AS3.19

The COVID-19 pandemic struck China in January 2020 and the rest of the world from February 2020 onwards. Public authorities enforced several kinds of lockdowns in order to limit the spread of the pandemic and reduce its impact on the health system: at the height of the first wave of the pandemic, more than one human in two was subjected to a lockdown, with associated disruption in local and international travel, industry, tourism etc. These lockdowns had a profound effect on anthropogenic emissions of aerosol, trace gases and greenhouse gases; in this work we focus on aerosols and a selection of trace gases.

The Integrated Forecasting System (IFS) of ECMWF is core of the Copernicus Atmosphere Monitoring Service (CAMS) to provide global analyses and forecasts of atmospheric composition, including reactive gases, as well as aerosol and greenhouse gases. In this work, we use two emission reduction scenario with an experimental version of the IFS in its CAMS configuration: a global and a European one.  Global simulations of aerosols were carried out with these two scenarii and compared to a reference simulation without any COVID-19 impact, and to worldwide observations of PM2.5, AOD and trace gases.

The simulated PM2.5 using the global emission reduction scenario were found to reproduce quite accurately the observed evolution over China, India and United States. Over Europe, the simulated PM2.5 using the European reduction scenario were closer to observations and appeared more realistic. India was the only place where a significant impact on AOD and on temperature and radiation from the COVID-19 lockdowns was simulated. These simulations also provided information on how the aerosol speciation was altered by the COVID-19 lockdowns: over Europe and the U.S., most of the decrease in surface aerosols was simulated to come from nitrate aerosols. Over the U.S., this matched well with observations of speciated aerosols at surface.

How to cite: Pelletier, S., Rémy, S., Kipling, Z., Vilardell, M. G., Bouarar, I., Engelen, R., Flemming, J., Huijnen, V., and Meziane, M.: Simulation of the impact of COVID-19 lockdowns on aerosols and radiation at a global and European scale in CAMS, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-16450, https://doi.org/10.5194/egusphere-egu21-16450, 2021.

To hinder the circulation of the COVID-19 virus, European governments implemented emergency measures going from light social distancing to strict lockdowns, depending on the country. As a consequence, many industries, businesses and transport networks were forced to either close down or drastically reduce their activity, which resulted in an unprecedented drop of anthropogenic emissions. This work presents the Copernicus Atmosphere Monitoring Service (CAMS) European regional emission reduction factors associated to the COVID-19 mobility restrictions (CAMS-REG_ERF-COVID19), an open source dataset of daily-, sector-, pollutant- and country-dependent emission reduction factors for Europe linked to the COVID-19 pandemic. The resulting dataset covers a total of six emission sectors, including: road transport, energy industry, manufacturing industry, residential and commercial combustion, aviation and shipping. The time period covered by the dataset includes the first and second waves of the disease ocurred during 2020, starting from 21 February, when the first European localised lockdown was implemented in the region of Lombardy (Italy), until 31 December, when COVID-19 transmission remained widespread and several countries had nationwide restrictions still in place. The CAMS-REG_ERF-COVID19 dataset is based on a wide range of information sources and approaches, including open access and measured activity data and meteorological data, as well as the use of machine learning techniques. We combined the computed emission reduction factors with the Copernicus CAMS European gridded emission inventory to spatially (0.1x0.05 degrees) and temporally (daily) quantify reductions in 2020 primary emissions from both criteria pollutants (NO_x, SO₂, NMVOC, NH₃, CO, PM₁₀ and PM_2.5) and greenhouse gases (CO₂ fossil fuel, CO₂ biofuel and CH₄), as well as to assess the contribution of each pollutant sector and country to the overall reductions. The resulting gridded and time-resolved emission reductions suggest an heterogeneous impact of the COVID-19 across pollutants, sectors and countries.

How to cite: Guevara, M., Jorba, O., Petetin, H., Denier Van Der Gon, H., Kuenen, J., Super, I., Peuch, V.-H., and Pérez García-Pando, C.: Quantification of the Emission Changes in Europe During 2020 Due to the COVID-19 Mobility Restrictions, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-5894, https://doi.org/10.5194/egusphere-egu21-5894, 2021.

The mobility restrictions implemented to slow down the transmission of the new coronavirus disease (COVID-19) drastically altered Spanish anthropogenic emissions in several sectors, leading to substantial impacts on air pollutant concentrations. In order to reliably quantify these changes, the confounding effects of meteorological variability need to be properly taken into account. We thus designed an innovative methodology relying on the use of machine learning (ML) models fed with ERA5 meteorological reanalysis data and other time features, to estimate more accurately the so-called business-as-usual (BAU) pollutant concentrations that would have been observed in the absence of lockdown (Petetin et al., 2020). The difference with concentrations actually observed during the lockdown give meteorology-normalized estimates of the AQ changes due to the altered anthropogenic emission forcing, independently from the meteorological variability. Importantly, our methodology includes a conservative estimation of the uncertainties, which allows to highlight statistically significant changes. This study focuses on NO2 and O3. We applied this analysis for a selection of urban background and traffic stations covering more than 50 Spanish provinces and islands. Validation results indicate that the method usually performs well for estimating BAU concentrations (mean absolute bias below +6%, root mean square error around 25-30% and correlation above 0.80).

The COVID-19-related lockdown has induced a strong reduction (-50% on average) of NO2 concentrations in Spanish urban areas, although with some spatial variability among the provinces. In largest cities, stronger reductions were found at traffic stations compared to urban background ones, reflecting the major impact of the lockdown on traffic emissions. Substantial discrepancies with changes obtained considering a climatological averaged NO2 concentrations were found, highlighting the interest of such ML-based weather-normalization method. Compared to NO2, the impact on O3 is lower and more heterogeneous. In many cities, O3 levels slightly increased (likely due to a reduced titration by NO), but these increments often remain within the (95% confidence level) uncertainties of our methodology. However, during the most stringent phase of the lockdown (beginning of April and the few following days), a clearer O3 increase is found, reaching the statistical significance in several Spanish cities (e.g. Albacete, Barcelona, Castellón, Mallorca, Murcia, Málaga).

These results are of strong interest for quantifying the corresponding health impacts of these AQ changes, especially for showing the potential trade-offs between health benefits induced by the reduction of NO2 and enhanced mortality due to higher O3.

Petetin, H., Bowdalo, D., Soret, A., Guevara, M., Jorba, O., Serradell, K., and Pérez García-Pando, C.: Meteorology-normalized impact of the COVID-19 lockdown upon NO₂ pollution in Spain, Atmos. Chem. Phys., 20, 11119–11141, https://doi.org/10.5194/acp-20-11119-2020, 2020.

How to cite: Petetin, H., Bowdalo, D., Achebak, H., Soret, A., Guevara, M., Jorba, O., Serradell, K., Quijal-Zamorano, M., Ballester, J., and Pérez García-Pando, C.: Meteorology-normalized impact of COVID-19 lockdown upon NO2 and O3 in Spain, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-14422, https://doi.org/10.5194/egusphere-egu21-14422, 2021.

Expediting urbanization has triggered an increase in cardiopulmonary diseases attributable to fine-particulate air pollution. Air Quality models simulate the dilution and dispersion of air pollutants that affect the atmosphere, contributing crucially to the comprehension of its processes. Air quality forecasts produced by the Copernicus Atmosphere Monitoring Service (CAMS) provide open access to accurate and reliable information but in a coarse resolution. Data-driven models can downscale the forecasts from deterministic air quality models on the basis of reliable measurements. Low-cost air quality sensors are widely known for their increased spatial coverage and economic operational costs, but usually, their reliability is in dispute. In this study, a dense network of calibrated PM2.5 measurements installed in the city of Patras is combined with CAMS forecasts and statistical approaches to generate 24h forecasts of PM_2.5 concentrations in an urban area of Greece. The implemented techniques are the analog ensemble (AnEn) and the Long Short-Term Memory (LSTM) network. Auxiliary variables of meteorological origin were also utilized. The required forecasts were retrieved from the European Center for Medium-Range Weather Forecasts (ECMWF), and were pin-pointed to the location of the air quality monitoring stations. The results showed that both methods had comparable performance, with low bias and relatively small errors. In the stations with high PM2.5 levels, AnEn performed better, whereas in the stations and seasons with moderate concentrations LSTM outperformed. A comprehensive validation is presented and discussed. AnEn and LSTM methods were proved reliable tools for air pollution forecasting and can be used for other regions with small modifications.

How to cite: Pappa, A. and Kioutsioukis, I.: High-resolution PM2.5 forecasting using CAMS predictions, low-cost sensors and ensemble techniques, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-15960, https://doi.org/10.5194/egusphere-egu21-15960, 2021.

In order to reach legal air quality limits, several municipalities in Germany have decided to take actions if concentrations of NO₂ and Particulate Matter (PM) exceed certain thresholds. The decision for concrete measures is usually based on observations or use the Direct Model Output (DMO) of air quality models. However, due to large biases of state-of-the-art numerical air quality models, the skill of DMO forecasts to predict periods of polluted air up to four days ahead is very limited.

The project LQ-WARN aims to develop a system for warning of poor air quality based on Model Output Statistics (MOS). Therefore, air quality related observations, model results provided by the Copernicus Atmosphere Monitoring Service (CAMS) and meteorological parameters from the ECMWF numerical weather prediction model are used as predictors to forecast the air quality by applying Multiple Linear Regression (MLR). In this way MOS equations are calculated for four seasons. The final forecast product will comprise post-processed probabilistic as well as deterministic (e.g. mass concentration) parameters for the species NO₂, O₃, PM₁₀ and PM_2.5. Forecasts will be available for several hundred German locations and cover lead times up to 96 hours.

Here, we show first results of our phase 1 MOS prototype, for which observational, meteorological and empirical predictors are applied. Despite of the preliminary exclusion of CAMS predictors, the verifications of the MOS equations imply a considerable reduction of variance and a significant reduction of RMSE (Root Mean Square Error) compared to the climatological values for all four species. Hence, the MOS system can already provide a reasonably good air quality forecast. Furthermore, our analysis of used meteorological predictors, enables a detailed analysis of the importance of specific meteorological parameters for improved statistical air quality forecasts.  As an outlook we will provide detailed information about the final phase 2 LQ-WARN product, which will also include the MOS predictors of CAMS and is expected to be launched in pre-operational mode by 2022.

How to cite: Robrecht, S., Lambert, A., and Gilge, S.: The LQ-WARN Project – Development of a Model Output Statistics Product for Air Quality Warnings, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-882, https://doi.org/10.5194/egusphere-egu21-882, 2021.

Regional atmospheric chemistry models are adopted for simulating concentrations of atmospheric components at high resolution and quantifying the impact of localized emissions (e.g. industrial and urban clusters) on the non-linear chemical processes, e.g. ozone production. However, their evaluation is challenging due to the limited availability of high spatiotemporally resolved reference datasets. For the same reason, the vertical distribution of pollutants simulated by the model is especially arduous to assess.

Here, we present regional atmospheric chemistry model studies with spatial resolution up to 2.2 × 2.2 km² focused around Germany for May 2018 using the MECO(n) model system. Using a network of surface concentration measurements at background, near traffic and industrial locations, we evaluate the spatial distribution of NO₂ simulated by the model. The highly resolved model together with a comparable resolution and up-to-date input emissions inventory, was found to perform best in reproducing the spatial distribution of NO₂ surface volume mixing ratios (VMRs). We propose a computationally efficient approach to account for the diurnal and day of the week variability of input anthropogenic emissions (e.g. from road transport), which proved to be crucial for resolving the temporal variability of NO₂ surface VMRs.

The simulated NO₂ tropospheric vertical column densities were evaluated by employing the measurements of a 4-azimuth MAX-DOAS instrument in Mainz. Generally, such comparisons do not account for the spatial sensitivity volume of the MAX-DOAS measurements, the change of sensitivity within this volume and the spatial heterogeneity of NO₂. We therefore apply a consistent approach of comparison of the differential slant column densities (dSCDs), which overcomes these limitations. Moreover, the dSCDs are obtained for several elevation and azimuth angles, which are characterized by distinctive sensitivity for different vertical levels within the boundary layer and different horizontal representativeness. Hence, also an evaluation of the model in simulating the vertical distribution of NO₂ can be performed with this approach using continuous MAX-DOAS measurements spanning long time periods. We found that the model performs well with respect to the measured dSCDs at low elevation angles (< 8°) with an overall bias between +14 and -9%, and Pearson correlation coefficients between 0.5 and 0.8 for the different azimuth viewing directions.

How to cite: Kumar, V., Remmers, J., Beirle, S., Kerkweg, A., Lelieveld, J., Mertens, M., Pozzer, A., Steil, B., and Wagner, T.: Evaluation of a high resolution regional atmospheric chemistry model using MAX-DOAS and in situ NO2 measurements, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-4989, https://doi.org/10.5194/egusphere-egu21-4989, 2021.

Reducing air pollution, which is the world's largest single environmental health risk, demands better-informed air quality policies. Consequently, multi-scale air quality models are being developed with the goal to resolve cities. One of the major challenges in such model systems is to accurately represent all large- and regional-scale processes that may critically determine the background concentration levels over a given city. This is particularly true for longer-lived species such as aerosols, for which background levels often dominate the concentration levels, even within the city. Furthermore, the heterogeneous local emissions, and complex dispersion in the city have to be considered carefully.

In this study, the impact of processes across a wide range of scales on background concentrations over Switzerland and the city of Zurich was modelled by performing one year of nested European and Swiss national COSMO-ART simulations to obtain adequate boundary conditions for gas-phase chemical, aerosol and meteorological conditions for city-resolving simulations. The regional climate chemistry model COSMO-ART (Vogel et al. 2009) was used in a 1-way coupled mode. The outer, European, domain, which was driven by chemical boundary conditions from the global MOZART model, had a 6.6 km horizontal resolution and the inner, Swiss, domain one of 2.2 km. For the city scale, a catalogue of more than 1000 mesoscale flow patterns with 100 m resolution was created with the model GRAMM, based on a discrete set of atmospheric stabilities, wind speeds and directions, accounting for the influence of land-use and topography. Finally, the flow around buildings was solved with the CFD model GRAL forced at the boundaries by GRAMM. Subsequently, Lagrangian dispersion simulations for a set of air pollutants and emission sectors (traffic, industry, ...) based on extremely detailed building and emission data was performed in GRAL. The result of this nested procedure is a library of 3-dimensional air pollution maps representative of hourly situations in Zurich (Berchet et al. 2017). From these pre-computed situations, time-series and concentration maps can be obtained by selecting situations according to observed or modelled meteorological conditions.

The results were compared to measurements from air quality monitoring network stations. Modelled concentrations of NO_x and PM compared well to measurements across multiple locations, provided background conditions were considered carefully. The nested multi-scale modelling system COSMO-ART/GRAMM/GRAL can adequately reproduce local air quality and help understanding the relative contributions of local versus distant emissions, as well as fill the space between precise point measurements from monitoring sites. This information is useful for research, policy-making, and epidemiological studies particularly under the assumption that exceedingly high concentrations become more and more localised phenomenon in the future.

How to cite: Suter, I., Emmenegger, L., and Brunner, D.: NESTED HIGH-RESOLUTION NOx AND PM SIMULATIONS OVER ZÜRICH, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-6202, https://doi.org/10.5194/egusphere-egu21-6202, 2021.

High spatial resolution (<100m) mapping of NO₂ at various temporal scales (e.g., hours of the week, month, or year) provides opportunities to study the relationship between personal air pollution exposure and health over large populations. Statistical modelling of NO₂ at the global scale provides high-resolution estimations for countries with deficient ground station measurements and provides air pollution maps and human exposures with consistent uncertainties for global health studies. Our objective is to develop spatiotemporally-resolved statistical learning models, understand the temporal dynamics of NO₂ and the contributing sources, and open-source our global NO₂ prediction maps at 100 m resolution. The global maps are provided at various temporal aggregations (e.g. separating between weekdays and weekends, day and night) and spatial aggregations (e.g. multiple gridded resolutions, administrative units) to facilitate global exposure assessment. To create these maps, we compiled from multiple sources a dataset of hourly NO₂ measurements from more than 7000 ground stations over the globe, considerably larger in size and spatiotemporal coverage than used in recent high-resolution NO₂ mapping studies. For statistical modelling, geospatial predictors include Sentinel-5 satellite (Tropomi instrument) measurements, variables relating to the emission sources (e.g., road network), dispersion processes (e.g., meteorological variables), elevation and Earth nightlights (from VIIRS nightlight data). We evaluate various statistical models including linear models, ensemble tree-based models, deep convolution models, stacked models with regularisation, and hierarchical modelling strategies and select the optimal model for mapping. Evaluation of models included uncertainty assessment as well as spatial validation methods.

How to cite: Lu, M., Schmitz, O., de Hoogh, K., Hystad, P., Knibbs, L., Kai, Q., and Karssenberg, D.: Global, high-resolution statistical modelling of NO2, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-6355, https://doi.org/10.5194/egusphere-egu21-6355, 2021.

Atmospheric chemistry is critical in determining air quality and thus impacts climate change. Anthropogenic species are released into the atmosphere, and undergo complex photochemical transformations leading to the production of secondary pollutants, among which ozone and particulate matter. This can induce adverse effects on human health, visibility, ecosystems and local meteorology.  The combination of state-of-the-art atmospheric models with accurate atmospheric measurements of atmospheric species abundances is needed to evaluate whether atmospheric models can successfully simulate the chemical and physical processes occurring, and hopefully monitor the emissions of anthropogenic compounds and help in the implementation and verification of abatement policies.

In this work, ground-based, airborne and spaceborne measuring techniques are used to evaluate the performance of the full chemistry on-line WRF-Chem model over Antwerp in Flanders, Belgium, one of the areas with the highest NO2 pollution in the world. The model is configured to allow two nested domains with spatial resolution changing from 5 to 1km, so as to pinpoint the most pollutant sources in the region, and applied to simulate the urban air quality over the Antwerp agglomeration.

We will briefly discuss the choices and adaptations made regarding the physical parameterizations, emission inventories and chemical mechanism. The model performance is evaluated through comparison with various observation types. The physics parameterizations in WRF model  are evaluated through comparison against ground-based data from two meteorological stations in the Antwerp region. The WRF-Chem NO2 distributions are evaluated against (1) hourly measured concentration values from monitoring stations in Flanders, (2) vertical columns measured by an airborne hyperspectral imager APEX, providing a 2-dimensional spatial mapping, on 27 and 29 June 2019, and (3) spaceborne NO2 columns over Belgium obtained from the high-resolution TROPOMI instrument aboard S5p. The consistency of the model biases across the three datasets will be discussed, and recommendations will be made for improving model performance in this region.

How to cite: Poraicu, C., Müller, J.-F., Stavrakou, T., Fonteyn, D., Tack, F., and Veldeman, N.: Developing high-resolution simulations of tropospheric NO2 over Flanders using WRF-Chem, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-2063, https://doi.org/10.5194/egusphere-egu21-2063, 2021.

Cooking operations can be an important fine PM source for urban areas. Cooking emissions are a source of pollution that has been often ignored and are not included or are seriously underestimated in urban emission inventories. However, several field studies in cities all over Europe suggest that cooking organic aerosol (COA) can be an important component of the total organic PM. In this study we propose and evaluate a methodology for the simulation of the COA concentration and its variability in space and time in an urban area. The city of Patras, the third biggest in Greece is used for this first application for a typical late summer period. The spatial distribution of COA emissions is based on the exact location of restaurants and grills, while the emissions on the meat consumption in Greece. We estimated COA emissions of 150 kg d^-1 that corresponds to 0.6 g d^-1 per person. The temporal distribution of COA was based on the known cooking times and the results of the past field studies in the area. Half of the daily COA is emitted during dinner time (21:00-0:00 LT), while approximately 25% during lunch time (13:00-16:00 LT). The COA is simulated using the Volatility Basis Set with a volatility distribution measured in the laboratory and is treated as semivolatile and reactive. The maximum average COA concentration during the simulation period is predicted to be 1.3 μg m^-3 in a mainly pedestrian area with a high density of restaurants. Peak hourly COA concentrations in this area exceed 10 μg m^-3 during several nights. The local production of secondary COA is predicted to be slow and it represents just a few percent of the total COA.

How to cite: Siouti, E., Skyllakou, K., Kioutsioukis, I., Ciarelli, G., and Pandis, S. N.: Simulation of the Cooking Organic Aerosol Concentration Variability in an Urban Area, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-10090, https://doi.org/10.5194/egusphere-egu21-10090, 2021.

Given the steep trajectory of the global climate crisis, current emission allowances following the 2015 Paris Agreement require that national GHG budgets (sources/sinks) are quantified more accurately and more timely. Large cities also play a key role in achieving the national objectives of emission reduction as urbanization reaches unprecedented levels. Atmospheric inversion approaches have the potential to produce a semi-independent assessment of these fluxes by combining atmospheric data and high-resolution inventories.  However, these approaches only provide an estimate of the total city flux, with no information on the per sector distribution, a major shortcoming for policy makers.

Multiple emission datasets have been developed worldwide at various spatial scales in order to provide a better understanding of the global carbon cycle, but also more locally for large cities and emission hot-spots. Due to the different methodologies and the quality of the surrogate data, large discrepancies are observed between these datasets, especially at the sectoral level. To allow for sectoral attribution in GHG inversions, we investigate Air Quality (AQ) data as additional information assimilated jointly with GHG’s to attribute atmospheric information to specific sectors of activity.

We focus here on the Paris metropolitan area and analyze ground-based observations as well as high-resolution emission inventory estimates for both GHG’s and other reactive pollutants. The observations were acquired by the ICOS GHG monitoring network and the Airparif AQMN. Bottom-up emission estimates were provided by three different emission products for CO₂, CO, NO_X.

We analyzed the atmospheric signals using a backward-in-time Lagrangian Particle Dispersion Model (LPDM) driven by meteorological variables from mesoscale simulations (WRF-FDDA) at 1-km resolution to represent the origin of the emissions (so-called tower footprints).

The modelled concentrations were compared to observations from March to June for the year 2019 to assess the validity of the temporal variations, for each emissions dataset and for both weekly and diurnal cycles. Furthermore, we estimated a correction factor for the modelled NO_X, CO, and CO₂ concentrations using a Monte-Carlo approach that optimizes the three inter-species ratios (NO_X/CO, NO_X/CO₂, and CO/CO₂) to quantify the actual emissions for these three species. We further look into the first Covid-19 lockdown period of 2020 to evaluate the applicability of the method, a first step toward providing process-based information from atmospheric observations, and determine the sectoral contributions to observed emissions changes.

How to cite: Abdallah, C., Lauvaux, T., Gros, V., jinghui, L., Bréon, F.-M., Ramonet, M., Ciais, P., Denier van der Gon, H. A. C., Perussel, O., Baudic, A., and Laurent, O.: From fluxes to signals: A joint analysis of GHG and Air Quality over the Paris Megacity, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-11915, https://doi.org/10.5194/egusphere-egu21-11915, 2021.

Climate regulations and satellite monitoring on increasingly high resolution creates a demand for an insight into emissions on an urban scale. The aim of the Ruisdael Observatory (www.ruisdaelobservatory.nl) is to provide just that: detailed and high-resolution modelling and measurements of weather and air quality in a domain covering the Netherlands.

The Ruisdael Observatory created a renewed impulse in the developments of the DALES Large-eddy simulation (LES) model (Heus et al., 2010, Ouwersloot et al. 2016) to find and push the limits of atmospheric modelling. Typical simulations with DALES will use a spatial resolution in the order of 100m in domain sizes spanning over 100x100 km. This high resolution justifies the complexity and the multitude of emission sources and resulting transport of pollutants in the atmospheric boundary layer.

The combination of high resolution and large domain sizes allows us to investigate how emissions disperse in a turbulent environment which is forced by large-scale flow at the same time. Parameterizations are no longer needed to calculate horizontal or vertical transport in the boundary-layer. This way, we can provide new insight into the transport of emissions in the boundary layer and the detrainment of gases out of the boundary layer into the free atmosphere.

We will discuss the construction of our emission database for the Netherlands with a 100-meter and 1-hourly resolution. For this, we started from the official E-PRTR reported emission inventories (www.emissieregistratie.nl) and enriched with high resolution activity data from mostly open-source datasets. Moreover, large emissions sources (accounting for e.g. >80% of CO2 emissions) are subject to mandatory registration and their locations are known exactly. Emissions from different source categories can be tracked individually and compared to measurements from the Ruisdael Observatory measurement sites. Examples of simulations of fair-weather summer days will be compared to surface measurements and showcase the data richness of our new model and combination to measurements from our network.

How to cite: de Bruine, M., Jansson, F., van Stratum, B., Rijsdijk, P., and Houweling, S.: Simulating the emission and transport of gases on 100-meter resolution in a 100-kilometer domain., EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-13055, https://doi.org/10.5194/egusphere-egu21-13055, 2021.

High air pollution levels pose a threat to both human health and ecosystem vitality in Hebei Province, NE China. Although air quality changes are monitored hourly with high-end equipment at the provincial scale (197 stations for 187,693 km²) it is difficult for individual counties or cities to improve local air quality based on regional-scale information. The Sino-Dutch Technology Transfer & Training Project established a monitoring network of 43 low-cost air-boxes and 11 standard meteorological stations in Shexian county, Handan city (~ 1500 km²) to measure atmospheric concentrations of PM₁₀, PM_2.5, CO, SO₂, NO₂ and O₃ at 1-min intervals from January 2020 onwards. Data from these stations were evaluated in real time using the TNO Gaussian plume model. The model provides point emission levels of PM₁₀, PM_2.5 and CO at 10-min intervals after calibration against measured concentrations. Based on a 2019 pollution source inventory, 21 major source areas were identified and used to derive an optimized source map for model input – including a large steel company, a coal-fueled power plant, different industrial complexes (cement, coking plant for ore smelting), as well as the densely populated city centre, rural residential areas, and a busy highway. The model performs source optimization using concentration data for all 43 stations and subsequently calculates the contributions of individual sources for each monitoring station to see to what extent the source map explained observed concentrations. Full network operation started in July 2020. Based on a one-month test period (August 2020), the steel company and coking plant were estimated to contribute ~25% of the total area’s PM-emissions. The central city area contributed ~10% and 17% of total PM- and CO-emissions, respectively, mostly due to construction activity and traffic. Repeating the exercise for the two provincial monitoring stations that also had high-end equipment in place in the downtown area gave inferred average urban contributions to measured concentrations as high as 60–62.5% for PM₁₀ and PM_2.5 versus 48% for CO. The steel factory contributed an estimated 9–11% for PM₁₀ and PM_2.5 at these locations and a cement factory 13% for CO. The combined results underline the importance of taking spatial variability of emission sources into explicit account in complex industrialized cities. Moreover, the combination of a low-cost airbox real-time monitoring network with emission apportionment modeling will allow local policy-makers to take proper actions towards reducing air pollution levels at the local scale.

How to cite: Zhang, J., Hensen, A., Seignette, P., and Yu, D.: Low-cost airbox network and Gaussian plume modelling to assist air quality policy-making at the local scale: Shexian County, Hebei Province, China, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-7296, https://doi.org/10.5194/egusphere-egu21-7296, 2021.

Despite efforts to reduce pollutant emissions in the UK, between 28,000 and 36,000 deaths a year are attributable to poor air quality and ambient air pollution is considered the UK’s biggest environmental threat to health. Characterising, quantifying and understanding air quality variability and the importance of different drivers is essential to guide policies to address the issue and its risks, for both the short and long term. Here we investigate a statistical modelling approach to characterise air quality variability and its key drivers, using Kalman filters. Kalman filters are a commonly used tool in air quality modelling but are seldom used in a statistic framework that accounts for uncertainty in a principled way. Kalman filtering allows us to take data which is noisy or partially recorded, such as air quality data, and help reveal the true underlying trends and dynamics of the data. This allows us to combine measurement information with the statistical model to obtain an air quality forecast, using the measurement information to reduce the statistical model errors and improve model results. We explore this approach using air quality monitoring data from the UK Automatic Urban and Rural Network (AURN), which consists of 150 sites focussed mainly in populated areas, leaving large areas unmonitored. AURN is primarily used for compliance reporting against national and European air quality standards and targets. Eventually, our aim is to provide short-term forecasts of pollutant levels from AURN, comparing this against process model forecasts and ultimately providing an optimised combination of process model, statistical model and measurement.   

How to cite: Duncan, R., Young, P., and Nemeth, C.: Forecasting UK site level air quality with a Kalman filtering approach, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-11249, https://doi.org/10.5194/egusphere-egu21-11249, 2021.

Only a few studies have reported sources, characteristics, and strategies for controlling severe air pollution events frequently occurring in several urban areas in India. For a detailed analysis of particulate matter (PM) and gaseous species for their temporal and spatial distribution, a high-resolution simulation through Weather Research and Forecasting with Chemistry (WRF-Chem) model was undertaken for the entire India. Emission Database for Global Atmospheric Research (EDGAR v2.2) was used. WRF-Chem model was used for predicting concentrations of NO₂, O₃, CO, SO₂, and PM_2.5 along with its components in major cities (Delhi, Lucknow, Patna, Kolkata, Ahmedabad, Mumbai, Hyderabad, Bangalore, Chennai) spread all over India. The model's performance was validated against observations that were available for a few large cities from national ambient air quality monitoring stations. Generally, O₃ predictions did not show an acceptable association with the measurements, but PM_2.5 predictions did meet the model performance criteria (root mean square error (RMSE), normalized mean bias (NMB), normalized mean error (NME), mean fractional bias (MFB) and mean fractional error (MFE)). Model performance was better for days with higher levels of PM_2.5. PM_2.5 showed the highest concentration levels for India's Northern and Eastern parts and a major portion of the Indo-Gangetic Plain (IGP). Concentrations of PM_2.5 were observed to be lower during monsoon and higher during the winter seasons. Nitrate levels were found to be 150–240% higher in winter than the yearly average. However, a decrease in solar radiation intensity and temperature during the winter season showed sulfate levels to be much lower than in other seasons. Except for South India, Primary Organic Aerosol (POA) contribution to PM_2.5 was highest for regional analysis. Analysis of model concentrations indicates the importance of controlling precursor gases for secondary pollutants in India. Conclusively, WRF-Chem predicted particulate and gaseous air pollutant levels can be used to develop control strategies for large regions that are part of the same airshed.

How to cite: Azmi, S., Nagar, P. K., and Sharma, M.: Regional emission loading of particulate and gaseous air pollutants over India using fine resolution WRF-Chem simulation technique, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-1135, https://doi.org/10.5194/egusphere-egu21-1135, 2021.

Due to fast industrialization and urbanization, air pollution is more and more serious in Taiwan. Generally, many anthropogenic factors can affect air quality; for example,  exhaust gas from automobiles and motorcycles, factory emissions, fossil fuels, burning straw, incinerators, etc. The factors are highly associated with land use. Previous studies typically used multiple linear regression model to analyze the relationships between air quality and land use. This study adopts multi-threshold land use logistic regression (LULR) models with several continuous and categorical variables to assess different levels of fine particulate matters (PM_2.5) in Taiwan and to determine key land-use factors controlling various levels of air PM_2.5 pollution. First, data on annual air PM_2.5 pollution in the Taiwan Island are collected in 2017. Four thresholds of 16.37, 18.68, 21.83, 25.83 µg/m³ are determined based on the 20th, 40th, 60th, and 80th percentiles, respectively, of observed data. Geographical information system is then adopted to analyze data on 29 environmental variables obtained from the three main dimensions–information of land-use categories, amounts of specified pollution sources in townships, and geographical locations adjacent to monitoring stations of air quality. Finally, data in 2017 are employed to establish the LULR model and significant land-use factors causing air PM_2.5 pollution are determined using stepwise LULR models for various levels of air PM_2.5 pollution. Moreover, data in 2018 are used to verify the established LULR models. The analyzed results reveal that correct responses of the LULR models range from 83.6% to 100%. For the 20th-percentile threshold, locations and the industry land-use area are positively contributed to air pollution, while tempt densities and building, agriculture, forest land-use areas are negatively contributed to air pollution. For the 40th-percentile threshold, locations, plains with an elevation of less than 150 m, and agriculture land-use areas are related to air pollution. For the 60th-percentile threshold, locations are positively related to air pollution, while forest land-use areas are negatively related to air pollution. For the 80th-percentile threshold, locations and industry park areas associated with air pollution. According to the research results, a feasible strategy of environmental management and outdoor activities is proposed.

How to cite: Jang, C.-S.: Applying land use logistic regression models to assess different levels of air quality and to determine key environmental factors in Taiwan, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-1182, https://doi.org/10.5194/egusphere-egu21-1182, 2021.

The online-coupled Weather Research and Forecasting model with Chemistry (WRF-Chem v4.0), was applied to evaluate the impact of using different anthropogenic emissions inventories on regional air quality in Argentina. For this purpose, we couple the Argentinian high-resolution emissions inventory (GEAA-AHRI) and the Emissions Database for Global Atmospheric Research – Hemispheric Transport of Air Pollution (EDGAR-HTAP) and introduce them into the model, with a local optimized configuration considering 3 nested domains with a horizontal grid size of 20 x 20 km, 4 x 4 km, and 1.3 x 1.3 km and the MOZART chemical scheme. The model output for NO2, PM10, PM2.5, and O3 concentrations over the innermost domain was compared against the existing surface and satellite-derived observations for the Buenos Aires Metropolitan Area (AMBA) during austral fall 2018. We found an overall good model performance for all simulations, and large discrepancies between the emission inventories, obtaining an improved urban-scale spatio-temporal representation when the high resolution GEAA-AHRI dataset is considered. Our results show that the daytime concentrations of air pollutants are strongly influenced by the shape and shift of the hourly emissions profile before sunrise and after sunset, especially for NO2 where the inclusion of the temporal profile decreased the mean bias by ~80%. Performance criteria for modeled PM10 and PM2.5 were in general satisfied, despite having an average underestimation of observations. When compared to NO2 tropospheric columns derived from TROPOMI, The general magnitude and spatial pattern of the NO2 tropospheric column is in agreement with the mean TROPOMI columns during the modeled period, obtaining correlation coefficients higher than 0.6 for all simulations. Our results highlight the benefits of using a time-dependent and high-resolution local inventory for addressing the background air quality in AMBA. The implementation and validation of local emissions and static fields with high spatial and temporal resolution carried out in this work, establishes a benchmark for forthcoming studies in other regions of South America where different modeling tools for air quality analysis are currently being used to complement the usually sparse and discontinuous air quality networks.

How to cite: Lopez-Noreña, A. I., Berná, L., Tames, M. F., Millán, E., Puliafito, E., and Fernandez, R. P.: Evaluation of NO2, O3, PM10, and PM2.5 in the city of Buenos Aires, Argentina using WRF-Chem model, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-1514, https://doi.org/10.5194/egusphere-egu21-1514, 2021.

Estimating personal exposure to air pollution is important in investigating the impact of air pollution on chronic diseases such as diabetes or cardiovascular disease. Long-term personal exposures estimates from large cohorts are required to reliably identify the relation between chronic air pollution exposure and non-communicable disease outcomes. Using e.g. yearly averaged concentrations at fixed locations such as the home address may result in incomplete quantification of personal exposure as persons move in space and time. An appropriate estimation involves mapping of space-time variation of concentrations as well as incorporating several activities of individuals at different locations and the mobility of individuals along their space-time paths. While for small surveys detailed information is often available (e.g. home and work address, GPS tracking data and travel mode), this abundance of data is not available for large-scale personal exposure assessment. Thus, for large-scale exposure assessment the first challenge is the design of model representations of individual mobility for which parameters can be identified with relatively limited observational data on individual mobility. The second challenge is the execution of such large-scale models over large populations.

We address these challenges by developing a modelling framework on top of Campo (https://campo.computationalgeography.org) that combines the space-time mapping of pollution and activity-based mobility simulation of individuals. To represent data sparse information on individuals, we use personal activity schedules. Air pollution is based on land use regression models. Our modelling approach contains the following key components: a) an activity schedule generator allowing to express the type, location and duration of an individual's activity as a function of a person's profile defined by e.g. age, gender or occupation, and b) a spatial context generator providing the location of an individual during a particular activity. Activities cover residence in certain areas (home, work, leisure) or along routes using different travel modes (car, bicycle, on foot), and c) an exposure estimator. Exposure estimation is subsequently the combination of the spatial contexts for each activity with air pollution concentrations at corresponding times.

Using these decoupled but interacting components provides the flexibility to express a broad range of representative time spans and spatial residences, required e.g. to represent uncertainty of unknown work locations or travelled routes. We present concepts and the model using a nationwide cohort from Switzerland.

How to cite: Schmitz, O., Lu, M., de Hoogh, K., Probst-Hensch, N., Jeong, A., Flückiger, B., Vienneau, D., Hoek, G., Kyriakou, K., Vermeulen, R. C. H., and Karssenberg, D.: Nationwide estimation of personal exposure to air pollution using activity-based field-agent modelling, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-7081, https://doi.org/10.5194/egusphere-egu21-7081, 2021.

Air pollution has a serious impact on health and this problem will be aggravated under the action of climate change. This climate penalty can play an important role when trying to assess future impacts of air pollution on several pathologies. Among these diseases, the scientific literature is scarce when referring to the influence of atmospheric pollutants on neurodegenerative diseases for future climate change scenarios. Under this framework, this contribution evaluates the incidence of dementia (Alzheimer's disease and vascular dementia) occurring in Europe due to exposure of air pollution (essentially NO₂ and PM2.5) for the present climatic period (1991-2010) and for a future climate change scenario (RCP8.5, 2031-2050). The GEMM methodology has been applied to climatic air pollution simulations using the chemistry/climate regional model WRF-Chem. Present population data were obtained from NASA's Center for Socioeconomic Data and Applications (SEDAC); while future population projections for the year 2050 were derived from the United Nations (UN) Department of Economic and Social Affairs-Population Dynamics.

Overall, the estimated incidence of Alzheimer's disease and vascular dementia associated to air pollution over Europe is 498,000 [95% confidence interval (95% CI) 348,600-647,400] and 314,000 (95% CI 257,500-401,900) new cases per year, respectively. An important increase in the future incidence is projected (around 72% for both types of dementia) when considering the effect of climate change together with the foreseen changes in the dynamics of population (expected aging of European population). The climate penalty has a limited effect on the total changes of Alzheimer's disease and vascular dementia (approx. 0.5%), since the large increase in new annual cases over southern Europe is offset by the decrease of the incidence associated to these pathologies over more northern countries, favored by an improvement of air pollution caused by the projected enhancement of rainfall.

How to cite: Jiménez-Guerrero, P., Guzmán, P., Tarín-Carrasco, P., and Morales-Suarez-Varela, M.: Effects of air pollution on neurodegenerative diseases (Alzheimer's disease and vascular dementia) over Europe for present and future climate change scenarios, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-14494, https://doi.org/10.5194/egusphere-egu21-14494, 2021.

Warm summer temperatures provide ideal conditions for the occurrence of extreme ground level ozone pollution episodes. Given the well-established negative impacts of ozone on human and plant health, understanding and attributing these extreme events is of importance to the scientific and wider community, particularly as heatwaves may become more frequent due to climate change. Extreme Value Analysis provides a powerful and flexible framework in which to statistically model unusually large observed values of ozone extracted from historical data. Here, a temperature dependent Peaks-Over-Threshold method based upon the Generalised Pareto Distribution is used to carry out a regional comparison of extreme ozone pollution episodes within the UK. Our analysis uses surface ozone observations from the UK’s extensive Automatic Urban and Rural Network. The statistical model was used to quantify the frequency and magnitude of extreme ozone events, including a probabilistic assessment of exceeding UK public health thresholds, conditional on temperature. Return levels are provided for each monitoring site demonstrating the expected future projections of extreme ozone pollution events across the UK. We find that across UK rural background sites, return periods for a daily maximum 8-hr ozone level of 100 ug/m3 (a 'moderate' level of air pollution in the UK's Air Quality Index) range from 32-147 days, based on analysis of the data in the decade 2010-2019. Similarly, for urban background sites the range is 36-869 days. An analysis of the spatio temporal variability in UK ozone extremes, along with their temperature dependence, will be presented.

How to cite: Gouldsbrough, L., Hossaini, R., Eastoe, E., and Young, P. J.: Investigating the occurrence, likelihood and regional variability of extreme ozone pollution episodes across the UK, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-7653, https://doi.org/10.5194/egusphere-egu21-7653, 2021.

The fifth version of the Emission Database for Global Atmospheric Research (EDGAR 5.0) provides an impressive inventory of various pollutants. Pollutants from different emission sectors are available with daily, monthly and yearly temporal profiles at a high global resolution of 0.1°×0.1°. Although this resolution has been sufficient for regional air quality studies, the emissions appeared to be too coarse for local air quality studies in areas with complex topography. With Switzerland as a case study, we present our approach for downscaling EDGAR emission data to a much finer resolution of 0.02°×0.02° with the aim of modelling local air quality.

We downscaled the EDGAR emissions using a combination of GIS tools including QGIS, ArcGIS, and a series of python scripts. We obtained the surface coverage of different land use features within the defined EDGAR emission sectors from Open Street Map (OSM) using the QuickOSM tool in QGIS. With the calculated local surface area coverage of the emissions sectors, we downscaled the EDGAR inventory data within ArcGIS using a set of developed Arcpy script tools.

The outcome was a much finer resolved emission dataset which we fed into the WRF-CHEM air quality model within a pilot project. A comparison of the modelled pollutant concentrations using the two datasets (original EDGAR data and the downscaled data) shows an improved agreement between the downscaled dataset and the measurement data.

Studies investigating the impact of urbanization, land use change or traffic pattern on air quality may benefit from our downscaling solution, which, thanks to the global coverage of OSM, can be globally applied.

How to cite: Edebeli, J., Spirig, C., and Anet, J.: Downscaling EDGAR emissions to local emission sectors for Switzerland, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-8208, https://doi.org/10.5194/egusphere-egu21-8208, 2021.

Quantification of the spatial and temporal variations in the sources of air pollutants, especially PM_2.5, can inform control strategies and, potentially, the understanding of PM_2.5 health effects. Three-dimensional chemical transport models (CTMs) are well suited to help address this problem, since they simulate all the major processes that impact PM_2.5 concentrations and transport. In this study we quantify the changes in the concentration, exposure, composition, and sources of PM_2.5 in the US from the early 1990s to the early 2010s. Significant reductions of emissions of SO₂, NO_x, VOCs and primary PM have taken place in the US during the last 20 years. We evaluate our understanding of the links between these emissions and concentration changes combining a chemical transport model (PMCAMx) with the Particle Source Apportionment Algorithm (PSAT) (Skyllakou et al., 2017). Results for 1990, 2001 and 2010 are presented. The reductions in SO₂ emissions (64% mainly from electrical generation units) during these 20 years have dominated the reductions in PM_2.5 leading to a 45% reduction of the sulfate levels. The predicted sulfate reductions were in excellent agreement with the available measurements. Also, the reductions in elemental carbon (EC) emissions (mainly from transportation) have led to a 30% reduction of EC concentrations. The most important source of OA through the years according to PMCAMx is biomass burning followed by biogenic SOA. OA from on-road transport has been reduced by more than a factor of 3, on the other hand changes in biomass burning OA and biogenic SOA have been modest. In 1990 90% of the US population was exposed to PM_2.5 concentrations to equal and higher than the suggested annual mean by the WHO (10 μg m^-3), but this reduced to 70% in 2010. Also, the predicted changes in concentrations were evaluated against the observed changes for 1990, 2001 and 2010, in order to understand if the model represents well the changes through the years.

Skyllakou, K., Fountoukis, C., Charalampidis, P., and Pandis, S.N. (2017). Volatility-resolved source apportionment of primary and secondary organic aerosol over Europe, Atmos. Environ., 167, 1–10.

How to cite: Skyllakou, K., Rivera, P. G., Dinkelacker, B., Karnezi, E., Kioutsioukis, I., Hernandez, C., Adams, P., and Pandis, S.: Changes of PM2.5 concentrations and their sources in the US from 1990 to 2010, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-10567, https://doi.org/10.5194/egusphere-egu21-10567, 2021.

Tropospheric O₃ pollution notably contributes to the deterioration of air quality in many metropolitan regions, resulting in detrimental effects on human health and ecosystem. Due to the moderate atmospheric lifetimes of O₃, horizontal transport, exchange between atmospheric boundary layer (ABL) and free troposphere (FT), and chemical process within the ABL all potentially play important roles in regional O₃ pollution. In this study, we developed a post-calculation tool to quantify the hourly contributions of these processes to the regional budget of O₃ mass and concentration variations within the ABL based on the modelling results of the Community Multiscale Air Quality (CMAQ) model. The new features of this tool include: (1) the contributions of ABL-FT exchange on O₃ pollution can be quantified; (2) horizontally, the targeted region can be freely defined by users and vertically, the volumes are non-fixed owing to the diurnal variations of ABL; and (3) the budgets of O₃ mass and concentration variations are separately calculated and analysed. The Pearl River Delta (PRD) region, located in the South China and faced with severe O₃ pollution, was selected as the target region in this study. Results show that the variations of total O₃ mass within the ABL of the PRD were controlled by ABL-FT exchange, that is, the increase (decrease) of O₃ mass in the morning (afternoon) was driven by O₃ inflow (outflow) through ABL-FT exchange. By contrast, it was the chemical process that drove the variations of regional-mean O₃ concentrations. Except that ABL-FT exchange contributed to the rise of O₃ concentrations in several hours after sunrise, O₃ transport did not lead to the notable variation of O₃ concentration in the remaining hours of the day. Combining source apportionment methods, we found that outside O₃ (including O₃ produced by emissions within the East and Central China and background O₃) entered the PRD mainly through ABL-FT exchange. For chemical process, local sources played a major part, but the contributions of outside emissions cannot be neglected, suggesting the contributions of precursor transport. The effects of typhoon periphery, the weather system most related to O₃ pollution in the PRD, were also examined by comparing the budget results on O₃ pollution days with and without the occurrence of typhoons. The usage of this tool will help to comprehensively understand the influence of transport and chemical process in O₃ pollution on the regional scale, which is crucial for effective and strategic O₃ control.

Acknowledgement. This work is sponsored by the National Key Research and Development Program of China (Grant No. 2018YFC0213204, 2018YFC0213506) and the National Science and Technology Pillar Program of China (Grant No. 2014BAC21B01).

How to cite: Qu, K., Wang, X., Cai, X., Yan, Y., Jin, X., Shen, J., Xiao, T., Zeng, L., and Zhang, Y.: The regional budget of O3 mass and concentration variations within the atmospheric boundary layer using the CMAQ model: An example from the Pearl River Delta, China, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-11180, https://doi.org/10.5194/egusphere-egu21-11180, 2021.

During the Air Quality and Climate Change in the Arabian Basin (AQABA) campaign a MAX-DOAS instrument was set up on board of the Kommandor Iona. The ship route covered a variety of regions with different atmospheric compositions: Clean air in the Mediterranean and the Arabian Sea, anthropogenic air pollution near the oil fields in the Arabian Gulf or in areas of dense ship traffic like the Suez Channel or the dust clouds of the nearby deserts in the Red sea. The measured spectra in the UV/VIS spectral range (302 to 467nm) provide sufficient information for the retrieval of aerosol and trace gas profiles. In this study, we focus on evidences of direct nitrous acid emission sources in harbor areas around Jeddah and Kuwait. Since HONO daytime chemistry is debated in recent literature and missing sources are being discussed, we compared the results of the MAX DOAS measurements to WRF-Chem model output in order to identify potential daytime sources in maritime/harbor regions.

How to cite: Dörner, S., Donner, S., Behrens, L., Beirle, S., Osipov, S., and Wagner, T.: Enhanced levels of nitrous acid during daytime derived from MAX-DOAS measurements during the AQABA campaign in late summer 2017, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-2571, https://doi.org/10.5194/egusphere-egu21-2571, 2021.