We invite contributions on all aspects of forecasting and data assimilation of aerosols, reactive gases, greenhouse gases, and weather or stratospheric dynamics across different time scales. Our focus is on the scientific, computational, and societal advantages of such integrated approaches. Specifically, but not exclusively, we invite papers addressing the following topics:
a) Improved weather predictions due to simulated feedback between aerosols and chemistry in radiation and cloud physics,
b) Improved AC prediction by improved representation of the meteorological variability,
c) Advancements in operational NWP-AC prediction systems, in particular using data-driven machine learning methods,
f) Data assimilation developments for AC and NWP systems,
g) Forecasting of stratospheric composition and dynamics after large volcanic eruptions such as the Hunga-Tonga,
h) Combined impact of environmental hazards on society, such as air pollution and high-impact weather, wildfires, dust storms and the underlying meteorological factors,
i) Evaluation, validation, and applications of NWP-AC prediction systems.
This Session is organized in cooperation with the Copernicus Atmosphere Monitoring Service (CAMS) and the Global Air Quality Forecasting and Information Systems (GAFIS) initiative of the WMO Global Atmosphere Watch (GAW) Program.
Machine learning has shown great success in numerical weather prediction. Here, we extend these advances to atmospheric composition forecasting by introducing AIFS-Compo, an AI-based system for predicting aerosols and reactive trace gases. Building on ECMWF’s AI weather forecasting framework, AIFS, we develop a large-scale graph-transformer model trained in two stages: first on CAMS EAC4 reanalysis data, and subsequently on a combination of CAMS analysis and forecast data. The resulting system produces 3-hourly forecasts and jointly uses prognostic variables from both numerical weather prediction and atmospheric composition.
When verifying against observations, AIFS-Compo achieves lower errors than the operational IFS-Compo system for 5-day forecasts of aerosol optical depth (AOD) and PM2.5, while showing comparable skill for reactive gases including ozone, carbon monoxide, nitrogen dioxide, and sulfur dioxide. Overall, AIFS-Compo delivers performance competitive with the operational system at a fraction of the computational cost. This efficiency for example enables extension to longer leadtimes, such as 10-day forecasts, supporting applications including early ozone hole prediction.
How to cite: Harder, P., Flemming, J., Alexe, M., and Chantry, M.: AIFS-Compo: A Data-Driven Atmospheric Composition Forecasting System , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-18643, https://doi.org/10.5194/egusphere-egu26-18643, 2026.
For thirty years, the European Research Infrastructure IAGOS has been equipping commercial aircraft with instruments to monitor the composition of the atmosphere on long-haul flights around the world. The aircraft measure a number of chemical species, meteorological parameters and cloud particles at cruise altitude in the upper troposphere/lower stratosphere and during landing and take-off at many international airports. The in-situ data on chemical composition of the atmosphere collected by IAGOS is used in the routine validation of the forecasts and analyses from the Copernicus Atmosphere Monitoring Service (CAMS). This evaluation by IAGOS now covers the CAMS global and regional forecasts for ozone, carbon monoxide, and nitrous oxides, and the CAMS global greenhouse gas forecasts for carbon dioxide and methane. We describe recent cut-off lows that led to severe flooding over Spain in Autumn 2024, characterised by the dynamical fields in the ERA-5 meteorological re-analysis. IAGOS measurements of the trace gases, ozone, water vapour and carbon monoxide in the cut-off low allowed us to identify stratosphere to troposphere transport. We describe the differences between the CAMS forecasts and IAGOS observations of these trace gases during this event. In addition, we use a new tool based on FLEXPART using ERA-5 winds and METEOSAT third generation's (MTG) lightning imager to determine the origin of elevated NOx observed at altitude.
How to cite: Clark, H., Sauvage, B., Bennouna, Y., Patuel, J., Mahnke, C., and Rohs, S.: Upper tropospheric composition during Iberian cut-off lows as seen by IAGOS and CAMS, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-5262, https://doi.org/10.5194/egusphere-egu26-5262, 2026.
Within the Copernicus Atmosphere Monitoring Service (CAMS), ECMWF operates the Integrated Forecasting System with atmospheric composition extension (IFS-COMPO) to provide global forecasts and reanalyses of aerosols and trace gases. In support of ongoing preparations for a new CAMS reanalysis, which will cover the years 2003-present, multi-decadal model simulations with a fixed IFS-COMPO model configuration have been produced for the same period. The model version for these evaluations is similar as planned for the next CAMS Reanalysis. Particularly, the CAMS-GLOB-ANT-M01 ‘Mosaic’ anthropogenic emissions are used. But different than planned for the new CAMS reanalysis we exclude composition data assimilation and perform the simulation on a comparatively coarse model resolution (~80 km).
In this contribution we evaluate the quality of sulfur, oxidized nitrogen and reduced nitrogen deposition fluxes in IFS-COMPO for the period 2003-2022, making use of E4C as described in Williams et al., GMD, 18, 9913–9943 (2025). We present evaluations against various observational networks, namely CASTNet (US), EMEP (Europe) and EANET (Eastern Asia). Also we compare our simulation results with those obtained for the existing CAMS Reanalysis that is based on an older model configuration.
We will show to what extent the simulated deposition fluxes follow the observed trends in the different parts of the world, thereby giving confidence particularly in the used sulfur and nitrogen emissions on a continental scale. Uncertainties due to modelling will be highlighted by assessing our simulation results with multi-model assessments as done, e.g., in HTAP.
How to cite: Huijnen, V., Rémy, S., Williams, J., Metzger, S., and Flemming, J.: Evaluation of a 20-year simulation of nitrogen and sulfur deposition fluxes in IFS-COMPO, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-11188, https://doi.org/10.5194/egusphere-egu26-11188, 2026.
Accurate aerosol prediction remains challenging due to uncertainties in atmospheric composition arising from imperfect initial conditions, errors in emission inventories, and our limited understanding of aerosol processes and properties interacting with atmospheric variables. Due to their short lifetime and strong spatial/temporal variability, global observations of aerosols and clouds rely heavily on satellite remote sensing.
The U.S. National Science Foundation (NSF) National Center for Atmospheric Research (NCAR) has recently developed the atmospheric Model for Prediction Across Scales (MPAS-A; Skamarock et al. 2012) coupled with the next-generation Goddard Chemistry Aerosol Radiation and Transport model (GOCART-2G; Collow et al. 2024) and interfaced with the Joint Effort for Data Assimilation Integration (JEDI) system. This integrated framework enables online-coupled data assimilation of multi-sensor, hyperspectral satellite aerosol retrievals and all-sky radiances across a wide spectral range within a unified atmosphere-aerosol analysis and forecasting system.
This talk introduces the new MPAS-GOCART2G-JEDI system, with an emphasis on the assimilation of NASA Plankton, Aerosol, Cloud, ocean Ecosystem (PACE) Aerosol Optical Depth (AOD) retrievals and their systematic evaluation against legacy AOD products such as MODIS, VIIRS, and AERONET.
How to cite: Ha, S., Kumar, R., Barth, M., Pfister, G., Wei, S.-W., Park, J., Duda, M., Dang, C., Lacey, F., Lu, C.-H., and Abdi-Oskouei, M.: The NSF NCAR Next-Generation Online-Coupled Air Quality and Weather Analysis and Forecasting System (MPAS-GOCART2G-JEDI), EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-1880, https://doi.org/10.5194/egusphere-egu26-1880, 2026.
Convective systems redistribute atmospheric trace gases due to the inherent high vertical wind velocities. Furthermore, the different convective cloud hydrometeors interact with soluble tracers, partially removing them from the atmosphere via precipitation. These processes have an impact on air quality, acid rain, and upper tropospheric composition and photochemistry via the outflow from the storms. Convection also affects the aerosol particle composition due to cloud processing and potentially enables the new particle formation in the upper troposphere, representing a feedback mechanism on the meteorology. Therefore, it is crucial to accurately represent convective transport and scavenging in models aiming to predict the chemical composition.
As current convection parameterising models struggle with this task, we developed a new parameterisation, Mainz Convective Transport and Scavenging (MCTS). MCTS calculates convective transport and scavenging quasi-simultaneously in one column. It considers tracer redistribution due to the high vertical wind velocity, uptake by droplets, and aqueous-phase chemistry. Retention and uptake by ice crystals are included as well.
To evaluate the novel scheme, a case study was performed for a convective situation observed during the NASA SEAC4RS campaign in the US in 2013. MCTS is compared to DC8 flight observations and to cloud-resolving WRF-Chem simulations performed by Cuchiara et al. (2020). MCTS performs reasonably and sufficiently reproduces the HCHO mixing ratios measured during the convective core intercept flights.
MCTS opens the path for a more consistent and accurate representation of convection composition interactions in large-scale models. Ensuring that the consequences of these interactions, i.e., new particle formation in the upper troposphere, radiative feedback, and air quality can be addressed with enhanced accuracy.
Reference
Cuchiara, G. C., Fried, A., Barth, M. C., Bela, M., Homeyer, C. R., Gaubert, B., et al. (2020). Vertical transport, entrainment, and scavenging processes affecting trace gases in a modeled and observed SEAC4RS case study. Journal of Geophysical Research: Atmospheres, 125, e2019JD031957. https://doi.org/10.1029/2019JD031957
How to cite: Jeske, A., Menken, J. M., Cuchiara, G. C., Ranocha, H., Barth, M. C., and Tost, H.: Mainz Convective Transport and Scavenging: A new parameterization of convection-chemistry-interaction in global chemistry-circulation models, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-1984, https://doi.org/10.5194/egusphere-egu26-1984, 2026.
This work assesses the impact that the assimilation of ozone profiles has on meteorological fields in NWP simulations of recent weather events that were influenced by stratosphere-troposphere interactions.
We use the European Centre for Medium-Range Weather Forecasts (ECMWF) IFS model with CAMS configurations, focusing on Northern Hemisphere winters within the period 2020-2023. We investigate the impact of Microwave Limb Sounder (MLS) ozone profiles, as MLS on the Aura satellite has been providing essential observations of ozone for the stratosphere and the upper troposphere-lower stratosphere (UTLS) regions.
Chemistry-dynamics interactions in these regions are key for winter weather and climate patterns, and for the coupling between troposphere and stratosphere. Our work highlights the capacity of MLS O3 to enhance weather forecasting and shows the need for alternatives once MLS is decommissioned.
Our study also explores alternatives to be used after MLS data will stop being available. And it shows the need for future observation platforms similar to the ESA-CAIRT EE11 candidate instrument, to provide atmospheric composition measurements that would enable better representation of stratospheric and UTLS processes and enhance stratosphere-troposphere coupling in weather forecast systems and reanalyses.
How to cite: Monge-Sanz, B., Inness, A., Errera, Q., and Sinnhuber, B.-M.: Stratospheric composition Limb observations to improve NWP forecasts and (re)analyses, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-19175, https://doi.org/10.5194/egusphere-egu26-19175, 2026.
Organic compounds constitute roughly half of the submicron aerosol mass in the troposphere, highlighting the need for accurate representation of organic aerosol (OA) in atmospheric composition (AC) modeling systems to improve aerosol forecasting. The secondary fraction of OA (SOA), formed through the oxidation of various volatile organic compounds (VOCs) from both natural and anthropogenic sources, complicates OA simulation. However, most global AC models either assume a nonvolatile SOA produced with a constant yield from known precursors or provide a simplistic parameterization of its volatility, treating the primary fraction of OA (POA) as nonreactive. This approach often fails to accurately reproduce observed OA atmospheric measurements. Additionally, primary biological aerosol particles are commonly identified as part of the supermicron OA mass, although most global AC models inadequately represent them.
In the context of the Copernicus Atmosphere Monitoring Service, we focus on improving the OA representation in the CAMS global forecasting system (IFS-COMPO). We here present simulations of the partitioning and chemical evolution of POA vapors, including their changes in volatility, as well as the incorporation of coarse organic carbon emissions from major bioaerosol species. The formation of SOA from semi-volatile organic compounds (SVOCs) and intermediate-volatility organic compounds (IVOCs) has been integrated into the SOA formation schemes from biogenic and anthropogenic VOCs in IFS-COMPO using a lite version of the well-documented aerosol module ORACLE, which allows for relatively limited computing resources. Additionally, fungal spores and pollen grains have been included in IFS-COMPO through interactive emission schemes that depend on ecosystem types, the leaf area index (LAI), and meteorological parameters. Overall, our efforts aim to bridge the gap between model simulations and observations, thereby enhancing our understanding of the atmospheric organic carbon burden.
How to cite: Myriokefalitakis, S., Rémy, S., Huijnen, V., Karathanasopoulos, O., Tsimpidi, A. P., and Karydis, V. A.: An Improved Representation of Organic Aerosol Composition in the ECMWF IFS-COMPO, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-4169, https://doi.org/10.5194/egusphere-egu26-4169, 2026.
Volcanic ash and SO2 clouds pose significant hazards to aviation, air quality, and downwind ecosystems, motivating rapid, physically consistent plume modeling. We present a new and improved volcanic capability in WRF-Chem v4.8 [1] that addresses key limitations of earlier implementations by (i) strengthening ash/SO2/sulfate mass conservation and diagnostics, (ii) correcting ash gravitational settling and refining removal pathways via added wet and dry deposition for ash and sulfate, (iii) improving SO2 oxidation to sulfate (gas-phase and in-cloud pathways), and (iv) enabling direct radiative effects of ash and sulfate for fully interactive aerosol–meteorology coupling. These developments are paired with an emission-preprocessing workflow that supports time- and height-varying volcanic source terms for rapid-response simulations.
We demonstrate the approach for the unusually explosive Hayli Gubbi eruption (Afar, Ethiopia) on 23 November 2025 by reconstructing emissions using backward-trajectory analysis and constraining the simulation with satellite and ground-based observations. The downwind plume was captured by an AERONET station in Oman, providing rare constraints on the ash size distribution, while TROPOMI retrievals constrain SO2 columns and plume properties. The enhanced WRF-Chem reproduces the observed plume timing and structure and yields best-fit total emissions of approximately 1.0 Mt of fine ash and 0.3 Mt of SO2. Ash is injected mainly at ~7-11 km with a brief pulse up to ~14 km, whereas SO2 is emitted higher (~8–16 km) and remains predominantly tropospheric during the first day, implying limited near-term climate impact for the inferred SO2 burden. Simulated deposition indicates measurable ash fallout to the southern Red Sea and Gulf of Aden within ~28 hours, consistent with satellite-observed chlorophyll anomalies suggestive of an early marine response. This workflow is readily transferable to other eruptions for near-real-time plume forecasting and impact assessment.
References:
[1] Ukhov, A., Stenchikov, G., Schnell, J., Ahmadov, R., Rizza, U., Grell, G., and Hoteit, I.: Enhancing volcanic eruption simulations with the WRF-Chem v4.8, Geosci. Model Dev., 18, 9805–9825, https://doi.org/10.5194/gmd-18-9805-2025, 2025.
How to cite: Ukhov, A., Masabathini, S., Pateraki, M., Papagiannopoulos, N., Rizza, U., and Hoteit, I.: Rapid-response simulation of the 2025 Hayli Gubbi eruption with an enhanced WRF-Chem v4.8 model, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-5190, https://doi.org/10.5194/egusphere-egu26-5190, 2026.
Aerosols are a ubiquitous part of the Earth’s climate system, where they influence radiative forcing, cloud microphysics, and air quality. Accurate modelling of their spatiotemporal evolution is needed for producing accurate simulations of climate and air quality impacts. Thus far, the aerosol description of ECMWF IFS-COMPO (Integrated Forecast System with atmospheric composition extension) has relied on a “bulk-bin” scheme (denoted IFS-AER) where only aerosol mass is simulated. However, a detailed representation of both mass and number concentrations of aerosols is required for a more accurate simulation of the climate effects and impact on air quality. For this work we show results from IFS-COMPO (Cy50r1) and OpenIFS (Cy48r1; portable and easy-to-use version of the IFS). Within these models we replaced the AER scheme with a modal aerosol scheme based on HAM-M7 (Hamburg Aerosol Model with M7 microphysics core) that is coupled to an aerosol composition module (E4C). We have used both models to simulate the global aerosol evolution and evaluate their performance against observational data.
The HAM-M7 module includes representations of aerosol processes such as new particle formation, emissions, sedimentation, deposition, and microphysical interactions across seven log-normal modes, including both mass and number concentrations as size-resolved properties for key aerosol species, including sulphate, black carbon, organic matter, sea salt, and dust supplemented with E4C compounds nitrate and ammonium. Furthermore, current implementation within OpenIFS Cy48r1 includes aerosol interactions with radiation and cloud microphysics. However, within IFS Cy50r1 only the coupling with radiation is included due to the expected strong influence of cloud activation on the forecast.
Models are run for one year (2018) with either CMIP (OpenIFS) or CAMS (IFS) emissions with one year of spinup. The simulated aerosol fields are compared with observed number and mass concentrations from the ACTRIS observational network as well as earlier simulations with the chemical transport model TM5. Furthermore, the simulated aerosol budgets and surface concentrations are compared with those provided by the aerosol models within the AeroCom project.
This work was supported by the European Union’s Horizon Europe projects CAMAERA - CAMS AERosol Advancement (number 101134927), CAMS2_35bis project and FOCI, Non-CO2 Forcers and Their Climate, Weather, Air Quality and Health Impacts (number 101056783).
How to cite: Bergman, T., Holopainen, E., Wu, L., Kokkola, H., Laakso, A., Halonen, H., Juurikkala, K., Le Sager, P., van Noije, T., Huijnen, V., Checa-Garcia, R., Tsikerdekis, A., Hill, A., Köhler, M., Rémy, S., and Metzger, S.: Evaluation of HAM-M7 within the ECMWF IFS and OpenIFS frameworks, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-12282, https://doi.org/10.5194/egusphere-egu26-12282, 2026.
Heavy rain episodes in the midlatitudes turn often into hazardous natural disasters, causing flooding events, infrastructure damage, and environmental repercussions. While the synoptic processes responsible for heavy rain are usually well understood, the aerosol–meteorology feedbacks sometimes remain uncertain despite their tremendous role in such episodes. The modeling study is conducted to identify key aerosol–cloud interactions during two heavy rain episodes that occurred in Europe in 2023: one over the Italy region in May 2023 and another over the Black Sea – Ukraine region in November 2023. The simulations were performed employing the Environment – HIgh-Resolution Limited Area Model (Enviro-HIRLAM) run at 15, 5 and 2 km horizontal resolutions for two model configurations: a reference run without aerosol effects (REF) and a run including indirect aerosol effects (IDAE).
Simulations with the IDAE configuration showed a significant increase in specific cloud ice and liquid water, a higher fraction of low-tropospheric cloud cover (but a reduction above 5 km), and an overall increase in total cloud condensate. Despite these changes, accumulated precipitation generally decreased by up to 3–4 mm per 6 h intervals when the aerosol effects were included. At finer spatial resolutions, localized areas with enhanced precipitation were identified, although the patterns differed regionally. In particular, in Italy, higher precipitation occurred mainly over marine areas, while in Ukraine it appeared predominantly over the land surface compared with the Black Sea aquatoria. Spatial correlation analysis between the aerosol fields and the differences between REF and IDAE configurations indicated that the dominant drivers of increased cloud water, ice content, and low-level cloud cover are soluble sulfate and sea-salt particles in the accumulation and coarse modes. Their influence was most pronounced within the 2–4 km layer. Less strong, although still significant, positive correlations were identified between the cloud cover and water/ice content and the presence of soluble coarse-mode dust, black carbon, and organic carbon. At finer resolution, these correlations weaken substantially, although the relationships for sulfate and sea-salt particles in the accumulation and coarse modes remained detectable. At the same time, no direct relationship was detected between aerosol fields and total precipitation, nor with the general decrease in rainfall in the IDAE relative to REF run. This suggests that complex atmospheric feedbacks govern precipitation formation, and that increased cloud water content does not necessarily translate into increased surface rainfall.
This study was conducted within the Horizon Europe programme under Grant Agreement No 101137680 CERTAINTY project (Cloud-aERosol inTeractions & their impActs IN The earth sYstem). The required infrastructure, computing and storage resources, and technical support were provided by the CSC – IT Center for Science (Finland) under PEEX Modelling Platform research and development through CSC HPC research projects (PEEX-MP-at-CSC). Initial and boundary conditions for meteorology, chemistry, aerosols, observations for data assimilation, other input required for the Enviro-HIRLAM simulations were provided by the European Centre for Medium-range weather Forecasting (ECMWF).
How to cite: Hrama, D., Pysarenko, L., Nadtochii, L., Rudas, M., Savenets, M., Mahura, A., and Petäjä, T.: Enviro-HIRLAM model simulations of aerosol–cloud interactions during two cases of heavy rain in Italy and Ukraine, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-520, https://doi.org/10.5194/egusphere-egu26-520, 2026.
Ozone (O3) is a secondary pollutant in the atmosphere formed by photochemical reactions that endangers human health and ecosystems. Since the mid-1990s, Asian regions have experienced the fastest O3 increase rate of 2–8 ppb per decade at remote surface sites and in the lower free troposphere across the world. Therefore, a deeper understanding of the long-term changes and causes of tropospheric O3 concentrations is of significance in both the environment and climate policy making.
In this study, to quantify the impacts of future climate change on O3 pollution, near-surface O3 concentrations over Asia in 2020–2100 are projected using a machine learning (ML) method along with multi-source data. The ML model is trained with assimilated O3 data from a global atmospheric chemical transport model and real-time observations. The ML model is then used to predict future O3 with meteorological fields from CMIP6 multi-model simulations under various climate scenarios. The climate penalty on future O3 is robust over most regions of Asia. The near-surface O3 levels are projected to increase by 5 %–20 % over South China, Southeast Asia, and South India under the high-forcing scenarios in the last decade of 21st century, compared to the first decade of 2020–2100. We also find that the summertime O3 pollution over eastern China will expand from North China to South China and extend into the cold season in a warmer future.
Unlike the traditional “black box” ML models, we predict near‐surface O3 concentrations in China in 2030 and 2060 based on a process‐based interpretable ML method, integrated with physical and chemical processes of O3, natural emissions of O3 precursors, and other multi‐source data. The direct (via changing physical and chemical processes of O3) and indirect (via changing natural emissions of O3 precursors) impacts of future climate change on O3 concentrations are quantitatively analyzed. The results suggest that the climate‐driven O3 levels are projected to decrease by more than 0.4 ppb in 2060 over eastern China under a carbon neutral scenario relative to a high emission scenario. The physical and chemical processes under climate change play a more important role in regulating O3 concentrations than natural emissions in the future under the carbon neutral scenario.
How to cite: Li, H.: Projecting future climate change impacts on ozone pollution with machine learning, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-5011, https://doi.org/10.5194/egusphere-egu26-5011, 2026.
Reanalyses of atmospheric composition (AC) combine atmospheric models with satellite retrievals to produce consistent, long‑term, gridded datasets that are widely used to assess trends and variability in atmospheric composition and air quality. However, changes in the availability of the assimilated observations can introduce artificial discontinuities, complicating the interpretation of long‑term trends. Bias correction and careful selection of the assimilated datasets are therefore essential to ensure temporal consistency.
The Copernicus Atmosphere Monitoring Service (CAMS) global AC reanalysis (EAC4) assimilates multiple retrievals of aerosol optical depth, ozone, and nitrogen dioxide, as well as total column carbon monoxide (TCCO) from the MOPITT instrument—the sole CO data source in the system. EAC4 spans 2003 to near‑present and has been extensively used for reporting AC anomalies and trends. The termination of MOPITT operations in January 2025 resulted in a substantial shift in CO fields in the subsequent EAC5 reanalysis, preventing its direct use for diagnosing TCCO anomalies in 2025.
To address this discontinuity, we developed a machine‑learning‑based method to emulate the MOPITT‑driven assimilation impact in EAC4. The ML model predicts monthly mean TCCO fields by learning the relationship between EAC4 and a corresponding control simulation without data assimilation. The control simulation uses the same meteorological fields and the same emissions as EAC4, including the CO wildfire emissions that dominate interannual variability.
We evaluate the agreement between EAC4 TCCO trends and annual anomalies and their ML‑based predictions. We also discuss alternative approaches for deriving TCCO anomalies for 2025, such as the use of the control simulation alone or the TCCO analysis from the operational CAMS forecasting system.
This work represents an initial step toward emulating AC data assimilation using machine learning, with the broader aim of improving the robustness of long‑term AC datasets in the presence of observational gaps.
How to cite: Flemming, J., Harder, P., and Inness, A.: Extending the CAMS Carbon Monoxide Reanalysis after the Loss of MOPITT: An ML‑Based Approach, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-13979, https://doi.org/10.5194/egusphere-egu26-13979, 2026.
In this study, simluation performance was evaluated using the WRF-CHEM model coupled with the Vegetation Photosynthesis and Respiration Model(VPRM) module, which accounts for the Net Echosystem Exchange (NEE) by terrestrial ecosystems. In the WRF-VPRM model, carbon dioxide (CO2) concentrations are simulated by dividing them into three components: background, anthropogenic, and biogenic.
To execute the WRF-VPRM model, various input datasets were utilized, including ERA5 meteorological fields, CAMS CO2 concentrations, EDGAR CO2 emission data, and MODIS surface reflectance data. To obtain high-resolution results for the Korean Peninsula, a nested grid system was configured, covering East Asia (27 km spatial resolution) and the Korean Peninsula (9 km spatial resolution). Fore Vegetation information - a critical input for VPRM - the Enhanced Vegetation Index (EVI) and Land Surface Water Index (LSWI), which represent vegetation activity and soil moisture derived from MODIS observations, were employed.
To evaluate the model's preformance, including seasonal variability, simulations were conducted over a three-year period from 2018 to 2020 for the East Asian region. Validation against ERA5 (meteorological variables) and CAMS data (CO2 concentration) usde for initial and boundary conditions confirmed an annual trend of increasing CO2 concentrations over the Korean Peninsula. Furthermore, the WRF-VPRM model successfully captured seasonal variability, showing lower concentrations during the summer - when vegetation effects are most prominent - compared to other seasons. This presentation introduces the validation of meteorological variable simulations and the analysis of characteristics related to CO2 concentration fluctuations.
How to cite: Chung, C.-Y., Kim, D.-H., Sung, H. M., Kim, M., Kim, T.-J., and Boo, K.-O.: Simulation and Evaluation of CO2 Concentration Variability over the Korean Peninsula using the WRF-VPRM Model., EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-18833, https://doi.org/10.5194/egusphere-egu26-18833, 2026.
The Integrated Forecasting System with atmospheric composition extension (IFS-COMPO) 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. The IFS-COMPO system is composed of tropospheric and stratospheric aerosol and chemistry components which are deeply intertwined. The composition model is updated regularly, aligned with updates of ECMWF’s operational meteorological model. Here we report on updates planned for the operational version after next, referred to as CY51R1. This concerns revisions on a large range of topics, as developed over the recent years, and therefore impacting many aspects of chemistry and aerosol composition in troposphere and stratosphere. The main aspects of the proposed upgrade concern:
- The representation of the life cycle desert dust, with the implementation of 6 size bins instead of three, and a new emission scheme that takes into better account high latitude dust sources,
- An updated representation of aerosol dry deposition,
- Optimized and online stratospheric photolysis, taking into the account the impact of clouds and aerosols,
- Use of prognostic aerosol information in stratospheric heterogeneous chemistry instead of climatological information,
- Update of Isoprene degradation chemistry and introduction of a tracer for ethyne and associated chemistry.
In this contribution we provide an overview of expected changes with emphasis on changes in composition modeling aspects and particulary the desert dust updates. We will present their expected impact on key atmospheric composition aspects, including air quality performance across major pollution regions across the world, aerosol optical depth, dust, and stratospheric composition products.
How to cite: Remy, S., Huijnen, V., Chabrillat, S., Minganti, D., Metzger, S., Russo, E., and Flemming, J.: Changes to the IFS-COMPO atmospheric composition mode in support to the CAMS update to cycle 51R1, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-11014, https://doi.org/10.5194/egusphere-egu26-11014, 2026.
Air pollution is one of the most critical environmental threats to human health and ecosystems, with major socio-economic impacts, and remains the leading environmental health risk in Europe, contributing to hundreds of thousands of premature deaths each year. Accurate one-day-ahead air quality forecasts are therefore essential to support timely mitigation actions and protect vulnerable populations under rapidly evolving atmospheric conditions.
This work develops and evaluates machine learning approaches for next-day prediction of PM₁₀ concentrations in Portugal, focusing on the Centro (NUTS II) region over the period 2003–2022. Two architectures were implemented and tested: a multilayer perceptron (MLP) and a deep learning long short-term memory (DL-LSTM) model, trained and cross-validated on data from 2003–2021, with 2022 reserved as an independent test year.
Model skill was assessed both for routine conditions and during two well-documented extreme events: the 2020 Oleiros wildfires and the 2022 Serra da Estrela wildfires, which produced intense PM₁₀ episodes in central Portugal. The models showed high predictive capability for daily PM₁₀, with the MLP achieving a correlation coefficient of 0.97 and slightly outperforming the DL-LSTM configuration.
These results highlight the potential of data-driven methods to anticipate short-term air quality degradation, including wildfire-driven pollution peaks, and to support operational warning systems at the regional scale. The proposed framework can be extended to other pollutants and regions, contributing to more effective environmental management and public health planning in Portugal.
How to cite: Oliveira, A., Brito, A., Durão, R., and Russo, A.: Data-driven one-day-ahead PM₁₀ prediction for Portugal: comparing MLP and LSTM models under extreme fire event, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-20341, https://doi.org/10.5194/egusphere-egu26-20341, 2026.
The Quick Look Content (QLC) suite provides automated forecast verification and analysis capabilities optimized for the Copernicus Atmosphere Monitoring Service (CAMS) and emerging AI-integrated forecasting systems. QLC addresses the growing need for systematic, reproducible evaluation of atmospheric composition forecasts through an end-to-end workflow from data retrieval to publication-quality visualizations.
The system integrates direct access to ECMWF's MARS archive with currently 16 observation networks including EBAS, AirNow, and GHOST-harmonized datasets, covering 7,855 atmospheric variables. Native GRIB support preserves forecast step information critical for analyzing temporal forecast evolution. QLC handles the complete verification workflow: automated MARS data retrieval, model-observation collocation with configurable spatial and temporal matching, statistical analysis including bias, RMSE, and correlation metrics, and generation of comprehensive visualizations (e.g., maps, time series, scatter plots, Taylor diagrams).
Recent development of an AIFS-specific workflow enables systematic comparison of AI-integrated forecasts against traditional IFS-COMPO runs and observational data. QLC supports multiple evaluation modes: single experiment validation, multi-experiment intercomparison, and observation-only analysis. Processing scales from single-station quick looks to continental-scale multi-variable assessments. Integration with the evaltools package (CNRM/Météo-France) provides advanced statistical diagnostics including diurnal cycles, station score maps, and exceedance analysis.
Here we introduce QLC and demonstrate its capabilities through verification examples comparing IFS-COMPO and AIFS-COMPO forecasts for ozone, nitrogen oxides, and particulate matter against multi-network observations. Results highlight the tool's utility for operational forecast monitoring, model development support, and scientific analysis. The open-source package (PyPI: rc-qlc) is designed for use on both HPC systems and workstations, with one-command installation and comprehensive documentation at docs.researchconcepts.io/qlc.
This work is supported by CAMS2_35_bis_KNMI: "Developments on reactive gases and aerosol in the global system" (https://atmosphere.copernicus.eu/).
How to cite: Metzger, S., Feigel, G., Karathanasopoulos, O., Myriokefalitakis, S., Elias, T., Rémy, S., Huijnen, V., Wing Yi Li, C., Harder, P., and Flemming, J.: QLC: An Automated Forecast Verification Suite for CAMS and AI-Integrated Weather Prediction Systems, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-20963, https://doi.org/10.5194/egusphere-egu26-20963, 2026.
We report on the development of an inversion system for biogenic emissions within ECMWF’s Integrated Forecasting System IFS-COMPO. As part of the Horizon Europe CAMEO (CAMS EvOlution) project, a satellite-retrieval based inversion system for surface fluxes of biogenic volatile organic compounds was implemented in the IFS global model. The scheme is based on formaldehyde (HCHO) satellite observations from multiple satellite instruments. As part of the work, an assimilation capacity for formaldehyde was developed for use in IFS-COMPO, which uses the 4DVAR data assimilation technique. The extension for HCHO assimilation applies the tangent linear and adjoint of recently developed simplified formaldehyde-isoprene chemistry scheme. The purpose of the adjoint simplified chemistry scheme is to enable a modification of the isoprene fields based on the assimilation of HCHO observations. The impact of the assimilation of HCHO in IFS-COMPO is analyzed using TROPOMI S5P formaldehyde observations, with a particular focus on its impact on HCHO as well as key atmospheric reactive trace gases, such as isoprene, ozone, and carbon monoxide. We further present first optimizations of the a priori biogenic isoprene emissions based on HCHO satellite observations. Particular focus is put on the sensitivity of the implemented simplified HCHO chemistry to isoprene emissions by systematically investigating the link of isoprene and formaldehyde both in the IFS standard configuration and when using the new simplified HCHO chemistry scheme. The latter is assessed by scaling climatological a-priori isoprene emissions by differing factors and consecutively analysing the resulting variance of full-IFS-COMPO and simplified-chemistry formaldehyde in respective IFS-COMPO forecasts.
How to cite: Kluge, F., Flemming, J., Huijnen, V., Inness, A., Kelly, C., Müller, J.-F., Oomen, G.-M., Ribas, R., Stavrakou, T., and van der Worp, M.: Development of an inversion system for biogenic isoprene emissions in IFS-COMPO, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-18359, https://doi.org/10.5194/egusphere-egu26-18359, 2026.
Accurate real-time prediction of surface ozone at high spatial and temporal resolution is critical for air-quality management, exposure assessment, and public health protection. However, developing an operational hourly ozone prediction system at the national scale remains challenging due to the need for continuous data acquisition, integration of heterogeneous data sources, and computationally efficient modeling frameworks. This study presents a real-time, data-driven framework for high-resolution hourly ozone prediction across South Korea by constructing an automated, nationwide atmospheric database that integrates satellite, meteorological, and ground-based observations.
We establish an operational data pipeline that collects and processes near-real-time meteorological observations from the Korea Meteorological Administration’s Operational Data Assimilation and Model (ODAM), atmospheric information derived from the GK2A geostationary satellite, and surface ozone measurements from the AIRKOREA monitoring network. All data streams are ingested on an hourly basis and systematically harmonized through temporal synchronization and spatial alignment to generate high-resolution predictors covering the entire Korean Peninsula. This integrated database enables consistent representation of rapidly evolving meteorological conditions, atmospheric composition, and surface-level air quality.
Using the constructed real-time database, we develop a data-driven prediction model to estimate hourly surface ozone concentrations, with AIRKOREA observations used as the target variable. The modeling framework is designed with operational feasibility in mind, supporting continuous updates, automated preprocessing, and near-real-time inference without reliance on computationally expensive chemical transport models. The resulting system provides high-resolution ozone predictions that capture fine-scale spatiotemporal variability at the national level.
Model evaluation demonstrates that integrating geostationary satellite data with real-time meteorological and surface observations substantially enhances the prediction of hourly ozone variability compared to single-source or static-input approaches. The proposed framework highlights the advantages of real-time, high-resolution, and nationwide data integration for operational ozone forecasting in South Korea. Beyond ozone, this scalable and extensible system provides a foundation for real-time prediction of additional atmospheric pollutants and supports the development of next-generation data-driven air-quality forecasting services.
How to cite: Kim, S., Kim, Y., and Lee, W.: A real-time, high-resolution framework for nationwide hourly ozone prediction in South Korea using integrated satellite, meteorological, and surface observations, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-18034, https://doi.org/10.5194/egusphere-egu26-18034, 2026.
Posters virtual: Tue, 5 May, 14:00–18:00 | vPoster spot 5
EGU26-12255 | Posters virtual | VPS3
Assimilation of lidar and ceilometer observations from the E-profile network of European ground-based stations into ECMWF’s Integrated Forecasting System (IFS-COMPO)Tue, 05 May, 14:03–14:06 (CEST) vPoster spot 5