Stefan Kaufmann

Contrails contribute significantly to the climate impact of aviation; however, their formation depends on fuel composition, engine type and meteorological conditions and current contrail formation models may be insufficiently constrained by observations. Here we revisit the 1D microphysics code MoMiE and compare its results with recent contrails measurements from the ECLIF and NEOFUELS/VOLCAN campaigns. The sensitivity of the model to different Fuel Sulfur Contents (FSC), from 0.5 to 195 ppmm, and to different engine combustion modes, i.e. lean-burn, intermediate and rich-burn, is studied, as well as to the emission index of organic compounds and ions, and to ambient temperature and relative humidity. For the first time, the recent NEOFUELS/VOLCAN data of total number of particles measured in the near-field are compared to model simulations. Adjusting the input parameters within their expected uncertainty ranges allows the model to approximate the data for different fuels and soot regimes, demonstrating a clear sensitivity to organic compounds. The model results are also compared to far-field measurements of ice crystals of ECLIF and NEOFUELS/VOLCAN. MoMiE generally captures the dependence of contrail ice crystals to non-volatile particle emissions and compares well with the measurements in the soot-rich regime. In the soot-poor regime, MoMiE results depend mainly on the ambient temperature and show little variability obtained for different FSCs. These results emphasize the need for further investigation on the role of the different parameters involved in the contrail formation, as well as the need for further measurements of the different plume components.

Lean-burn engines are designed to reduce emissions, and sustainable alternative fuels provide additional potential for lowering particle emissions. However, the magnitude of the benefits from these mitigation levers remains poorly constrained, limiting robust predictions of aviation’s climate impact. We present in-flight measurements of particle emissions from modern lean-burn LEAP-1A engines operated on Airbus A319neo and A321neo aircraft, burning conventional and sustainable aviation fuels. Emissions of nonvolatile particles larger than 14 nanometers from lean-burn engines were reduced by three orders of magnitude compared to rich-burn operation. Under rich-burn conditions, soot particle number emissions show an exponential dependence on fuel hydrogen content, with engine-specific variations. Volatile particle emissions scale with fuel sulfur content, and substantial differences were observed even between two engines of the same type. Optimizing fuel composition and combustor technology may reduce particle emissions, improve airport air quality, and mitigate contrail radiative forcing.

Chemically active chlorine species (ClOx) play a central role in the catalytic depletion of ozone in the polar winter and spring stratosphere. Together with the reservoir species HCl and ClONO2, they form inorganic chlorine (Cly). The amount and partitioning of Cly strongly influences the magnitude of polar ozone loss. Trends and variability in the upper troposphere lower stratosphere region (UTLS) are of particular importance, as this region is a large contributor to lower-stratospheric ozone change. We present new in-situ measurements from the HALO campaign ASCCI (Arctic Springtime Chemistry and Climate Investigations), obtained in spring 2025. The total organic chlorine (CCly) is directly derived from the major chlorine species measured by the Gas Chromatograph for Observational Studies using Tracers (GhOST). From these measurements, inorganic chlorine is derived, allowing an observational assessment of the chlorine budget in the Arctic UTLS. Additional in situ observations of the reservoir species HCl and ClONO2 from the Airborne Chemical Ionisation Mass Spectrometer (AIMS) are used to constrain the abundance of chemically active chlorine (ClOx) and to investigate chlorine activation and deactivation processes during spring 2025. The new ASCCI data are compared with observations from previous aircraft campaigns, including PGS (Arctic measurements in 2015) and SOUTHTRAC (Antarctic measurements from 2019) both performed during hemispheric springtime, allowing for an assessment of hemispheric differences between Arctic and Antarctic conditions as well as temporal changes in chlorine loading and partitioning. These comparisons place the new measurements in the context of declining stratospheric chlorine and ozone recovery.

The persistence of aircraft contrails and their climate impact are strongly controlled by the extent, lifetime, and properties of ice-supersaturated regions (ISSRs). Reliable prediction of such conditions remains challenging due to uncertainties in the representation of water vapor in numerical weather prediction models at typical cruising altitudes in the upper troposphere. In-situ airborne observations are therefore essential for evaluating model performance and assessing the potential benefit of additional humidity data sources.This study employs three complementary data sources: (1) high-resolution in-situ water vapor measurements obtained with the Sophisticated Hygrometer for Atmospheric Research (SHARC) and a modified Water Vapor Sensing System II (WVSS-II) aboard the High Altitude and Long-Range Research Aircraft (HALO); (2) routine aircraft observations from the Aircraft Meteorological Data Relay (AMDAR) program using WVSS-II sensors; and (3) numerical weather model output from the ICON-DREAM reanalysis of the German Weather Service (DWD).To evaluate the data quality of WVSS-II sensors on commercial aircraft, comparisons between HALO reference measurements and AMDAR observations with spatial overlap during the Arctic Springtime Chemistry Climate Investigations (ASCCI) field campaign are conducted. An indirect comparison uses the entire dataset to analyze water vapor concentration (H₂O) and relative humidity with respect to ice (RHi) as a function of potential temperature. In addition, a short parallel flight segment of HALO and a commercial flight at the same altitude was performed which allows for a direct comparison of both sensors. For both comparison approaches, agreement and variability between the datasets are assessed using statistical metrics. Furthermore, the HALO dataset is used to evaluate and quantify the representation of RHi in ICON first-guess fields, covering multiple campaign periods between 2012 and 2025 with a total of 878 flight hours.Building on the previous steps, ongoing work investigates the use of AMDAR humidity observations in dedicated ICON data assimilation experiments to evaluate their impact on humidity prediction. Differences between routine model simulations and assimilation runs are analyzed to assess potential improvements in upper-tropospheric humidity forecasts relevant for ISSR prediction.

Alternative aviation fuels represent a promising approach to reduce contrails climate effect. In the frame of VOLCAN (“VOL avec Carburants Alternatifs Nouveaux”) project (DGAC funding, collaboration with AIRBUS, Safran Aircraft Engines and DLR, financed by Neofuels), the influence of Sustainable Alternative Fuels (SAF) composition on exhaust plumes emission, and therefore on contrails, is investigated using the 1D detailed microphysical code MoMiE (Modèle Microphysique pour Effluents) developed at ONERA1,2. The VOLCAN measurement campaigns have been able to provide estimations of ice particle number emission indexes within contrails formed by different fuel types (classical kerosene Jet A-1 and biofuel HEFA) and different combustion modes (“rich” and “lean” burn). These are complementing the observations obtained for Sustainable Alternative Fuels with Emission and “CLimate Impact of alternative Fuels” (ECLIF) campaigns3,4, recently compared to the results of the “Aerosol and Contrail Microphysics” (ACM) model developed at the University of Albany5.In its most recent version ONERA’s code MoMiE has been adapted to Sustainable Alternative Fuels (SAF)2. It includes heterogenous freezing with soot activation by sulfur and organic species, as well as homogeneous freezing of liquid droplets of hydrated sulfates and organics, accounting for the competition between both nucleation modes. Chemiionization, brownian coagulation of particles, ice sublimation and condensation are also represented. The code computes the different aerosols distributions (size and number) of sulfates, organics, dry soot, activated soot, and ice particles, homogeneously (no solid nucleus) and heterogeneously (soot solid nucleus) formed.The work proposed here aims first at presenting and analyzing the results obtained with the model in comparison to some of the VOLCAN measurements. The sensitivity of contrail formation to the different fuel types, combustion modes and emission characteristics, as ion emission index, which is known to play a significant role in the coagulation process, are studied. The model is also confronted to the ECLIF measurements3,4 and the microphysics model results from University of Albany5. Advancement and results of this study will be presented and discussed during the conference.1Vancassel X. et al., Numerical simulation of aerosols in an aircraft wake using a 3D LES solver and a detailed microphysical model, International Journal of Sustainable Aviation, 20142Rojo C. et al., Impact of alternative jet fuels on aircraft-induced aerosols, Fuel, 20143Voigt C. et al., Cleaner burning aviation fuels can reduce contrail cloudiness, communications earth & environment, 20214Märkl R. S. et al., Powering aircraft with 100% sustainable aviation fuel reduces ice crystals in contrails, Atmospheric Chemistry and Physics, 20245Yu F. et al., Revisiting Contrail Ice Formation: Impact of Primary Soot Particle Sizes and Contribution of Volatile Particles, Environmental Science & Technology, 2024

Knowledge of humidity in the upper troposphere and lower stratosphere (UTLS) is of special interest due to its importance for cirrus cloud formation and its climate impact. However, the UTLS water vapor distribution in current weather models is subject to large uncertainties. Here, we develop a dynamic-based humidity correction method using an artificial neural network (ANN) to improve the relative humidity over ice (RHi) in ECMWF numerical weather predictions. The model is trained with time-dependent thermodynamic and dynamical variables from ECMWF ERA5 and humidity measurements from the In-service Aircraft for a Global Observing System (IAGOS). Previous and current atmospheric variables within ±2 ERA5 pressure layers around the IAGOS flight altitude are used for ANN training. RHi, temperature, and geopotential exhibit the highest impact on ANN results, while other dynamical variables are of low to moderate or high importance. The ANN shows excellent performance, and the predicted RHi in the UT has a mean absolute error (MAE) of 5.7 % and a coefficient of determination (R²) of 0.95, which is significantly improved compared to ERA5 RHi (MAE of 15.8 %; R² of 0.66). The ANN model also improves the prediction skill for all-sky UT/LS and cloudy UTLS and removes the peak at RHi = 100 %. The contrail predictions are in better agreement with Meteosat Second Generation (MSG) observations of ice optical thickness than the results without humidity correction for a contrail cirrus scene over the Atlantic. The ANN method can be applied to other weather models to improve humidity predictions and to support aviation and climate research applications.

Convective processes play a critical role in the atmosphere’s energy balance. In high updraft regimes, turbulent and diabatic processes redistribute moisture, heat, and aerosols, which lead to cloud formation affecting the radiation budget of the atmosphere.During the HALO airborne CIRRUS-HL mission in summer 2021, the outflow of a convective system over Northern Italy was probed at different altitudes. The system had an overshooting top accompanied by lightning and icing conditions. A suit of in-situ (aerosol, cloud probes, trace gases) and remote sensing (Lidar) instruments deployed on the research aircraft HALO combined with satellite observations provided the opportunity to investigate the system from different perspectives.The in-situ H2O -O3 correlation revealed unexpected insights in the extra-tropical tropopause transition layer (exTL), characterized by enhanced water vapor and ice crystal number in the exTL. The convective system penetrated into the exTL with O3 up to 450ppb. In contrast, the CO -O3 correlation shows minor influence, indicating that this convection event was little impacted by large scale mixing processes.The potential temperature around the upper cloud edge ranged from 330K to 350K. At higher potential temperatures (377-392K) no H2O enhancements were observed. Nevertheless, the irreversible injection of water vapor could lead to transport of moisture into the lower stratosphere in the following hours and days downwind of the system.Within the upper cloud part and in the vicinity of the cloud, water vapor and ice crystals are enhanced in comparison to the undisturbed surrounding, as visible in the Lidar curtain. Both, water vapor and ice crystals influence the hydration and dehydration of the exTL. While larger ice crystals sediment, smaller ice crystals may sublimate and contributing to a locally enhanced water vapor budget. Our measurements show, that strong convective systems can act as a potential moisture source of the lowermost stratosphere.

Knowledge of humidity in the upper troposphere and lower stratosphere (UTLS) is of special interest due to its importance for cirrus cloud formation and its climate impact. However, the UTLS water vapor distribution in current weather models is subject to large uncertainties. Here, we develop a dynamic-based humidity correction method using artificial neural network (ANN) to improve the relative humidity over ice (RHi) in ECMWF numerical weather predictions. The model is trained with time-dependent thermodynamic and dynamical variables from ECMWF ERA5 and humidity measurements from the In-service Aircraft for a Global Observing System (IAGOS). Previous and current atmospheric variables within ±2 ERA5 pressure layers around the IAGOS flight altitude are used for ANN training. RHi, temperature and geopotential exhibit the highest impact on ANN results, while other dynamical variables are of minor importance. The ANN shows excellent performance and the predicted RHi in the UT has a mean absolute error MAE of 6.6 % and a coefficient of determination R² of 0.93, which is significantly improved compared to ERA5 RHi (MAE of 15.7 %; R² of 0.66). The ANN model also improves the prediction skill for all sky UT/LS and cloudy UTLS and removes the artificial peak at RHi = 100 %. The contrail predictions are in better agreement with MSG observations of ice optical thickness than the results without humidity correction for a contrail cirrus scene over the Atlantic. The ANN method can be applied to other weather models to improve humidity predictions and to support aviation and climate research applications.

Knowledge of humidity in the upper troposphere and lower stratosphere (UTLS) is of special interest due to its importance for cirrus cloud formation and its climate impact. However, the UTLS water vapor distribution in current weather models is subject to large uncertainties. Here, we develop a dynamic-based humidity correction method using artificial neural network (ANN) to improve the relative humidity over ice (RHi) in ECMWF numerical weather predictions. The model is trained with time-dependent thermodynamic and dynamical variables from ECMWF ERA5 and humidity measurements from the In-service Aircraft for a Global Observing System (IAGOS). Previous and current atmospheric variables within ±2 ERA5 pressure layers around the IAGOS flight altitude are used for ANN training. RHi, temperature and geopotential exhibit the highest impact on ANN results, while other dynamical variables are of minor importance. The ANN shows excellent performance and the predicted RHi in the UT has a mean absolute error MAE of 6.6 % and a coefficient of determination R² of 0.93, which is significantly improved compared to ERA5 RHi (MAE of 15.7 %; R² of 0.66). The ANN model also improves the prediction skill for all sky UT/LS and cloudy UTLS and removes the artificial peak at RHi = 100 %. The contrail predictions are in better agreement with MSG observations of ice optical thickness than the results without humidity correction for a contrail cirrus scene over the Atlantic. The ANN method can be applied to other weather models to improve humidity predictions and to support aviation and climate research applications.

Knowledge of humidity in the upper troposphere and lower stratosphere (UTLS) is of special interest due to its importance for cirrus cloud formation and its climate impact. However, the UTLS water vapor distribution in current weather models is subject to large uncertainties. Here, we develop a dynamic-based humidity correction method using artificial neural network (ANN) to improve the relative humidity over ice (RHi) in ECMWF numerical weather predictions. The model is trained with time-dependent thermodynamic and dynamical variables from ECMWF ERA5 and humidity measurements from the In-service Aircraft for a Global Observing System (IAGOS). Previous and current atmospheric variables within ±2 ERA5 pressure layers around the IAGOS flight altitude are used for ANN training. RHi, temperature and geopotential exhibit the highest impact on ANN results, while other dynamical variables are of minor importance. The ANN shows excellent performance and the predicted RHi in the UT has a mean absolute error MAE of 6.6 % and a coefficient of determination R² of 0.93, which is significantly improved compared to ERA5 RHi (MAE of 15.7 %; R² of 0.66). The ANN model also improves the prediction skill for all sky UT/LS and cloudy UTLS and removes the artificial peak at RHi = 100 %. The contrail predictions are in better agreement with MSG observations of ice optical thickness than the results without humidity correction for a contrail cirrus scene over the Atlantic. The ANN method can be applied to other weather models to improve humidity predictions and to support aviation and climate research applications.