AS3.2

Enhanced emissions of both greenhouse gases and aerosols generate climate responses on a wide range of time scales. An initial radiative response triggers a set of rapid adjustments, which are eventually followed by surface-temperature-driven feedbacks. While a lot happens during the first days and months after a perturbation, the monthly mean data typically used in climate studies are too coarse to show the temporal evolution of responses. In these analyses, we take a closer look at how the climate system responds during the very first hours and days after a sudden increase in carbon dioxide (CO2), in black carbon (BC) or in sulfate (SO4). Five models have performed PDRMIP simulations with hourly output, and we also compare results to monthly PDRMIP and CMIP6 results. We find that the effect of increasing ocean temperatures kicks in after a couple of months. Rapid precipitation reductions are for all three climate perturbations established after just a couple of days, and does for BC not differ much from the full-time response.  For CO2 and SO4, the magnitude of the precipitation response gradually increases with surface warming, and for CO2 the sign of the response changes for negative to positive after two years. Rapid cloud adjustments are typically established within the first 24 hours and while the magnitude of cloud feedbacks for CO2 and SO4 increases over time, the latitude-height pattern of the total cloud changes is clearly present after one year. While previously known that climate responses to BC are dominated by rapid adjustments, this work underlines the swiftness of the processes involved.

How to cite: Myhre, G., Stjern, C., Samset, B., Forster, P., Quaas, J., Takemura, T., Voulgarakis, A., Jia, H., Jouan, C., Sand, M., and Olivie, D.: The timescales of climate responses to carbon dioxide and aerosols, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-4108, https://doi.org/10.5194/egusphere-egu22-4108, 2022.

• We find that the ocean heat transport decreases significantly in the Northern Hemisphere extratropics, which is counter-intuitive given the increasing temperature gradient between the tropics and the polar regions due to the enhanced aerosol emissions.
• When doubling SO₂ emissions, we find increases in Southern Hemisphere sea surface temperatures in contrast to cooling in the Northern Hemisphere. 

How to cite: Pietschnig, M., Olivié, D., Moseid, K. O., Hofer, S., Madan, G., Lacasce, J. H., and Storelvmo, T.: Effects of increased aerosol emissions over Asia on global sea-surface temperatures, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-13274, https://doi.org/10.5194/egusphere-egu22-13274, 2022.

The Sun’s radiation is the primary energy source for chemical, biological, and physical processes that happen in the climate system. Thus, the earth-atmosphere radiative balance system is one of the main aspects of climate change. Natural fluctuations in incident solar radiation caused by the sunspot cycle can influence the energy balance. In addition, human activities can also affect this balance. Changes in gases and aerosols emissions to the atmosphere can modify its composition because they are involved in complex chemical reactions, such as ozone concentrations. Gases and aerosols can absorb, scatter, and reflect incident solar radiation, thereby affecting the balance. This study assesses the direct impact of aerosol (through changing the optical depth of aerosol-AOD in the radiation subroutine) on the surface atmospheric temperature and radiation balance. Three simple experiments for 1998-2017 were carried out through numerical modeling using BAM-v1.2 (Brazilian Global Atmospheric Model), the operational weather and climate forecasting model at CPTEC/INPE (Center for Weather Forecasting and Climate Studies/National Institute for Space Research). These experiments were zero aerosols, fixed AOD over the land and ocean, and AOD climatology with a spatial and temporal variation (AOD-C). The surface atmospheric temperature was validated against the ERA5 reanalysis from December to February (DJF) and June to August (JJA). Also, the downward shortwave solar radiation on the clear-sky variable was validated against CERES-EBAF satellite data. We performed the bias, the difference between the model and the reanalysis data (ERA5) and EBAF-CERES, correlation and RMSE of the model results against ERA5 and EBAF-CERES for surface temperature and downward shortwave solar radiation on clear-sky respectively. Our results have shown a positive bias atmospheric surface temperature for the northern hemisphere continent and a negative bias for the southern hemisphere continent during JJA. We observed a decrease in this positive bias in the northern hemisphere in the experiments with fixed aerosols, but an important improvement (vies, correlation, and RMSE) was observed in the experiment with AOD-C. On the other hand, during the DJF period, the model has a positive bias only in some continental areas, such as southwestern South America and South Africa, North Africa, and the Australian continent. Similarly, to JJA, we observed improvements in these regions in the experiments that use fixed and AOD-C. The downward shortwave solar radiation on clear-sky results for both DJF and JJA showed an inversion from the positive bias to a negative bias in the model version without aerosols to the model with fixed aerosols and AOD-C, due to the presence of the aerosol, which reduces short wave flow. An important improvement (vies, correlation, and RMSE) in the downward shortwave solar clear sky was observed in the version that uses AOD-C during JJA.

How to cite: Herdies, D., Alvim, D., Basso, L., Pendharkar, J., Castilho, D., Oyerinde, G., Costa, S., Kubota, P., and Figueroa, S.: Validadion of three Brazilian Global Atmospheric Model experiments (without, fixed and monthly climatological aerosol) against ERA5 and CERES-EBAF, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-6250, https://doi.org/10.5194/egusphere-egu22-6250, 2022.

Uncertainty in cloud feedback in climate models is a major limitation in projections of future climate. We analyse cloud biases and trends in climate models relative to satellite observations, and relate them to equilibrium climate sensitivity, transient climate response and cloud feedback. For this purpose, we develop a deep convolutional artificial neural network for determination of cloud types from low-resolution daily mean top of atmosphere shortwave and longwave radiation images, corresponding to the World Meteorological Organization (WMO) cloud genera recorded by human observers in the Global Telecommunication System. We train this network on a satellite top of atmosphere radiation dataset, and apply it on the Climate Model Intercomparison Project phase 5 and 6 (CMIP5 and CMIP6) historical and abrupt-4xCO2 experiment model output and the ERA5 and Modern-Era Retrospective Analysis for Research and Applications Version 2 (MERRA-2) reanalyses. We compare these with satellite observations, link cloud type occurrence biases and trends to climate sensitivity, and compare our cloud types with an existing cloud regime classification based on the Moderate Resolution Imaging Spectroradiometer (MODIS) and International Satellite Cloud Climatology Project (ISCCP) satellite data. We show that there is a strong linear relationship between the root mean square error of cloud type occurrence and model equilibrium climate sensitivity, transient climate response and cloud feedback (Bayes factor 7×10², 4×10² and 13, respectively). This indicates that models with a better representation of the cloud types have a more positive cloud feedback and higher climate sensitivity. Along with other studies, our results point to a choice between two explanations: either high sensitivity models are plausible, contrary to combined assessments of climate sensitivity and cloud feedback in previous review studies, or the accuracy of representation of present-day clouds in models is negatively correlated with the accuracy of representation of future projected clouds.

How to cite: Kuma, P., Bender, F., Schuddeboom, A., and McDonald, A.: Cloud type machine learning shows better present-day cloud representation in climate models is associated with higher climate sensitivity, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-4247, https://doi.org/10.5194/egusphere-egu22-4247, 2022.

In this contribution, statistical and machine learning techniques are applied to quantify the response of clouds to changes in environmental factors.

Clouds play a key role for the Earth’s energy balance; however, their response to climatic and anthropogenic aerosol emission changes is not clear, yet. Here, 20 years of satellite cloud observations are analyzed together with reanalysis data sets in multivariate-regression and machine-learning approaches to quantitatively link the variability of observed cloud radiative effects to changes in environmental factors, or cloud-controlling factors (CCFs). In this data-driven approach, a large number of CCFs, including aerosol proxies, are used as predictors at a large spatial and temporal scale typical of CCF analyses. The analysis reveals distinct regional patterns of CCF importance for shortwave and longwave cloud radiative effects. In stratocumulus cloud regions, the main controls of shortwave CRE are the sea surface temperature and the estimated inversion strength, but also zonal winds in the lower free troposphere are relevant controls of CRE. Aerosol proxies are shown to be most important for shortwave CRE in the regions of stratocumulus to cumulus transition. Future analyses of interactions between different CCFs and comparisons to global climate models are outlined.

How to cite: Andersen, H. and Cermak, J.: A data-driven approach to understanding global controls of cloud radiative effects, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-11653, https://doi.org/10.5194/egusphere-egu22-11653, 2022.

A wide range of aerosol effects on precipitation have been proposed, from the scale of individual clouds to that of the globe.

This presentation, based on the findings of an expert workshop under the umbrella of the GEWEX Aerosol Precipitation initiative, reviews the evidence and scientific consensus behind these effects and the underlying set of physical mechanisms, categorised into i) radiative effects via modification of radiative fluxes and the energy balance and ii) microphysical effects via modification of cloud droplets and ice crystals.

There exists broad consensus and strong theoretical evidence that, because global mean precipitation is constrained by energetics and surface evaporation, aerosol radiative effects (aerosol-radiation interactions and aerosol-cloud interactions) act as drivers of precipitation changes. Likewise, aerosol radiative effects cause well-documented shifts of large-scale precipitation patterns, such as the Inter-Tropical Convergence Zone (ITCZ). The extent to which aerosol effects on precipitation are applicable at smaller scales and driven or buffered by compensating microphysical and dynamical mechanisms and budgetary constraints is less clear. Although there exists broad consensus and strong evidence that suitable aerosol perturbations increase cloud droplet numbers, reducing the efficiency of warm rain formation across cloud regimes, the overall aerosol effect on cloud microphysics and dynamics as well as the subsequent impact on local, regional and global precipitation is less constrained.

This presentation provides a review of the physical mechanisms of aerosol effects on precipitation backed up by evidence from recent cloud-resolving and global modelling simulations as well as from satellite observations.

How to cite: Stier, P. and the GEWEX Aerosol Precipitation (GAP) initiative expert workshop team: Multifaceted Aerosol Effects on Precipitation, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-2856, https://doi.org/10.5194/egusphere-egu22-2856, 2022.

In recent decades, WRF has been widely used in regional rainfall simulations, and many studies have shown it has good performance in reproducing rainfall distribution. However, the WRF simulated rainfall amounts are often significantly underestimated which might be due to insufficient consideration of aerosol-cloud-precipitation-meteorology interaction in its mdoelling. WRF-chem as a meteorology-chemistry coupling model is expected to improve such a shortcoming. In this study, we carry out a series of WRF and WRF-chem simulations of a large-scale extreme rainfall event (occurred from October 11th to 15th, 2018) over the UK to explore whether different aerosol effects could help improve the rainfall simulation performance.

To compare and evaluate the influences of different aerosol effects, four types of simulations using WRF and WRF-chem were conducted. The baseline simulation (called WF_B) was simulated by WRF without any emission data and chemical boundary conditions. The sensitivity simulation (WC_NE) was simulated by WRF-chem with emission data and chemical boundary conditions as well as used a chemical mechanism. But it turned off aerosol direct and indirect effects. The other two sensitivity simulations WC_DE and WC_DAIE were conducted by turning on the direct aerosol effect and turning on all (direct and indirect) aerosol effects, respectively. All simulations used the same domain configurations, physical schemes, and meteorological boundary conditions. Through comparing the difference between the four simulated rainfall distributions and amounts, the impact of aerosol direct effect, indirect effect, and net (direct + indirect) effect on extreme rainfall simulation were estimated.

The simulation results were compared with UK radar observations. The sensitivity study shows that the rainfall intensity performance greatly improved with the inclusion of the aerosol-cloud interaction in the modelling (indirect effect). However, aerosol-radiation feedback (direct effect) does not have a significant impact on rainfall intensity estimations. One of the reasons was because the aerosol indirect effect has a great influence on droplet/particle concentration, precipitation efficiency and cloud life in nature. Statistics show that there are 115 grids in radar observation with rainfall greater than 100 mm, while WF_B, WC_NE, WC_DE and WC_DAIE simulations have respectively 44, 44, 44 and 117 grids with rainfall greater than 100 mm. In addition, the Root Mean Square Error of WF_B, WC_NE, WC_DE and WC_DAIE accumulated rainfall is 2.501, 2.501, 2.484 and 0.779 respectively. On the other hand, the rainfall spatial performances of the four simulations are relatively close, which were not improved obviously with the inclusion of aerosol effects. Their probability of detection (POD), frequency bias index (FBI), critical success index (CSI), and false alarm ratio (FAR) performances were averaged at 0.941, 0.946, 0.936, and 0.006, respectively. Finally, using the chemical mechanism and chemical data but turning off aerosol effects resulted in similar rainfall estimations of WRF-chem and original WRF. In summary, it is highly recommended to turn on WRF aerosol effects, especially the indirect aerosol effects in extreme rainfall simulations.

How to cite: Liu, Y., Zhuo, L., and Han, D.: Investigating the influence of aerosol effects on extreme rainfall simulations over the UK using WRF and WRF-chem model, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-5393, https://doi.org/10.5194/egusphere-egu22-5393, 2022.

The Sichuan Basin, which is located in southwest China, has become one of the most polluted regions in China. The frequency of atmospheric boundary layer in the Sichuan Basin are analyzed based on radiosonde profile from 2014 to 2016. The occurrence probability of multi-layer temperature inversions is about 46.86% in winter over the Sichuan Basin. The radiosonde data are divided into four equal parts according to the previous 12-hour average visibility and PM_2.5 mass concentration. As the PM_2.5 mass concentration increases (visibility decreases), the surface-based inversion frequency increases more consistently in the morning, while the elevated inversion increases more consistently in the evening. The mechanisms of the aerosol radiative effect on the boundary layer temperature inversion are further investigated by using the radiometer, the Mie scattering lidar, and the microwave radiometer in Chengdu. 1D-SBDART simulation is performed to better clarify the mechanisms. The simulation results show that: from clean to heavy pollution conditions, in clear-sky the surface shortwave radiation reduces by 135.04 w·m^-2 and the heating rate increases by 0.75 k·d^-1; in cloudy sky the surface shortwave radiation reduces by 46.15 w·m^-2 and the heating rate increases by 0.35 k·d^-1. Aerosols can enhance the boundary layer temperature inversion at both daytime and nighttime due to the radiative effect, while clouds mitigate the enhancement by aerosol effects. 

How to cite: Jiang, M., Yang, Y., Ni, C., and Chen, Q.: The impact of aerosols on the temperature inversion in the boundary layer over southwest China, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-9554, https://doi.org/10.5194/egusphere-egu22-9554, 2022.

The long-lived and widespread nature of smoke, coupled with its ability to perturb the atmosphere simultaneously via aerosol-cloud and aerosol-radiation interactions, has proven a challenge to observe and simulate; as such, the impact of smoke on regional and global scales remains uncertain.

In this study we use an 18-year climatology from multiple instruments onboard AQUA and TERRA satellites to identify and characterise the relationships between aerosol-optical-depth (AOD) and the large-scale properties of the clouds, precipitation, and top-of-atmosphere radiation over the Amazon rainforest during the biomass burning season.

Our analysis provides robust evidence that localised smoke production drives widespread modification to the cloud regime over the region: in the morning (TERRA) cloud liquid water path increases with AOD, whereas in the afternoon (AQUA) convective activity is initially enhanced then supressed when AOD exceeds 0.4. During both time periods there is an increasingly pronounced presence of high-altitude, optically thin, clouds.

The result is a sharp contrast in the cloud-field properties and vertical distribution between low-AOD days and high-AOD days and a pronounced top-of-atmosphere radiative effect of -50 Wm­^-2 (for AOD = 1.4), which persists throughout the day.

How to cite: Herbert, R. and Stier, P.: Amazon fires drive widespread changes to diurnal cloud regimes and radiation, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-2863, https://doi.org/10.5194/egusphere-egu22-2863, 2022.

Aerosol–cloud interactions in marine stratocumulus clouds (Sc) are among the most challenging frontiers in cloud–climate research. In particular, the cloud cover susceptibility to droplet concentration remained under-represented in the literature. We developed methodologies to estimate what would have been the cloud cover and the associated radiative effect of currently observed Sc, but in a hypothetical cleaner world. The first methodology uses a realistic Lagrangian large eddy simulation coupled with satellite observations and provides a process-oriented analysis. The other uses a simple model and provides a global estimate of the radiative impact. We found that overcast Sc decks would have broken up sooner had they not been influenced by anthropogenic aerosol, thereby causing a significant effective radiative forcing.

How to cite: Goren, T., Feingold, G., Gryspeerdt, E., Kazil, J., and Quaas, J.: Exploring the Effect of Aerosol on Marine Cloud Cover Using a Counterfactual Approach, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-12323, https://doi.org/10.5194/egusphere-egu22-12323, 2022.

How to cite: Delbeke, L. and Wang, C.: Influence of Different Atmospheric Aerosol Compositions on the Life Cycle of Stratiform Clouds over Southern West Africa, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-11779, https://doi.org/10.5194/egusphere-egu22-11779, 2022.

Using in-situ aircraft observations from six flights over Zhejiang on Sep. 1 and Sep. 4, 2016, this study investigates differences in aerosol and cloud properties between daytime and nighttime. The samples were divided into marine type and continental type based on the backward air mass trajectories and aerosol characteristics. The results show that the aerosol number concentration (N_a) near the ground during daytime is higher than that at nighttime. During daytime, Na has a significant decreasing trend near the top of the planetary boundary layer (PBL), which is not obvious during nighttime. There may be still a relative high concentration of aerosols remaining in the transition zone between the PBL and the free troposphere. Under similar liquid water content (LWC) conditions, the cloud droplet number concentration (N_c) at night is lower, and the cloud droplet effective diameter (cloud ED) is larger. The total N_a of marine type aerosols is generally lower than that of continental type aerosols, but for aerosols with particle diameters greater than 1 μm, the marine type aerosols are higher. The study shows a strong negative N_a-cloud ED relationship for marine type aerosols, but no obvious Na-cloud ED relationship for continental type aerosols. The number of cloud condensation nuclei (CCN) is higher under high-N_a conditions; the ratio of CCN to N_a reveals that the activation efficiency of marine type aerosols is higher than that of continental type aerosols. There is no obvious difference in activation efficiency between day and night.

How to cite: Che, Y., Zhang, J., Zhao, C., Zhou, X., and Duan, J.: Aerosol and cloud properties over a coastal area from aircraft observations in Zhejiang, China, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-17, https://doi.org/10.5194/egusphere-egu22-17, 2022.

High-resolution modeling of aerosol-cloud interactions typically applies aerosol perturbations for the duration of the simulation, which may be anywhere from a few hours to a few days.

In reality, however, natural and anthropogenic aerosol perturbations have characteristic durations, along with concomitant changes in meteorology and associated cloud conditions. In this talk we will explore the effect of the duration of aerosol perturbations on the cloud radiative responses using idealized large eddy simulations. We will also consider observed seasonal cycles in meteorology, clouds, and aerosol, and how they affect cloud albedo responses. These exercises will help to assess the radiative effect of aerosol-cloud interactions.

How to cite: Feingold, G., Prabhakaran, P., Zhang, J., Zhou, X., and Hoffmann, F.: Exploring the Influence of the Duration of Aerosol Perturbations on Cloud Responses, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-13537, https://doi.org/10.5194/egusphere-egu22-13537, 2022.

The so-called autoconversion is a key numerical process used to describe the coalescence growth of cloud droplets to drizzle and rain in atmospheric models. Together with further growth of drizzle through accretion, these processes are typically represented by relatively simple two-moment bulk parameterizations. One of the shortcomings of this approach is that most often the parameterizations do not explicitly consider the impact of coarse mode aerosol and giant cloud condensation nuclei (GCCN), even though they are known to be important in marine cloud regimes. More elaborate sectional models, such as the Sectional Aerosol Module for Large-Scale Applications (SALSA) solve the coalescence growth equations and are able to account for the effect of large aerosol particles on droplet growth and drizzle formation.

In this work, the autoconversion and accretion rates diagnosed from SALSA are compared with the process rates from bulk parameterizations run simultaneously within a large-eddy simulation model (UCLALES-SALSA). The model is used to create an ensemble of simulations comprising varying aerosol conditions in terms of the coarse mode particles in marine stratocumulus and shallow cumulus regimes. The difference between SALSA and bulk process rates is taken as the approximation error in the bulk parameterizations. The dependence of this error term on the coarse mode aerosol concentration is shown by a multivariate sensitivity analysis based on the ensemble data. Further, machine learning methods, notably the neural networks, are used to represent the approximation error term. The trained networks are shown to successfully capture the main features of the model-based approximation error. As an outlook, the machine learning-based representation of the approximation error allows to enhance the existing bulk microphysical parameterizations for drizzle formation and to introduce an explicit dependence on coarse model aerosol and GCCN.

How to cite: Tonttila, J., Romakkaniemi, S., and Kokkola, H.: Approximation error correction for drizzle formation in bulk microphysical parameterizations, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-7676, https://doi.org/10.5194/egusphere-egu22-7676, 2022.

The transition from stratocumulus to cumulus clouds that takes place as air is advected from the subtropics towards the equator causes a decrease in cloud radiative effect, with cloud fraction halving from start to finish. The transition is initiated by increasing sea surface temperatures, and it is widely agreed that the lower tropospheric stability plays a key role in the timing of the transition. In this work, we study the relative importance of five atmospheric initial conditions: specific humidity in the boundary layer and free troposphere, free tropospheric potential temperature, inversion height and initial aerosol distribution. We simulate a Lagrangian trajectory of a stratocumulus-to-cumulus transition, using the Met Office/NERC cloud model coupled with a bulk microphysics scheme and a radiation scheme. From this base simulation we make 60 perturbations to simulate the transition under different combinations of the atmospheric initial conditions mentioned. Additionally, we include a model parameter from the Khairoutdinov and Kogan autoconversion parameterisation from 2000. We discuss here the relative importance of these so-called parameters, in particular the role of aerosol, and we explore whether a much faster transition by drizzle takes place in simulations with lower aerosol concentrations. 

How to cite: Sansom, R., Lee, L., Johnson, J., Regayre, L., and Carslaw, K.: Exploring a Stratocumulus-to-Cumulus Transition: A Perturbed Parameter Ensemble of Large-Eddy Simulations, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-12007, https://doi.org/10.5194/egusphere-egu22-12007, 2022.

In this contribution geographic patterns of fog and low stratus (FLS) formation and dissipation over central Europe are presented using a novel satellite-based data set. 
Formation and dissipation of FLS are the results of complex interactions of meteorological and land-surface processes. Furthermore, the timing of FLS formation and dissipation has implications for traffic and the production of solar energy. Yet, little is known about the spatial and temporal patterns of both in central Europe. To improve this situation, this study analyzes these patterns, as well as the meteorological drivers and their modifications by the land surface. The basis of the analysis is a novel FLS formation and dissipation data set, derived based on satellite FLS products and logistic regression.
Very distinct and contrasting spatial and seasonal patterns of FLS formation and dissipation are found across the study area. In large-scale river valleys, FLS forms most frequently in the morning and dissipates in the afternoon. In mountainous areas and on the coast, FLS forms during the night and dissipates in the morning. FLS persists longer in winter compared to other seasons. The quantitative analysis of meteorological drivers shows that the large-scale meteorological conditions, in particular mean surface pressure and wind speed, substantially influence the timing of FLS formation and dissipation. Local variations in topography modulate these conditions, leading to local differences in the observed patterns.

How to cite: Pauli, E., Cermak, J., and Andersen, H.: A geographic perspective on fog and low stratus formation and dissipation over central Europe, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-9720, https://doi.org/10.5194/egusphere-egu22-9720, 2022.

Ship exhausts have historically been significant sources of sulfur dioxide and aerosols to the marine atmosphere and some global models suggest the emissions cause a large negative radiative forcing by modifying cloud properties. International Maritime Organisation (IMO, an agency of the UN) regulations require that ships in international waters reduce their sulfur emissions from a maximum of 3.5% to 0.5% from January 2020. The ACRUISE project, taking advantage of this unique large-scale aerosol perturbation, investigates the impacts of the IMO’s 2020 sulfur regulations on aerosols, clouds, and radiation in the North Atlantic and globally. Here I summarise our findings so far from intensive aircraft observations, high-resolution model simulations, and deep learning-based satellite cloud analysis. 

Aerosol-cloud interaction near shipping lanes was studied from an aircraft in the northeast Atlantic in 2019 as well as in 2021. Aerosol chemical and physical properties were markedly different between the two years, with much lower sulfur content, smaller, and less hygroscopic aerosols in 2021. A detailed analysis of the aerosol and cloud microphysics observations within/immediately outside the ship plumes will be performed to determine whether some clouds appeared to be strongly impacted by ship plumes, while other clouds were not. To help interpret the aircraft data and provide context, we ran nested regional domain simulations of the Met Office Unified Model for all flight campaigns. These high-resolution simulations (few hundred metres) show a generally diffuse pattern of perturbed trace gases and aerosols that are not apparent as individual ship tracks, suggesting that analysis of tracks alone may underestimate the climatic effects of ship emissions.

We have trained a deep learning model to detect ship-tracks in satellite imagery with good skill and applied it to the whole MODIS mission in order to develop a global climatology. We will discuss the spatial and temporal distribution of shiptracks relative to the underlying ship emissions, and particularly focus on the effects of the IMO regulation as well as the global COVID-19 pandemic. Ongoing work that combines airmass trajectory modelling with known positions of ships will enable us to assess the impact of ship emissions on all pixels, and not just those identified as ship tracks.

How to cite: Yang, M., Bell, T., Bower, K., Carslaw, K., Choularton, T., Christensen, M., Coe, H., Grosvenor, D., Lee, J., Watson-Parris, D., Stier, P., and Yoshioka, M.: Insights from ACRUISE (Atmospheric Composition and Radiative forcing changes due to UN International Ship Emissions regulations) from aircraft, modelling, and satellite perspectives, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-7078, https://doi.org/10.5194/egusphere-egu22-7078, 2022.

The sensitivity of clouds to anthropogenic aerosol perturbations remains one of the largest uncertainties in the human forcing of the climate system. A key difficulty is in isolating the impact of aerosols from large-scale covariability of aerosol and cloud properties. Natural experiments, where aerosol is produced independently of the cloud and meteorological properties, provide a pathway to address this issue. These aerosol sources often modify cloud properties, leaving linear cloud features known as shiptracks (when formed by a ship) or pollution tracks (more generally).

In this work, we use a database of point sources of aerosol over both land and ocean to identify clouds that are sensitive to aerosol and to measure their response. Using a neural network to identify when a point source is modifying the cloud, we are able to measure the sensitivity of individual clouds to aerosol at a global scale, looking at over 400 million cases.

We find the probability of track formation is strongly dependent on the background cloud and meteorological state, similar to previous regional studies. With our global database, we identify regions that are strongly susceptible to aerosol perturbations, even where aerosol sources are rare. We find that there are several regions that are highly susceptible to aerosol, but that have been previously overlooked due to a low frequency of pollution tracks.    

How to cite: Gryspeerdt, E., Louro Coelho, M., Smith, T., Suarez De La Fuente, S., Quilelli Correa Rocha Ribeiro, R., and van Reeuwijk, M.: Measuring cloud sensitivity to aerosols at a global scale using isolated aerosol sources, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-9756, https://doi.org/10.5194/egusphere-egu22-9756, 2022.

The understanding of aerosols and their impact on climate via their interactions with clouds remains one of the largest uncertainty to our current estimates of future climate warming. A change in the amount of aerosol in the atmosphere, whether natural or artificial, can have a direct impact on cloud particle nucleation. However, anthropogenically induced aerosol-cloud interactions (aci) are thought to counteract some of the effects of global warming, quantifying their effect on total radiative forcing is vital for future climate predictions.

To tackle this problem, recent research has focused on so-called natural laboratories in order to better understand aci, meaning experiments where aerosol emissions are relatively well localized and understood, hence removing one important aspect of the aci uncertainties. For instance, volcano eruptions, ship tracks, industrial tracks, or contrails represent such laboratories. The purpose of the study is to assess if emission restriction events can act as a natural laboratory for aci studies, almost in a reverse manner as industrial tracks. Here, we will compare regional cloud properties observed during the lockdown periods to climatologies using MODIS satellite cloud droplet number concentration (N_d) retrievals. CAMS global reanalysis and CAM-chem model simulations are used to study CCN activity and aerosol emissions during the lockdown period. This study will focus on different industrial regions to examine if there are any clear signals of aci. It is found that there is no significant decrease in N_d during lockdown compared to climatology at a regional scale. Although there is a reduction in various anthropogenic activities such as industrial emissions, motor vehicle emissions, etc but the background CCN conditions played an important role in the influence of N_d.

How to cite: Saiprakash, A. and Sourdeval, O.: Satellite-based investigation of the impact of COVID-19 restrictions on cloud properties in industrial regions, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-8581, https://doi.org/10.5194/egusphere-egu22-8581, 2022.

Volcanic aerosol heating perturbs the tropical tropopause layer (TTL) and with it the stratospheric water budget. Whereas the effect of increased cold point temperatures on water slowly ascending into the TTL can be studied using general circulation models (GCM), the robustness of changes in convection after volcanic eruptions in these models is unclear as the TTL is tuned to unperturbed conditions and the simulations highly rely on parametrizations. Estimating the changes in the contributions of temperature effects, overshoots and vertical diffusion after a volcanic eruption or in a geoengineering study accordingly remains a challenge in GCM simulations. The emerging cloud resolving simulations however offer the unique possibility to gain insight into the sensitivity of the TTL to external forcings.

They allow in particular to study potential changes of convection, where two processes counteract each other after volcanic eruptions: the downwards shift of the lapse rate tropopause favoring overshooting convection in combination with increased stability in the TTL region suppressing overshooting convection.

We analyze these effects employing convection-resolving simulations for the atmosphere with the Icosahedral Nonhydrostatic Weather and Climate Model (ICON-A) at 10 km horizontal resolution in two scenarios: a control run and a volcanically perturbed run. The perturbed run has an aerosol layer in the lower stratosphere corresponding to the peak loading of an injection of 20 Tg sulfur using sea surface temperatures from the control run. In addition to a downwards shift of the lapse rate tropopause, we find that the level of neutral buoyancy, based on the temperature difference in convective areas and their surroundings, reaches the TTL more often in the volcanically perturbed simulations. This allows – contrary to previous assumptions - for more overshoots into the region above the tropical lapse rate tropopause.

How to cite: Kroll, C., Kornblueh, L., Dauhut, T., Schmidt, H., Timmreck, C., and Fueglistaler, S.: The volcanic impact on convection and stratospheric ice, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-10341, https://doi.org/10.5194/egusphere-egu22-10341, 2022.

Clouds play a critical role in the climate system by modifying incoming and outgoing radiation. In turn, clouds are affected by climatic changes. Cloud fraction (CF, the part of the sky covered by clouds) is one of the most reliable observed cloud properties and can be regarded as a good first approximation for cloud radiative effects.

Different types of clouds modify the energy budget and respond to climatic changes differently. Current climate models predict a negative trend in CF of shallow clouds (with low top height), which will further warm the climate by reducing the reflection of solar radiation. For other types of clouds, the predicted trend in CF has a net-zero radiative effect. The modeled results suffer from large variability and it contributes greatly to the uncertainty of climate projections. Therefore, observational constraints are badly needed.

Here, we divide the cloud records into three classes: total clouds (including all clouds), shallow clouds (with top pressure > 700 hPa), and free-atmosphere clouds (other than shallow). For each class, we decompose the general CF into two parts: one which is related to clouds’ horizontal size (CF under cloudy conditions) and a second part, related to the frequency of occurrence (the ratio of cloudy days to all days). 17 years (2003-2019) of satellite observations (MODIS aboard Aqua) and reanalysis data (ERA5) are used in this work. Satellite records show significant regional CF trends. They show that: (1) the trend in shallow clouds’ size dominates the trend in total CF, (2) there are opposite trends between the shallow and free-atmosphere clouds’ occurrence, and (3) the trend in shallow clouds’ occurrence compensate the trend in shallow clouds’ size and lead to a weak trend in shallow CF. It can indicate the development of shallow clouds into free-atmosphere clouds and it can relate to an overlapping problem, where MODIS cannot detect shallow clouds under other clouds. Reanalysis data reveals that after considering a correction for the overlapping problem of cloudy layers, the observed opposite trends (point no. 2 above) are still detected, but to a smaller extent, indicating that this relationship in MODIS records is impacted by the clouds overlapping problem. This means that when considering a correction for the overlapping problem, larger local trends in shallow clouds’ CF (point no. 3 above) are expected. Our findings provide new statistical relationships between clouds’ trends by high-quality observational records and shows that the overlapping problem biases systemically the trends in cloud properties.

How to cite: Liu, H., Koren, I., and Altaratz, O.: Observed Relationships between Cloud Fraction Trends of Shallow and Free-atmosphere Clouds, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-5334, https://doi.org/10.5194/egusphere-egu22-5334, 2022.

Measuring the level of cloud coverage (CC) is important when analysing the absorption of solar radiation, the cloud formation process, and also to improve the forecast in general. This is becoming a more common task with the availability of digital sensors and their usage in monitoring the sky.

In this study we present a method of estimating the level of cloud coverage using Deep Learning (DL) applied to all-sky images of Meteorites Orbits Reconstruction by Optical Imaging (MOROI) network.

To label the data, a supervised validation was employed on the daytime images captured during the course of two years, recorded on Galati, Romania These were divided into three groups; CC <20%, CC between 20-80%, and CC >80%. Next, a set of DL models were trained, optimized and tested towards accurately classifying images according to each group.

We found that the classification accuracy can range between 89-95 % depending on how the cloud coverage is labeled and how the daytime is defined. This is mostly due to thin cirrus clouds, tricking the models, or poor sky illumination during sunrise or sunset. We discuss these methods and present a few strategies which circumvent classification problems, to further increase the accuracy of models.

The next step is to extend the study on the rest of the network, and also combining it with other sensors (e.g. satellite data) in order to understand the cloud-circulation coupling and its impact across climate models.

How to cite: Anghel, S., Voiculescu, M., Nedelcu, D.-A., and Birlan, M.: Cloud coverage estimation via deep learning applied to all-sky data in Romania : First Results, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-415, https://doi.org/10.5194/egusphere-egu22-415, 2022.

In this study we investigate the level of accuracy of scaling methods, and analytical approximations, commonly used in fast radiative transfer routines of weather and climate models. Specifically, we focus on Chou’s approximation (Chou et al., 1999), and a simple scaling method based on the similarity principle. The former one is widely implemented in existing fast radiative codes to solve the radiative transfer problem in the infrared spectral region. At this regard, updated Chou backscatter parameters are computed based on realistic particle size distributions of liquid water and ice particles and by exploiting state of the art optical properties databases.

The assessment of the accuracy of approximate methodology all over the infrared spectrum is obtained by considering a widespread collection of atmospheric scenarios. Top of the atmosphere synthetic spectral radiances are computed for each scenario by considering alternatively an accurate and time-consuming methodology, such as the discrete ordinate solution (DISORT), or the approximate methodologies. The residuals are evaluated at far- and mid-infrared wavelengths and compared with the goal noise of the future 9th Earth Explorer FORUM satellite sensor (Palchetti et al., 2020). Results are discussed and analyzed in terms of geometrical, microphysical, and optical properties of the clouds layers (Martinazzo et al., 2021). In case of both water and ice cloud scenarios, the approximate solutions perform well in the mid infrared for most of the cases studied. When the far infrared region is considered, not negligible inaccuracies are observed.

To reduce the computational errors of basic scaling methods, a correction term is modelled and computed using the solution proposed by Tang et al. (2018) which assumes a downward radiance term not necessarily equal to the blackbody radiance, as it is done in Chou’s approximation. The Tang methodology, originally implemented for flux computations, is here adapted to the simulation of radiance fields, and refined by computing the appropriate multiplicative coefficients in the far and mid-infrared separately.  Results show that the range of validity of the new methodology is extended with respect to Chou's approximation and covers most of the cloud cases encountered in nature. It represents an accurate solution for the computation of radiance fields in presence of cirrus clouds which are one of the targets of the FORUM mission.

Finally, the whole set of radiative parameters needed to solve the radiative transfer equation using the Chou and Tang approximations is parametrized by mean of polynomial functions of the effective dimension of the cloud particle size distribution. The parametrized parameters are then implemented in the sigma-FORUM code a forward model designed for the fast calculation of radiance and its derivatives with respect to atmospheric and spectroscopic parameters of nadir-looking hyperspectral instruments. The σ-FORUM model is an updated version of sigma-IASI model (Amato et al., 2002), a monochromatic radiative transfer model based on a look-up table of optical depths parametrized as a polynomial concerning the atmospheric temperature and constituents. The strategy enables fast, accurate radiance and analytical derivatives calculations.

How to cite: Martinazzo, M., Maestri, T., Cossich, W., Serio, C., Masiello, G., and Venafra, S.: A New Cloud Properties Parameterization Implemented in the Fast Radiative Transfer Code sigma-FORUM, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-10593, https://doi.org/10.5194/egusphere-egu22-10593, 2022.

The response of cloud properties to aerosol perturbations is one of the greatest uncertainties associated with anthropogenic climate forcing. Clouds play a central role in the Arctic surface energy budget, so understanding how aerosols influence their radiative properties is important for projecting future changes in the region. This is especially relevant as increases in temperature and reductions in sea ice extent allow for more industrial activity in the region, thereby introducing fresh sources of aerosol.  

As it is difficult to retrieve reliable satellite observations in polar regions due to issues such as high solar zenith angles, the impact of aerosols on Arctic clouds is particularly uncertain. However, by carefully filtering the satellite data to remove cases associated with retrieval biases, this work uses multiple satellite and reanalysis datasets to develop new constraints for the influence of aerosols on cloud properties.

The factors which influence the droplet number concentration-liquid water path (N_d-LWP) relationship, a key component of the aerosol-liquid water path relationship, are investigated. The N_d-LWP relationship varies significantly geographically, with increases in LWP with N_d observed at high latitudes. A range of meteorological factors are investigated, and it is shown that the lower tropospheric stability (LTS) is the driving force behind the variability in the N_d-LWP relationship in the Arctic. At high stability, the relationship is significantly more positive, producing LWP increases in more polluted environments, in contrast to the response of clouds to aerosol perturbations seen at lower latitudes.

As the Arctic warms, the boundary layer stability is projected to decrease. Additionally, industrial activity is expected to increase in the region, which may increase the aerosol burden. When considered individually, these two effects would lead to increases in LWP in marine Arctic clouds. However, when working together they may produce clouds with lower water paths, leading to a weaker negative cloud feedback in a more polluted environment. 

How to cite: Murray-Watson, R. and Gryspeerdt, E.: Stability dependent increases in liquid water with droplet number in the Arctic, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-2956, https://doi.org/10.5194/egusphere-egu22-2956, 2022.

Recent observations in the field measurements and laboratory studies suggest that organic aerosol particles may exist as highly viscous semi-solids or amorphous glassy solids under certain conditions, with important implications for atmospheric chemistry, climate, and air quality. A number of complementary techniques have been developed to probe the viscosity of aerosol particles in the last ten years. However, none of the available techniques is sufficiently versatile to determine aerosol viscosity at atmospherically relevant conditions for a range of particle sizes, chemical compositions, and sample sizes. Here we present a novel way to measure the viscosity of levitated droplets suspended in an electrodynamic balance under atmospheric conditions. Capillary oscillations are induced in a levitated droplet by the application of an external AC field. These oscillations are monitored using a high-speed camera or light scattering, and the viscosity is determined by fitting these oscillations using a suitable theoretical model. The model is based on the asymptotic analysis of surface oscillations of a charged drop, carried out using the viscous-potential flow theory.

How to cite: Singh, M., Jones, S., Duft, D., Kiselev, A., and Leisner, T.: Measurement of viscosity at low temperature from resonating droplet levitated in an electrodynamic balance, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-10270, https://doi.org/10.5194/egusphere-egu22-10270, 2022.

The susceptibility of cloud droplet number concentration (CDNC) to cloud condensation nuclei (CCN) number concentration is one of the major factors controlling the aerosol indirect forcing. In this study we investigate the sensitivity of CDNC to CCN concentrations using long term in-situ observations from three stations (Puijo, Pallas, Zeppelin) locating in Finland and Arctic. These stations represent semiurban, remote and Arctic remote environments with differences in typical updraft velocity conditions as well as in aerosol number concentrations. We compare the in-situ observations with three large scale models (ECHAM-M7, ECHAM-SALSA and NorESM) having differences in aerosol presentation while the activation parametrization is the same in all three model setups. In the comparison we use CDNC and CCN model outputs of the gridbox corresponding to the location and the height for each station. In addition, we compare the updraft velocities from the models and stations when they are available. Our current observational results show very high susceptibility of CDNC and CCN in all investigated stations. The agreement between the large scale models and observations was very good for Puijo and Pallas stations, but for the Arctic station (Zeppelin) the modelled CDNC susceptibility to CCN was much lower than the observed. This might be related to the recent results demonstrating that Aitken mode particles can active to cloud droplets at Zeppelin station (Bulatovic et al., 2021; Karlsson et al., 2021). In addition, at Zeppelin CDNC exhibits very low values which are below the lower bound imposed by ECHAM.

References:

Bulatovic, I., Igel, A. L., Leck, C., Heintzenberg, J., Riipinen, I., and Ekman, A. M. L.: The importance of Aitken mode aerosol particles for cloud sustenance in the summertime high Arctic – a simulation study supported by observational data, Atmos. Chem. Phys., 21, 3871–3897, https://doi.org/10.5194/acp-21-3871-2021, 2021.

Karlsson, L., Krejci, R., Koike, M., Ebell, K., and Zieger, P.: A long-term study of cloud residuals from low-level Arctic clouds, Atmos. Chem. Phys., 21, 8933–8959, 2021.

How to cite: Virtanen, A., Joutsensaari, J., Kokkola, H., Seland, Ø., Zieger, P., Karlsson, L., Riipinen, I., Krejci, R., Hyvärinen, A., Lihavainen, H., and Romakkaniemi, S.: Cloud droplet number susceptibility to CCN concentrations in low level boundary layer clouds: comparison of in-situ observations and large-scale models, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-4368, https://doi.org/10.5194/egusphere-egu22-4368, 2022.

As large-eddy simulations (LES), which explicitly simulate aerosol-cloud interactions, are often considered as benchmark simulations in climate science, it is necessary to critically evaluate if these high-resolution models can skillfully represent expected physical phenomena.

Here, we focus on the first aerosol indirect aerosol effect in a warm stratocumulus cloud. We investigate if the MIMICA LES (Savre et al., 2014) with a widely used bulk two-moment microphysical scheme (Seifert and Beheng, 2006) can reproduce the susceptibility regimes identified by Reutter et al., (2009). Using a parcel model, Reutter et al. (2009) showed that the cloud droplet number (N_d) responds differently to an increase in aerosol number (N_a) depending on ambient updraft strength (w). In the aerosol-limited regime, enough supersaturation can be generated by the updraft motions in the atmosphere so that increasing N_a leads to an increase in N_d. Conversely, in the updraft-limited regime, adding aerosol will not increase as activation is limited by the updraft strength and only increasing w will lead to an increase in N_a.

In the standard setup, the LES cannot simulate the transition from the aerosol- to the updraft-limited regime. Only when implementing a renormalization procedure following Reisin et al., (1996) and, at the same time, increasing the initial droplet radius of newly activated droplets (r_dⁱ) to values large than r_dⁱ>1µm, a regime transition emerges. However, a clear recommendation for the choice of r_dⁱ cannot be made upon physical arguments at this point. Interestingly, the “arbitrarily chosen” droplet mass by Seifert and Beheng (2006) of 1*10^-12kg, which corresponds to r_dⁱ≈6µm, seems to agree quite well with the expectations from parcel model simulations. The choice is, however, still arbitrary and therefore physically questionable.

A potential way to avoid this problem, which mainly occurs at high aerosol concentrations, would be to run the LES with a small enough temporal resolution (Δt≈0.1s) to explicitly resolve all relevant microphysical processes.

References

Reisin, T., Levin, Z., Tzivion, S., 1996. Rain Production in Convective Clouds As Simulated in an Axisymmetric Model with Detailed Microphysics. Part I: Description of the Model. Journal of Atmospheric Sciences 53, 497–520.

Reutter, P., Su, H., Trentmann, J., Simmel, M., Rose, D., Gunthe, S.S., Wernli, H., Andreae, M.O., Pöschl, U., 2009. Aerosol- and updraft-limited regimes of cloud droplet formation: influence of particle number, size and hygroscopicity on the activation of cloud condensation nuclei (CCN). Atmospheric Chemistry and Physics 9, 7067–7080.

Savre, J., Ekman, A.M.L., Svensson, G., 2014. Technical note: Introduction to MIMICA, a large-eddy simulation solver for cloudy planetary boundary layers. Journal of Advances in Modeling Earth Systems 6, 630–649.

Seifert, A., Beheng, K.D., 2006. A two-moment cloud microphysics parameterization for mixed-phase clouds. Part 1: Model description. Meteorol. Atmos. Phys. 92, 45–66.

How to cite: Schwarz, M., Savre, J., Sudhakar, D., Quaas, J., and Ekman, A. M. L.: Modeling the transition from the aerosol- to the updraft-limited cloud droplet susceptibility regime in large-eddy simulations with bulk microphysics, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-4978, https://doi.org/10.5194/egusphere-egu22-4978, 2022.

Scaled-up DNS and implicit LES simulations are used to study turbulent cloud base CCN activation and early growth of cloud droplets. The simulation framework includes a triply periodic computational domain ~1,000 cubic meters filled with inertial-range homogeneous isotropic turbulence. The domain experiences decrease of the mean air temperature and reduction of the mean pressure, both mimicking the rise of an adiabatic air parcel through the cloud base. Results of turbulent simulations are compared to CCN activation and droplet growth within a classical nonturbulent rising parcel. The key difference is a blurriness of the separation between activated and nonactivated (haze) CCN, especially for weak mean ascent rates, when CCN activate and subsequently some deactivate instead of becoming cloud droplets above the cloud base. This leads to significantly larger spectral widths in turbulent parcel simulations compared to the adiabatic nonturbulent parcel once CCN activation is completed. Further increase of the spectral width in the turbulent parcel is similar to that for the initially-monodisperse droplets in the inertial-range homogeneous isotropic turbulence that we and others studied previously, with the standard deviation of the radius squared increasing approximately as the square root of time. This contrasts with the classical nonturbulent parcel framework for which the radius squared standard deviation above the cloud base remains constant because of the parabolic growth of cloud droplets once surface tension and dilute effects can be neglected.

How to cite: Grabowski, W. W., Thomas, L., and Kumar, B.: Impact of turbulence on CCN activation and early growth of cloud droplets, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-13190, https://doi.org/10.5194/egusphere-egu22-13190, 2022.

On Earth, atmospheric instabilities drive the aerosolization of distinct microorganisms directly from the biosphere. Particle diameters of these (bio)aerosols typically range from ∼1nm to ∼100μm, and their residence times in the atmosphere have been suggested to range from a few days to weeks or even months. Found even beyond the troposphere, they have been shown to, in some cases, be metabolically active as well as involved in cloud processing mechanisms, depending on their surface properties, by acting as cloud condensation nuclei (CCN) or ice nuclei (IN), and thus effectively changing cloud life time and optical properties via a change in the droplet size distribution within the cloud.

How to cite: Garrido Zornoza, M., Mitarai, N., and Haerter, J. O.: Observational implications on the presence of selected bioaerosols in exoplanetary atmospheres, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-13218, https://doi.org/10.5194/egusphere-egu22-13218, 2022.