Mexican national emissions inventories for base years 2010, 2011, and 1999, respectively. Since 2010, however, emissions of many criteria air pollutants have declined in both Canada and the U.S.. We recently tested new input emissions files based on a 2013 Canadian inventory, a projected 2017 U.S. inventory, and a 2008 Mexican inventory in the ECCC regional AQ prediction system. For Canada, the switch from the 2010 inventory to the 2013 inventory reduced SO2, NOx, and VOC annual anthropogenic emissions by 12%, 2%, and 4%, respectively. For the continental U.S., adoption of the projected 2017 inventory reduced SO2, NOx, and VOC annual anthropogenic emissions relative to the 2011 inventory by 65%, 33%, and 11%, respectively, suggesting the importance of emissions base-year representativeness for AQ forecasting. Moreover, the use of these new input emissions fields for 2016 and 2017 test periods improved AQ forecasts in comparison to the operational model for Canada and the U.S., in particular for summertime ozone forecasts over the eastern U.S.. A new version of the ECCC forecast system that uses these updated input emissions files was accepted for operational implementation in mid 2018.

In this study we present the influence of assimilation of ground PM2.5 measurements on forecasted concentrations of particulate matters for low air quality episode observed in the year 2017 over Poland. The episode was not reproduced by a standard forecasting system, based on the Weather Research and Forecasting with Chemistry model (WRF-Chem), working operationally without data assimilation. Here, we used Grid point Statistical Interpolation (GSI) system to assimilate ground observations from 42 stations measuring PM2.5 concentrations. The results show that the assimilation of PM2.5 concentrations has a positive impact on modelled concentration of PM2.5 and PM10. The greatest positive impact is noticed for the period with the high measured concentrations of pollutants. The results also show that for some stations the assimilation of PM2.5 and PM10 may lead to overestimation of concentrations at the warmer period characterised by lower overall PM concentrations. Further study with an application of degree-day factors for residential emissions and GSI assimilation is planned as the next step.

In this work we analyse the impact of meteorological data assimilation on the performance of the air quality forecasting system for the area of Poland (Central Europe). The forecasting system uses the WRF-Chem model, which is online integrated meteorology and air chemistry transport model. The forecasts are run each day for the next 48 h, using two nested domains of 12 km × 12 km (Europe) and 4 km × 4 km (Poland) and 35 vertical levels. In this work we analyse the period of 11–25 February 2017, during which poor air quality was observed at the beginning, followed by unusually warm days with low concentrations of pollutants. Two sets of forecasts are compared. In the first group, we use the forecasts with no data assimilation. Secondly, we use the community Gridpoint Statistical Interpolation system (GSI) to assimilate surface and radiosonde meteorological data. Both sets of forecasts are compared with hourly measurements of PM10 and PM2.5 for Poland. Assimilation of meteorological data overall improves the air quality forecasts, but not always leads to better representation of high-concentration episode.

An analysis of pollution homogeneity has been conducted to distinguish between remote and local pollution which contributes to month to month changes in aerosol optical depth (AOD) over the Taiwan area. This was carried out using both AERONET measurements at six monitoring sites in Taiwan and NASA MERRA-2 reanalysis, over the 15-year period from 2002 to 2017. As a measure of air pollution homogeneity we used the AOD standard deviation: the more homogeneous the spatial distribution of air pollution, the lower the AOD standard deviation is. Using this approach, we found that, over Taiwan, in autumn, inhomogeneous local air pollution is predominant, while, in spring, homogeneous remote air pollution from south-east Asia dominates. In autumn, when inhomogeneous aerosols from local sources dominate, the AOD standard deviation is essentially higher than that in spring, when homogeneous aerosols from remote sources dominate. Our approach allowed us to distinguish between homogeneous remote and inhomogeneous local sulfate air pollution of similar optical properties and chemical composition.

Since the last ITM in October 2016, the Canadian operational Regional Air Quality Deterministic Prediction System (RAQDPS) has been ported to a new high-performance computing system and has been updated to use a new meteorological initialization method, a new meteorological “piloting” model, a new and faster version of the GEM-MACH code, and a new set of input emissions files and to produce an expanded set of output fields. These updates are briefly described and some examples are given of their impact on RAQDPS forecast performance, including improved NO2 forecasts and a large reduction (~10 ppbv) in summertime ozone overpredictions for the eastern United States.

A new MSC initiative, named PROGNOS, aims to provide a more versatile, modular and innovative weather and air quality post-processing system to replace the current operational system (UMOS). PROGNOS has extensible statistical modeling capabilities. Currently in development, it issues real-time experimental air quality and temperature forecasts for cities across Canada and will eventually be applied to other meteorological fields and numerical models. The batch updates of the statistical models occur weekly using parallel processing in a cluster computing environment. Less flexible but more computationally efficient, online updating methods are also being evaluated. Several statistical modeling approaches have been explored including multiple linear regression, random forest, and Kalman filter prototypes for air quality forecasts. Logging, parameterisation, diagnostic and visualization features are also being explored. Medium to long term milestones include integrating seasonal and other transitional schemes as well as gridded post-processing.

Hierarchical clustering (HC) analysis groups datasets into clusters based on their (dis)similarity, and can be used to assess air-quality monitoring networks representability. The methodology describe here is a new approach to designing optimized air-quality monitoring networks by combining Kolmogorov-Zurbenko filtering (KZ) and HC of observed and modelled time series. Here we present the optimization of the air quality network in the province of Alberta, Canada, for NO2, SO2, PM2.5 and O3. The study suggests that network optimization will vary depending on chemical species due to different emissions sources and/or the results of secondary chemistry. Making use of hourly and time-filtered time series allows identifying emission sources, with much of the signal identifying sources emissions residing in shorter time scales (hourly to daily) due to short-term variation of concentrations, and background concentrations can be identified by larger time scales (monthly or over). The methodology is also capable of generating maps of sub-regions within which a single station will represent the entire sub-region, to a given level of dissimilarity, when applied to gridded datasets such as chemical transport modelling output.

Participants in the Air Quality Model Evaluation International Initiative (AQMEII) have conducted three rounds of model evaluation and intercomparison activities over Europe and North America since 2010, resulting in a large dataset of modeled meteorology, air quality and deposition fields for 2006–2010 that is available to the community for ongoing research on model evaluation. This study presents a brief analysis of some of the deposition fields generated during these past phases of AQMEII, quantifying both model-to-model variability and the level of agreement with available wet deposition measurements. We also discuss potential future AQMEII work focused on evaluating deposition and using modeled deposition fields for applications such as producing maps of estimated total deposition.

Emission data are a key driver and a major source of uncertainty to atmospheric chemistry transport models. As part of the international model-intercomparison study AQMEII chemistry transport models (CTMs) with harmonized input data have been used to evaluate the impact of emission datasets on different species and compare it to the effect of meteorology and parametrization of the CTM.

The effect of different bromine and chlorine species on ozone is here evaluated using a System for Integrated modeLling of Atmospheric coMposition, SILAM. The new stratospheric chemistry module includes 15 chlorine and 9 bromine species together with their gas-phase and heterogeneous reactions with other species. Evaluation of ozone concentrations is obtained by using the steady-state, which is obtained after spin-up time of several decades. The amount of halogens is estimated by including bromine emissions from sea (e.g. sea-salt), from biomass burning and from terrestrial sources. For chlorine we used the emissions from the GEIA inventory, with additional yearly scaling of the different CFC emissions.

Nitrogen oxides (NOX: NO + NO2) play a critical role in controlling atmospheric chemistry, especially for the tropospheric ozone (O3) formation and distribution.

Many papers are published in which a dispersion model’s predictions are compared with field observations and/or with other models’ predictions. Standard model performance measures are used such as Fractional Bias (FB). Many times, subjective statements are made such as “The model has very small scatter” or “Model A is performing better than Model B”. About 30 years ago, we developed the BOOT model evaluation software, which has two main components: 1. Calculation of model performance measures such as FB; and 2. Calculation of confidence limits (e.g., 95%) on performance measures and on the difference in a performance measure between two models. Bootstrap or Jackknife resampling methods are employed. We briefly review the methodology in BOOT’s Component 2, which is seldom used by researchers. We present an example from a project where several urban puff models’ predictions are compared with JU2003 field data, and where assessments are carried out regarding whether, for example, it can be concluded, with 95% confidence, that the difference in FB for two models is not significantly different from zero.

Limited area models applied at higher resolution than global models and using datasets of higher resolution are generally expected to more accurately represent the spatiotemporal variability of key meteorological and climate parameters such as near surface temperature, pressure, wind speed and atmospheric composition. However, limited area models require boundary conditions and the accuracy of such datasets reflects on the accuracy of the mesoscale simulations of atmospheric composition, in particular of the longer-lived species. The effects of various resolution meteorological data and of different chemical boundary and initial conditions on the simulated concentrations of the chemical gases and aerosols in the Mediterranean have been here investigated. Two different simulations in three domains of progressively increasing horizontal resolution were performed for the year 2016 using the mesoscale Weather Research and Forecasting (WRF) meteorological model with the chemistry module (WRF-Chem) and an additional one with different chemical boundary and initial conditions. Meteorology from the Global Forecast System (GFS) at two different horizontal resolutions (1 × 1° and 0.25 × 0.25°) available from NOAA has been used in the model to investigate their impact on the simulated air pollutants. The higher resolution meteorological input improves the comparison of model results to observations. The results are found sensitive to the chemical boundary and initial conditions.

FLEXPART is a Lagrangian particle dispersion transport model which is originally designed for calculating the long-range and mesoscale dispersion of air pollutants from point sources. Through the years, these type of models have proven to be a very useful tool in an operational context for the protection of the population in case of accidents in a nuclear power plant. In the meantime, FLEXPART has evolved into a more comprehensive tool for atmospheric transport modelling and analysis, and it can be used for a wide range of applications. The model can be used in a forward or backward mode, making it possible to trace back the source pollution contribution of a certain pollutant. To perform the FLEXPART dispersion simulations under consideration, we will use meteorological data from the European Centre for Medium Range Forecasts (ECMWF), more specifically the new ERA5 10-member climate data reanalysis at a 63 km resolution. We will explore how good we can access the model uncertainty in an objective way by taking advantage of ensemble weather forecasts.

Detailed measurements and dispersion modeling were conducted to develop more accurate integrated metrics to assess exposure to potentially high pollutant levels of primary traffic emissions. A 13-week intensive sampling campaign was conducted at six monitoring sites surrounding one of the busiest highway segment in the US with the study area focusing on the Georgia Institute of Technology campus to capture the heterogeneity in pollutant concentrations related to primary traffic emissions. A dispersion model (RLINE) was used to develop spatial concentration fields at a fine-spatial resolution over the area of primary exposures. Initial RLINE results were highly biased, due either to errors in the emissions or the model. Analysis suggests that both may be important, depending upon species, though the largest errors were due to how the model represents near-source dispersion, especially when the wind aligns with the road segments. To correct for high near-road bias, the RLINE results were calibrated using measurement observations after the urban background was removed. Performing the calibration hourly also reduced the bias observed in the diurnal profile. Both the measurement observations and dispersion modeling results show that the highway has a substantial impact on primary traffic pollutant (particularly elemental carbon and carbon monoxide) concentrations and captures the prominent spatial gradients across the campus domain, though the gradients were highly species dependent. These improved concentration fields were used to enhance the characterization of pollutant spatial distribution around a traffic hotspot and to quantify personal exposure to primary traffic emissions.

The impacts of air pollution on the environment and health is a very actual topic in Hungary. By understanding the connections between the pollution that humanity produces and the meteorological situations, we can make the changes that are necessary to prevent smog episodes and to pass on a more sustainable world to future generations. Primarily during winter and fall seasons, episodes of poor air quality related to high concentrations of particulate matter are frequent, especially in the eastern part of Hungary. These situations are often connected to special meteorological conditions—such as cold air cushion—which do not help the mixing and dilution of air pollutants. Usually this type of meteorological condition is coupled with very low ambient air temperature, which can urge the usage of more solid fuel (wood and coal), therefore result in an increase of PM10 emission from domestic heating. Such a situation developed at the end of January 2017 in the Sajó Valley, Hungary. The aim of the present study is to examine the meteorological background of the extreme high PM10 concentrations which could contribute to the developing of the smog situation, and to estimate the growth rate of the emission from residential combustion due to the low ambient air temperature.

AERMOD and CALPUFF are two Ontario approved regulatory air dispersion models used to assess air quality compliance to provincial standards. Modelled results from these dispersion models may not always be representative of actual concentrations due to their inherent assumptions and atmospheric simplifications. This research aims to assess the near field performance of these two models by comparing modelled concentrations predictions with monitored observations. The case study analyzed was for 24 h average nickel (Ni) from a Facility located in Northern Ontario. The Facility has a monitoring program set up to measure dust, from which the monitored metal concentrations were speciated. Statistical analysis of the modelled results demonstrated that CALPUFF 24 h average modelled concentrations showed better agreement with monitored results than those modelled using AERMOD, however, even these CALPUFF results were often 2x (or 0.5x) the monitored value.

The aim of this paper is to present a new model of arrival time for gas clouds. To create such a model, simulations of short-term gas leakages were conducted in a wind tunnel with a neutrally stratified boundary layer. Into the tunnel, a model of an idealized urban canopy in scale 1:400 was placed. For simulations of the short-term gas discharges, ethane was utilized. Concentration time series were measured by a fast flame ionisation detector. The experiments were repeated about 400 times to get statistically representative datasets. The ensembles of concentration time series were measured at about 50 individual positions. From these data, puff arrival times were computed. The results showed that a suitable probability distribution to describe the variability in values at individual positions for arrival time is lognormal. Moreover, the parameters of this distribution do not change randomly with the change in the measurement position but their change can be described by functions. Utilizing them, probability density functions of arrival time can be constructed and whatever quantile of arrival time at a chosen position can be computed. Such a model could help emergency services to estimate how the situation could look like during the accident not only in the most frequently occurred but also in the extreme cases.

The 2D-VBS framework describing the organic aerosol (OA) distribution as a function of its volatility and O:C is used in PMCAMx-Trj, a one–dimensional Lagrangian chemical transport model. The model is used to simulate the atmospheric OA during two PEGASOS campaigns in the Po Valley in Italy during 2012 and in Hyytiälä, Finland during 2013. Po Valley is an area with many industrial and agricultural sources and Hyytiälä is characterized by high biogenic secondary OA (SOA) levels. The simulations are evaluated with AMS measurements both at the ground and aloft with a Zeppelin airship. There were seven aging schemes with different assumptions about functionalization, biogenic SOA aging, and fragmentation that succeeded in reproducing the AMS measurements within measurement error. For all seven schemes, the assumed enthalpy of evaporation had a surprisingly small effect on the diurnal and vertical profiles of O:C and OA concentration. Even if the seven schemes have relatively different characteristics and assumptions, all provide a rather similar picture about the different sources and processes contributing to the total OA in these two very different areas.

During night-time, the heterogeneous hydrolysis of N2O5 on the surface of deliquescent aerosol particles represents a major source for the formation of HNO3 and leads to an important reduction of NOx in the atmosphere. In Chen et al., Atmos. Chem. Phys. 18:673–689, 2018 [5], we investigate an improved parameterization of the heterogeneous N2O5 hydrolysis. This approach is based on laboratory experiments and takes into account the temperature, relative humidity, aerosol particle composition as well as the surface area concentration. The parametrization was implemented in the online coupled model system COSMO-MUSCAT (Consortium for Small-scale Modelling and Multi-Scale Chemistry Aerosol Transport, https://cosmo-muscat.tropos.de ). In Chen et al., Atmos. Chem. Phys. 18:673–689, 2018 [5], the modified model was applied for the simulation of the HOPE-Melpitz campaign (10–25 September 2013) where especially the nitrate prediction over western and central Europe was analysed. The modelled particulate nitrate concentrations were compared with filter measurements over Germany. In this first study, the particulate nitrate results are significantly improved by using the developed N2O5 parametrization, particularly if the particulate nitrate was dominated by the local chemical formation (September 12, 17–18 and 25). The aim of the current study consists in an evaluation over a longer time period for different meteorological conditions and emission situations. For this reason, we have simulated the period from March to November 2010. The results were compared with other approaches and evaluated by filter measurements. The improvement was confirmed for the results in spring and autumn, but nitrate is strongly over-predicted also for the new parametrization during the summer time.

Chemical transport model (CTM) simulation of organic aerosol (OA) is always challenged by numerous sources and complicated formation processes of secondary organic aerosol. In this study, we conducted a source-specific, whole-year (2011) simulation of organic aerosol in Europe using the air quality model CAMx v6.3 with volatility basis set (VBS) scheme after implementing new findings from experimental studies. The VBS module was parameterized based on the latest data for gasoline and diesel vehicles and wood combustion from the smog-chamber experiments. The model performance was evaluated using OA measurements from the ACSM (Aerodyne Chemical Speciation Monitor) and AMS (Aerodyne Aerosol Mass Spectrometer) network and contributions from 6 different anthropogenic (gasoline and diesel vehicles with old or new technologies, biomass burning, and other sources (OP)) and biogenic sources were estimated. The modified VBS scheme improved the model performance on OA simulation during the whole period by reducing the bias between model and measurements by up to 52%. The OA concentrations were dominated by biomass burning in winter, while biogenic emissions were the main sources in summer. The contribution of road traffic was relatively lower compared to studies in the USA. The contribution of new gasoline and diesel vehicles (after Euro IV emission standards or equipped with diesel particle filters) to the total OA was negligible.

The air-water exchange processes are considered as complicated but very important for the boundary layer structure. The complex physical procedures related to momentum, moisture and thermal exchanges, render their numerical description into a quite challenging task. Ocean surface roughness and spray droplets have an impact on the lower atmospheric layers with profound effects. These involve alterations in the atmospheric stability profiles and microphysical processes such as the formation of sea salt particles. Water vapor produced by sea spray can alter humidity and temperature profiles close to the ocean surface, leading to modified stability conditions. This cascade of effects can be further extended to wind profiles. Sea salt particles derived from water droplet evaporation are a major source of CCN and GCCN that may lead in low cloud formation over the sea and/or the coastal boundary layer. To describe these processes numerically, an atmospheric-wave-spray coupled system is introduced. The RAMS/ICLAMS atmospheric and the WAM wave spectral model are interfaced, including schemes for ocean drag, water droplet thermodynamics and salt particles considered as predictive quantity. Model simulations performed in the Atlantic shoreline showed that droplet evaporation near the surface modifies the atmospheric stability and affects sea salt dispersion.

Aerosol-cloud interactions constitute a major contributor of uncertainty in projections of anthropogenic climate change. The fraction of aerosol that activates to form cloud droplets (cloud condensation nuclei, CCN) is at the heart of aerosol cloud interactions. Towards this, we investigate the role of organic mass in the formation and evolution of CCN using the global 3-dimensional chemistry transport model TM4-ECPL coupled with the M7 aerosol microphysics module. The contribution of organics to the CCN levels is quantified by comparing the global surface distribution of aerosol particles and CCN computed with and without organic aerosol mass considerations, to the surface CCN observations. We also calculate the dynamical behavior of the CCN by computing their persistence times in atmosphere, i.e. the period over which the CCN concentrations show autocorrelation. It is found that organic species in aerosol modulate CCN concentrations by 50–90%—with a higher influence over land; furthermore, simulations compare better with observations when the impact of organics on CCN levels is taken into account.

Microorganisms are present in the atmosphere and can survive in cloud droplets. They are able to transform organic compounds and can consequently compete with radical chemistry. Because the cloud system is a complex multiphase medium, the efficiency of biodegradation by cloud microorganisms has to be evaluated by numerical models of different complexity simulating multiphase chemical processes in clouds. However, only abiotic processes are taken into account in these numerical tools. The objective of this work was thus to integrate biological data in an atmospheric cloud chemistry model and to evaluate the effect of microorganisms in the transformation of chemical compounds. For this, experimental biodegradation rates of acetic and formic acids, formaldehyde and hydrogen peroxide by various bacterial representative of cloud biodiversity were experimentally measured. These values have been implemented in a cloud chemistry model describing aqueous phase chemistry with the explicit CLEPS (Cloud Explicit Physico-chemical Scheme) mechanism. Several simulations with and without biodegradation processes have been performed changing temperature (5 °C or 17 ℃) and actinic flux to simulate summer or winter conditions. The chemical scenario (gas concentrations and emission/deposition) is representative of low-NOx emission with significant isoprene emissions.

Using a fully coupled air quality prediction model, simulations were carried out to investigate the impact of aerosol indirect effect on air pollution-meteorology interaction in an urban environment. We found that the aerosol indirect effect results in an increase in cloud droplet number concentration, a reduction in cloud droplet size, and an increase in cloud water. While, as a result, precipitation production is suppressed in low-level clouds, we found that, in a case of deep convective clouds, there is an enhancement of cloud ice and precipitation production at higher levels due to the increase in abundance of smaller drops being carried up in the updraft. There is also an indication of enhanced convective activity due to urban heating.

The NMMB-MONARCH is a fully online integrated meteorology-chemistry model for short- and mid-term Chemical Weather Forecasts (CWF) developed at the Barcelona Supercomputing Center (Earth Sciences). The meteorological core is the Nonhydrostatic Multiscale Model on the B-grid (NMMB). The global aerosol module includes five natural and anthropogenic species: mineral dust, sea salt, organic matter, black carbon and sulfate. The full online coupling between the aerosol module and the model radiation scheme (RRTMG) has been recently implemented. We present in this work a first evaluation of the aerosol intensive optical properties required by the coupling mechanism: single scattering albedo (ω) and asymmetry factor (g). New aerosol refractive indexes from recent literature have also been tested. We found that the model performance improves when considering a less absorbing mineral dust. Moreover, in anthropogenic areas a purely scattering organic matter and a higher real index for sulfate partially reduce the model systematic biases.

Aerosol acidity plays an important role by affecting aerosol formation, concentration, human health, and nutrient bioavailability. This study used thermodynamic model (ISORROPIA-II) and measurement data in Atlanta, GA, USA from 2009 to 2013. Aerosol pH in this research domain kept low (average 2.61) and did not change much over the past decade with the SO2 emissions reduction. But the trend still shows significant seasonal patterns (high in the winter and low in the summer) and spatial variation (higher in urban and lower in rural area). Model performance were evaluated by comparing estimated and observed NH3 gas phase partitioning. Also based on the source apportionment results for the same domain, multiple-linear-regression models were developed to show the source impacts on fine particle pH. With strong correlations (average R2 = 0.52), in most cases, models indicate vehicles have the largest positive impact on pH and ammonia bisulfate contribute most negative impact on pH, even though specific results are different for different seasons and locations.

An extreme dust event occurred over southern Israel, where meteorological and air pollution observations suggested that the front was about 100 km wide and the extreme dust concentration formed an advancing “wall” of more than 1 km high. In this study, the atmospheric model RAMS was employed with nested grids over the event domain, where strong turbulence was demonstrated by the model. The simulations showed some turbulent features which supported the observations, yet weren’t sufficient to explain the full development of the dust front.

Uncertainties in climate simulations can strongly propagate to estimates of climate change impacts, including its effects on air pollution. Here we use a coupled modeling framework to evaluate the role of climate model response in projections of climate-induced impacts on air quality. Within integrated economic, climate, and air pollution projections, climate model response is altered by modifying the climate sensitivity of the framework’s Earth system component. We find that variations in climate sensitivity ranging from 2.0 to 4.5 °C per doubling of CO2 can change projections of the climate penalty on O3 and PM2.5 pollution in the U.S. by more than 2 ppb and 0.5 µg m−3. The impact of uncertainty due to climate model response can be as important as that related to greenhouse gas emissions scenario or natural variability.

Tropospheric ozone (O3) can damage vegetation, affecting productivity and quality of the crops. Vines, in particular, have an intermediate sensitivity to ozone. Moreover, an increase of ozone levels is foreseen under climate change scenarios. The Douro Demarcated Region is one of the most productive wine areas in Portugal; thus studying the ozone deposition over this region and assessing its potential effects is a nowadays concern. This work aims to evaluate the risk of Douro vineyards exposure to ozone in present and future climates. The chemical transport model CHIMERE, with a spatial resolution of 1 km2, fed by meteorological data from the WRF model, was applied for the years 2003–2005 (present climate), for 2049 and 2064 (mid-term future) and for 2096 and 2097 (long-term future). The assessment of the potential damage in terms of productivity and quality was done through the analysis of ozone deposition and the application of concentration-response functions. The exposure indicator AOT40 (accumulated concentration of ozone above 40 ppb) for the period established in the Air Quality Framework Directive 2008/50/CE was also estimated. The model results show, for present and future climate, that the AOT40 levels in the entire Douro region are above the target value for the protection of vegetation. The results of the exposure-response functions suggest that the tropospheric ozone levels in the future, in the region, would influence the quality and productivity of the wine.

Extreme ozone pollution episodes can cause acute stress on the human respiratory system and damage vegetation. This study leverages long-term ozone measurement records from the United States Environmental Protection Agency Air Quality System (AQS) and the Environment and Climate Change Canada National Air Pollution Surveillance Program (NAPS) to identify ozone pollution episodes in summertime North America and assess meteorological patterns typical for those events. For this purpose, methods are adapted from studies of heat waves that similarly rely on station-based measurement records. Clustering methods are used to group ozone monitoring stations into large regions according to their likelihood of recording simultaneous ozone extremes. Multiday periods with abnormally high cluster-wide ozone concentrations are then identified as ozone episodes. Composite meteorological patterns associated with the ozone enhancement episodes are separately derived for each region. The composite patterns exhibit features commonly associated with elevated ozone, such as high temperatures and reduced cloud cover. In most regions, these features appear alongside anomalous synoptic-scale anticyclonic circulation in the mid-troposphere. These systems are themselves embedded in large-scale wave trains extending from the Pacific. The wave trains develop more than a week in advance of ozone episode onset and are plausibly related to sea surface temperature patterns with subseasonal persistence that also emerge in the composite meteorology. The persistence of these ozone episode circulation patterns could potentially be leveraged to improve forecasting and long-term projections of air quality.

Model results in this study show that the modelled maximum ground level concentrations could vary significantly with the choice of meteorological data periods and source configurations, mainly due to changes in the climatology of wind speeds, their distribution and temperature. Modelled maximum ground-level concentrations could vary by as much as 30% during the historical period (1996–2016), and could decrease by over 50% for most low-level sources in the future (2051–2055) with projected climate change.

The impact of air pollution on premature mortality in Europe and the United States (U.S.) in 2010 is modelled by a multi-model ensemble of regional models in the framework of the third phase of the Air Quality Model Evaluation International Initiative (AQMEII3). Introducing 20% emission reductions both globally and regionally in Europe, North America and East Asia were performed in order to calculate the domestic and foreign contributions to air quality and related premature mortality. Total premature mortality was estimated to be 414 000 in Europe and 160 000 in the U.S., using multi-model mean pollutant concentrations. The number of premature mortality cases calculated using concentration inputs from different air quality models can vary by up to a factor of three. Results show that the domestic emissions have the largest impacts on premature death, while foreign sources are a minor contributor to adverse impacts of air pollution.

We used linked and coupled physical models to identify relationships among environmental variables across multiple sources and pathways to examine the impact of nitrogen loadings on chlorophyll α concentrations.

Commonly, population exposure is evaluated by crossing data of atmospheric pollution and population density maps. The former are usually actual measurements or simulated concentrations; depending on the approach or the model resolution, very different patterns may appear both in space and time, so that conclusions can vary significantly. The latter are usually based on residency information, and for many of us, do not reflect the typical wanderings, thus actual exposure. With the rise of portable devices, we are given the unprecedented opportunity to measure pollutant concentrations at a high time rate and to know the exact location of a subject. Moreover, the increase of computational capacities allows one to perform operational runs at spatial and temporal resolutions of about 10 m and 1 h respectively. Furthermore, if the subject writes an activity log, it is also possible to discriminate indoor and outdoor situations. In this ongoing work, we investigate the discrepancy in the evaluation of population exposure when using, on one hand different pollutant concentration maps e.g. yearly, daily or hourly average values, more or less sophisticated and/or refined models, different information related to the population e.g. static or dynamic and on the other hand actual data. Our region of interest for this test case is the city of Liège in Belgium.

Poor air quality results in important health effects. However, the understanding of the problem by the public at large and by the policy makers is sometimes severely lacking. Therefore, it could be important to boil down the knowledge on air quality to some main points which then can be communicated to the stakeholders. The paper present such a list and is based on interviews and discussions with several air quality experts in the field. For the public at large, the main message is twofold. First of all, citizens have to acknowledge that every action they take has an influence on the quality of the air they breathe. Secondly, they have to understand the impact of air pollution on their own health and their neighbors, without blindly relying on rules like ‘natural/green is good for air quality’, which are often wrong. For the policy makers, a more heterogeneous set emerged from the discussions between the experts. First of all, policy makers have to know how much they (on their government level) can influence the air quality, and what effect certain actions can have on the population for which they are responsible. Secondly, they need to understand the uncertainties on the numbers as they exist now. Finally, stakeholders are encouraged to take action, from the local scale on, in order to get the actions at other levels moving (act local, think global).

Dr. Yitzhak Mahrer, an Israeli atmospheric scientist, was one of the earliest contributors to the Regional Atmospheric Modeling System (RAMS), a leading model with abilities on a wide range of atmospheric scales. He was involved in many complex-terrain and coastal atmospheric dynamic studies, and was among the pioneers of air-pollution modeling, especially over the Eastern Mediterranean. Dr. Mahrer deceased on September 2017, and RAMS community has lost one of its founders, with his shy smile, funny remarks, and bright mind. While he also led multiple fields observational campaigns and graduated many students as a Professor at the Hebrew University of Jerusalem, we hereby present only a brief overview of his scientific contribution to the atmospheric modeling community.