Articles | Volume 17, issue 9
https://doi.org/10.5194/gmd-17-4017-2024
© Author(s) 2024. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Special issue:
https://doi.org/10.5194/gmd-17-4017-2024
© Author(s) 2024. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Development and technical paper
| 
16 May 2024
Development and technical paper |
| 16 May 2024
| 16 May 2024
Efficient and stable coupling of the SuperdropNet deep-learning-based cloud microphysics (v0.1.0) with the ICON climate and weather model (v2.6.5)
German Climate Computing Center (DKRZ), Hamburg, Germany
Helmholtz-Zentrum Hereon, Geesthacht, Germany
Helmholtz AI, Munich, Germany
Helmholtz-Zentrum Hereon, Geesthacht, Germany
Helmholtz AI, Munich, Germany
International Max Planck Research School on Earth System Modelling, Hamburg, Germany
Tobias Weigel
German Climate Computing Center (DKRZ), Hamburg, Germany
Helmholtz-Zentrum Hereon, Geesthacht, Germany
Helmholtz AI, Munich, Germany
David S. Greenberg
Helmholtz-Zentrum Hereon, Geesthacht, Germany
Helmholtz AI, Munich, Germany
Related authors
No articles found.
Venkatramani Balaji, Karl E. Taylor, Martin Juckes, Bryan N. Lawrence, Paul J. Durack, Michael Lautenschlager, Chris Blanton, Luca Cinquini, Sébastien Denvil, Mark Elkington, Francesca Guglielmo, Eric Guilyardi, David Hassell, Slava Kharin, Stefan Kindermann, Sergey Nikonov, Aparna Radhakrishnan, Martina Stockhause, Tobias Weigel, and Dean Williams
Geosci. Model Dev., 11, 3659–3680, https://doi.org/10.5194/gmd-11-3659-2018, https://doi.org/10.5194/gmd-11-3659-2018, 2018
Short summary
Short summary
We present recommendations for the global data infrastructure needed to support CMIP scientific design and its future growth and evolution. We follow a dataset-centric design less prone to systemic failure. Scientific publication in the digital age is evolving to make data a primary scientific output, alongside articles. We design toward that future scientific data ecosystem, informed by the need for reproducibility, data provenance, future data technologies, and measures of costs and benefits.
Related subject area
Atmospheric sciences
Three-dimensional variational assimilation with a multivariate background error covariance for the Model for Prediction Across Scales – Atmosphere with the Joint Effort for Data assimilation Integration (JEDI-MPAS 2.0.0-beta)
FUME 2.0 – Flexible Universal processor for Modeling Emissions
DEUCE v1.0: a neural network for probabilistic precipitation nowcasting with aleatoric and epistemic uncertainties
Evaluation of multi-season convection-permitting atmosphere – mixed-layer ocean simulations of the Maritime Continent
Investigating the impact of coupling HARMONIE-WINS50 (cy43) meteorology to LOTOS-EUROS (v2.2.002) on a simulation of NO2 concentrations over the Netherlands
Balloon drift estimation and improved position estimates for radiosondes
Emission ensemble approach to improve the development of multi-scale emission inventories
What is the relative impact of nudging and online coupling on meteorological variables, pollutant concentrations and aerosol optical properties?
Diagnosing drivers of PM2.5 simulation biases in China from meteorology, chemical composition, and emission sources using an efficient machine learning method
Validation and analysis of the Polair3D v1.11 chemical transport model over Quebec
Assimilation of GNSS tropospheric gradients into the Weather Research and Forecasting (WRF) model version 4.4.1
Identifying atmospheric rivers and their poleward latent heat transport with generalizable neural networks: ARCNNv1
Assessing acetone for the GISS ModelE2.1 Earth system model
Bergen metrics: composite error metrics for assessing performance of climate models using EURO-CORDEX simulations
A dynamic approach to three-dimensional radiative transfer in subkilometer-scale numerical weather prediction models: the dynamic TenStream solver v1.0
Evaluation and development of surface layer scheme representation of temperature inversions over boreal forests in Arctic wintertime conditions
Modelling wind farm effects in HARMONIE–AROME (cycle 43.2.2) – Part 1: Implementation and evaluation
Analytical and adaptable initial conditions for dry and moist baroclinic waves in the global hydrostatic model OpenIFS (CY43R3)
Challenges of constructing and selecting the “perfect” boundary conditions for the large-eddy simulation model PALM
A machine learning approach for evaluating Southern Ocean cloud radiative biases in a global atmosphere model
Decision Support System version 1.0 (DSS v1.0) for air quality management in Delhi, India
How non-equilibrium aerosol chemistry impacts particle acidity: the GMXe AERosol CHEMistry (GMXe–AERCHEM, v1.0) sub-submodel of MESSy
A grid model for vertical correction of precipitable water vapor over the Chinese mainland and surrounding areas using random forest
MEXPLORER 1.0.0 – a mechanism explorer for analysis and visualization of chemical reaction pathways based on graph theory
Advances and prospects of deep learning for medium-range extreme weather forecasting
An overview of the Western United States Dynamically Downscaled Dataset (WUS-D3)
cloudbandPy 1.0: an automated algorithm for the detection of tropical–extratropical cloud bands
PyRTlib: an educational Python-based library for non-scattering atmospheric microwave radiative transfer computations
Deep learning applied to CO2 power plant emissions quantification using simulated satellite images
Sensitivity of the WRF-Chem v4.4 simulations of ozone and formaldehyde and their precursors to multiple bottom-up emission inventories over East Asia during the KORUS-AQ 2016 field campaign
Optimising urban measurement networks for CO2 flux estimation: a high-resolution observing system simulation experiment using GRAMM/GRAL
Assessment of climate biases in OpenIFS version 43r3 across model horizontal resolutions and time steps
High-resolution multi-scaling of outdoor human thermal comfort and its intra-urban variability based on machine learning
Effects of vertical grid spacing on the climate simulated in the ICON-Sapphire global storm-resolving model
Development of the tangent linear and adjoint models of the global online chemical transport model MPAS-CO2 v7.3
Impacts of updated reaction kinetics on the global GEOS-Chem simulation of atmospheric chemistry
Spatial spin-up of precipitation in limited-area convection-permitting simulations over North America using the CRCM6/GEM5.0 model
Sensitivity of atmospheric rivers to aerosol treatment in regional climate simulations: insights from the AIRA identification algorithm
The implementation of dust mineralogy in COSMO5.05-MUSCAT
Implementation of the ISORROPIA-lite aerosol thermodynamics model into the EMAC chemistry climate model (based on MESSy v2.55): implications for aerosol composition and acidity
Evaluation of surface shortwave downward radiation forecasts by the numerical weather prediction model AROME
GEO4PALM v1.1: an open-source geospatial data processing toolkit for the PALM model system
Modeling collision–coalescence in particle microphysics: numerical convergence of mean and variance of precipitation in cloud simulations using the University of Warsaw Lagrangian Cloud Model (UWLCM) 2.1
Modeling below-cloud scavenging of size-resolved particles in GEM-MACHv3.1
Impacts of a double-moment bulk cloud microphysics scheme (NDW6-G23) on aerosol fields in NICAM.19 with a global 14 km grid resolution
Sensitivity of air quality model responses to emission changes: comparison of results based on four EU inventories through FAIRMODE benchmarking methodology
A simple and realistic aerosol emission approach for use in the Thompson–Eidhammer microphysics scheme in the NOAA UFS Weather Model (version GSL global-24Feb2022)
Representing effects of surface heterogeneity in a multi-plume eddy diffusivity mass flux boundary layer parameterization
On the formation of biogenic secondary organic aerosol in chemical transport models: an evaluation of the WRF-CHIMERE (v2020r2) model with a focus over the Finnish boreal forest
The first application of a numerically exact, higher-order sensitivity analysis approach for atmospheric modelling: implementation of the hyperdual-step method in the Community Multiscale Air Quality Model (CMAQ) version 5.3.2
Byoung-Joo Jung, Benjamin Ménétrier, Chris Snyder, Zhiquan Liu, Jonathan J. Guerrette, Junmei Ban, Ivette Hernández Baños, Yonggang G. Yu, and William C. Skamarock
Geosci. Model Dev., 17, 3879–3895, https://doi.org/10.5194/gmd-17-3879-2024, https://doi.org/10.5194/gmd-17-3879-2024, 2024
Short summary
Short summary
We describe the multivariate static background error covariance (B) for the JEDI-MPAS 3D-Var data assimilation system. With tuned B parameters, the multivariate B gives physically balanced analysis increment fields in the single-observation test framework. In the month-long cycling experiment with a global 60 km mesh, 3D-Var with static B performs stably. Due to its simple workflow and minimal computational requirements, JEDI-MPAS 3D-Var can be useful for the research community.
Michal Belda, Nina Benešová, Jaroslav Resler, Peter Huszár, Ondřej Vlček, Pavel Krč, Jan Karlický, Pavel Juruš, and Kryštof Eben
Geosci. Model Dev., 17, 3867–3878, https://doi.org/10.5194/gmd-17-3867-2024, https://doi.org/10.5194/gmd-17-3867-2024, 2024
Short summary
Short summary
For modeling atmospheric chemistry, it is necessary to provide data on emissions of pollutants. These can come from various sources and in various forms, and preprocessing of the data to be ingestible by chemistry models can be quite challenging. We developed the FUME processor to use a database layer that internally transforms all input data into a rigid structure, facilitating further processing to allow for emission processing from the continental to the street scale.
Bent Harnist, Seppo Pulkkinen, and Terhi Mäkinen
Geosci. Model Dev., 17, 3839–3866, https://doi.org/10.5194/gmd-17-3839-2024, https://doi.org/10.5194/gmd-17-3839-2024, 2024
Short summary
Short summary
Probabilistic precipitation nowcasting (local forecasting for 0–6 h) is crucial for reducing damage from events like flash floods. For this goal, we propose the DEUCE neural-network-based model which uses data and model uncertainties to generate an ensemble of potential precipitation development scenarios for the next hour. Trained and evaluated with Finnish precipitation composites, DEUCE was found to produce more skillful and reliable nowcasts than established models.
Emma Howard, Steven Woolnough, Nicholas Klingaman, Daniel Shipley, Claudio Sanchez, Simon C. Peatman, Cathryn E. Birch, and Adrian J. Matthews
Geosci. Model Dev., 17, 3815–3837, https://doi.org/10.5194/gmd-17-3815-2024, https://doi.org/10.5194/gmd-17-3815-2024, 2024
Short summary
Short summary
This paper describes a coupled atmosphere–mixed-layer ocean simulation setup that will be used to study weather processes in Southeast Asia. The set-up has been used to compare high-resolution simulations, which are able to partially resolve storms, to coarser simulations, which cannot. We compare the model performance at representing variability of rainfall and sea surface temperatures across length scales between the coarse and fine models.
Andrés Yarce Botero, Michiel van Weele, Arjo Segers, Pier Siebesma, and Henk Eskes
Geosci. Model Dev., 17, 3765–3781, https://doi.org/10.5194/gmd-17-3765-2024, https://doi.org/10.5194/gmd-17-3765-2024, 2024
Short summary
Short summary
HARMONIE WINS50 reanalysis data with 0.025° × 0.025° resolution from 2019 to 2021 were coupled with the LOTOS-EUROS Chemical Transport Model. HARMONIE and ECMWF meteorology configurations against Cabauw observations (52.0° N, 4.9° W) were evaluated as simulated NO2 concentrations with ground-level sensors. Differences in crucial meteorological input parameters (boundary layer height, vertical diffusion coefficient) between the hydrostatic and non-hydrostatic models were analysed.
Ulrich Voggenberger, Leopold Haimberger, Federico Ambrogi, and Paul Poli
Geosci. Model Dev., 17, 3783–3799, https://doi.org/10.5194/gmd-17-3783-2024, https://doi.org/10.5194/gmd-17-3783-2024, 2024
Short summary
Short summary
This paper presents a method for calculating balloon drift from historical radiosonde ascent data. The drift can reach distances of several hundred kilometres and is often neglected. Verification shows the beneficial impact of the more accurate balloon position on model assimilation. The method is not limited to radiosondes but would also work for dropsondes, ozonesondes, or any other in situ sonde carried by the wind in the pre-GNSS era, provided the necessary information is available.
Philippe Thunis, Jeroen Kuenen, Enrico Pisoni, Bertrand Bessagnet, Manjola Banja, Lech Gawuc, Karol Szymankiewicz, Diego Guizardi, Monica Crippa, Susana Lopez-Aparicio, Marc Guevara, Alexander De Meij, Sabine Schindlbacher, and Alain Clappier
Geosci. Model Dev., 17, 3631–3643, https://doi.org/10.5194/gmd-17-3631-2024, https://doi.org/10.5194/gmd-17-3631-2024, 2024
Short summary
Short summary
An ensemble emission inventory is created with the aim of monitoring the status and progress made with the development of EU-wide inventories. This emission ensemble serves as a common benchmark for the screening and allows for the comparison of more than two inventories at a time. Because the emission “truth” is unknown, the approach does not tell which inventory is the closest to reality, but it identifies inconsistencies that require special attention.
Laurent Menut, Bertrand Bessagnet, Arineh Cholakian, Guillaume Siour, Sylvain Mailler, and Romain Pennel
Geosci. Model Dev., 17, 3645–3665, https://doi.org/10.5194/gmd-17-3645-2024, https://doi.org/10.5194/gmd-17-3645-2024, 2024
Short summary
Short summary
This study is about the modelling of the atmospheric composition in Europe during the summer of 2022, when massive wildfires were observed. It is a sensitivity study dedicated to the relative impacts of two modelling processes that are able to modify the meteorology used for the calculation of the atmospheric chemistry and transport of pollutants.
Shuai Wang, Mengyuan Zhang, Yueqi Gao, Peng Wang, Qingyan Fu, and Hongliang Zhang
Geosci. Model Dev., 17, 3617–3629, https://doi.org/10.5194/gmd-17-3617-2024, https://doi.org/10.5194/gmd-17-3617-2024, 2024
Short summary
Short summary
Numerical models are widely used in air pollution modeling but suffer from significant biases. The machine learning model designed in this study shows high efficiency in identifying such biases. Meteorology (relative humidity and cloud cover), chemical composition (secondary organic components and dust aerosols), and emission sources (residential activities) are diagnosed as the main drivers of bias in modeling PM2.5, a typical air pollutant. The results will help to improve numerical models.
Shoma Yamanouchi, Shayamilla Mahagammulla Gamage, Sara Torbatian, Jad Zalzal, Laura Minet, Audrey Smargiassi, Ying Liu, Ling Liu, Forood Azargoshasbi, Jinwoong Kim, Youngseob Kim, Daniel Yazgi, and Marianne Hatzopoulou
Geosci. Model Dev., 17, 3579–3597, https://doi.org/10.5194/gmd-17-3579-2024, https://doi.org/10.5194/gmd-17-3579-2024, 2024
Short summary
Short summary
Air pollution is a major health hazard, and chemical transport models (CTMs) are valuable tools that aid in our understanding of the risks of air pollution at both local and regional scales. In this study, the Polair3D CTM of the Polyphemus air quality modeling platform was set up over Quebec, Canada, to assess the model’s capability in predicting key air pollutant species over the region, at seasonal temporal scales and at regional spatial scales.
Rohith Thundathil, Florian Zus, Galina Dick, and Jens Wickert
Geosci. Model Dev., 17, 3599–3616, https://doi.org/10.5194/gmd-17-3599-2024, https://doi.org/10.5194/gmd-17-3599-2024, 2024
Short summary
Short summary
Global Navigation Satellite Systems (GNSS) provides moisture observations through its densely distributed ground station network. In this research, we assimilate a new type of observation called tropospheric gradient observations, which has never been incorporated into a weather model. We develop a forward operator for gradient-based observations and conduct an assimilation impact study. The study shows significant improvements in the model's humidity fields.
Ankur Mahesh, Travis A. O'Brien, Burlen Loring, Abdelrahman Elbashandy, William Boos, and William D. Collins
Geosci. Model Dev., 17, 3533–3557, https://doi.org/10.5194/gmd-17-3533-2024, https://doi.org/10.5194/gmd-17-3533-2024, 2024
Short summary
Short summary
Atmospheric rivers (ARs) are extreme weather events that can alleviate drought or cause billions of US dollars in flood damage. We train convolutional neural networks (CNNs) to detect ARs with an estimate of the uncertainty. We present a framework to generalize these CNNs to a variety of datasets of past, present, and future climate. Using a simplified simulation of the Earth's atmosphere, we validate the CNNs. We explore the role of ARs in maintaining energy balance in the Earth system.
Alexandra Rivera, Kostas Tsigaridis, Gregory Faluvegi, and Drew Shindell
Geosci. Model Dev., 17, 3487–3505, https://doi.org/10.5194/gmd-17-3487-2024, https://doi.org/10.5194/gmd-17-3487-2024, 2024
Short summary
Short summary
This paper describes and evaluates an improvement to the representation of acetone in the GISS ModelE2.1 Earth system model. We simulate acetone's concentration and transport across the atmosphere as well as its dependence on chemistry, the ocean, and various global emissions. Comparisons of our model’s estimates to past modeling studies and field measurements have shown encouraging results. Ultimately, this paper contributes to a broader understanding of acetone's role in the atmosphere.
Alok K. Samantaray, Priscilla A. Mooney, and Carla A. Vivacqua
Geosci. Model Dev., 17, 3321–3339, https://doi.org/10.5194/gmd-17-3321-2024, https://doi.org/10.5194/gmd-17-3321-2024, 2024
Short summary
Short summary
Any interpretation of climate model data requires a comprehensive evaluation of the model performance. Numerous error metrics exist for this purpose, and each focuses on a specific aspect of the relationship between reference and model data. Thus, a comprehensive evaluation demands the use of multiple error metrics. However, this can lead to confusion. We propose a clustering technique to reduce the number of error metrics needed and a composite error metric to simplify the interpretation.
Richard Maier, Fabian Jakub, Claudia Emde, Mihail Manev, Aiko Voigt, and Bernhard Mayer
Geosci. Model Dev., 17, 3357–3383, https://doi.org/10.5194/gmd-17-3357-2024, https://doi.org/10.5194/gmd-17-3357-2024, 2024
Short summary
Short summary
Based on the TenStream solver, we present a new method to accelerate 3D radiative transfer towards the speed of currently used 1D solvers. Using a shallow-cumulus-cloud time series, we evaluate the performance of this new solver in terms of both speed and accuracy. Compared to a 3D benchmark simulation, we show that our new solver is able to determine much more accurate irradiances and heating rates than a 1D δ-Eddington solver, even when operated with a similar computational demand.
Julia Maillard, Jean-Christophe Raut, and François Ravetta
Geosci. Model Dev., 17, 3303–3320, https://doi.org/10.5194/gmd-17-3303-2024, https://doi.org/10.5194/gmd-17-3303-2024, 2024
Short summary
Short summary
Atmospheric models struggle to reproduce the strong temperature inversions in the vicinity of the surface over forested areas in the Arctic winter. In this paper, we develop modified simplified versions of surface layer schemes widely used by the community. Our modifications are used to correct the fact that original schemes place strong limits on the turbulent collapse, leading to a lower surface temperature gradient at low wind speeds. Modified versions show a better performance.
Jana Fischereit, Henrik Vedel, Xiaoli Guo Larsén, Natalie E. Theeuwes, Gregor Giebel, and Eigil Kaas
Geosci. Model Dev., 17, 2855–2875, https://doi.org/10.5194/gmd-17-2855-2024, https://doi.org/10.5194/gmd-17-2855-2024, 2024
Short summary
Short summary
Wind farms impact local wind and turbulence. To incorporate these effects in weather forecasting, the explicit wake parameterization (EWP) is added to the forecasting model HARMONIE–AROME. We evaluate EWP using flight data above and downstream of wind farms, comparing it with an alternative wind farm parameterization and another weather model. Results affirm the correct implementation of EWP, emphasizing the necessity of accounting for wind farm effects in accurate weather forecasting.
Clément Bouvier, Daan van den Broek, Madeleine Ekblom, and Victoria A. Sinclair
Geosci. Model Dev., 17, 2961–2986, https://doi.org/10.5194/gmd-17-2961-2024, https://doi.org/10.5194/gmd-17-2961-2024, 2024
Short summary
Short summary
An analytical initial background state has been developed for moist baroclinic wave simulation on an aquaplanet and implemented into OpenIFS. Seven parameters can be controlled, which are used to generate the background states and the development of baroclinic waves. The meteorological and numerical stability has been assessed. Resulting baroclinic waves have proven to be realistic and sensitive to the jet's width.
Jelena Radović, Michal Belda, Jaroslav Resler, Kryštof Eben, Martin Bureš, Jan Geletič, Pavel Krč, Hynek Řezníček, and Vladimír Fuka
Geosci. Model Dev., 17, 2901–2927, https://doi.org/10.5194/gmd-17-2901-2024, https://doi.org/10.5194/gmd-17-2901-2024, 2024
Short summary
Short summary
Boundary conditions are of crucial importance for numerical model (e.g., PALM) validation studies and have a large influence on the model results, especially when studying the atmosphere of real, complex, and densely built urban environments. Our experiments with different driving conditions for the large-eddy simulation model PALM show its strong dependency on boundary conditions, which is important for the proper separation of errors coming from the boundary conditions and the model itself.
Sonya L. Fiddes, Marc D. Mallet, Alain Protat, Matthew T. Woodhouse, Simon P. Alexander, and Kalli Furtado
Geosci. Model Dev., 17, 2641–2662, https://doi.org/10.5194/gmd-17-2641-2024, https://doi.org/10.5194/gmd-17-2641-2024, 2024
Short summary
Short summary
In this study we present an evaluation that considers complex, non-linear systems in a holistic manner. This study uses XGBoost, a machine learning algorithm, to predict the simulated Southern Ocean shortwave radiation bias in the ACCESS model using cloud property biases as predictors. We then used a novel feature importance analysis to quantify the role that each cloud bias plays in predicting the radiative bias, laying the foundation for advanced Earth system model evaluation and development.
Gaurav Govardhan, Sachin D. Ghude, Rajesh Kumar, Sumit Sharma, Preeti Gunwani, Chinmay Jena, Prafull Yadav, Shubhangi Ingle, Sreyashi Debnath, Pooja Pawar, Prodip Acharja, Rajmal Jat, Gayatry Kalita, Rupal Ambulkar, Santosh Kulkarni, Akshara Kaginalkar, Vijay K. Soni, Ravi S. Nanjundiah, and Madhavan Rajeevan
Geosci. Model Dev., 17, 2617–2640, https://doi.org/10.5194/gmd-17-2617-2024, https://doi.org/10.5194/gmd-17-2617-2024, 2024
Short summary
Short summary
A newly developed air quality forecasting framework, Decision Support System (DSS), for air quality management in Delhi, India, provides source attribution with numerous emission reduction scenarios besides forecasts. DSS shows that during post-monsoon and winter seasons, Delhi and its neighboring districts contribute to 30 %–40 % each to pollution in Delhi. On average, a 40 % reduction in the emissions in Delhi and the surrounding districts would result in a 24 % reduction in Delhi's pollution.
Simon Rosanka, Holger Tost, Rolf Sander, Patrick Jöckel, Astrid Kerkweg, and Domenico Taraborrelli
Geosci. Model Dev., 17, 2597–2615, https://doi.org/10.5194/gmd-17-2597-2024, https://doi.org/10.5194/gmd-17-2597-2024, 2024
Short summary
Short summary
The capabilities of the Modular Earth Submodel System (MESSy) are extended to account for non-equilibrium aqueous-phase chemistry in the representation of deliquescent aerosols. When applying the new development in a global simulation, we find that MESSy's bias in modelling routinely observed reduced inorganic aerosol mass concentrations, especially in the United States. Furthermore, the representation of fine-aerosol pH is particularly improved in the marine boundary layer.
Junyu Li, Yuxin Wang, Lilong Liu, Yibin Yao, Liangke Huang, and Feijuan Li
Geosci. Model Dev., 17, 2569–2581, https://doi.org/10.5194/gmd-17-2569-2024, https://doi.org/10.5194/gmd-17-2569-2024, 2024
Short summary
Short summary
In this study, we have developed a model (RF-PWV) to characterize precipitable water vapor (PWV) variation with altitude in the study area. RF-PWV can significantly reduce errors in vertical correction, enhance PWV fusion product accuracy, and provide insights into PWV vertical distribution, thereby contributing to climate research.
Rolf Sander
Geosci. Model Dev., 17, 2419–2425, https://doi.org/10.5194/gmd-17-2419-2024, https://doi.org/10.5194/gmd-17-2419-2024, 2024
Short summary
Short summary
The open-source software MEXPLORER 1.0.0 is presented here. The program can be used to analyze, reduce, and visualize complex chemical reaction mechanisms. The mathematics behind the tool is based on graph theory: chemical species are represented as vertices, and reactions as edges. MEXPLORER is a community model published under the GNU General Public License.
Leonardo Olivetti and Gabriele Messori
Geosci. Model Dev., 17, 2347–2358, https://doi.org/10.5194/gmd-17-2347-2024, https://doi.org/10.5194/gmd-17-2347-2024, 2024
Short summary
Short summary
In the last decades, weather forecasting up to 15 d into the future has been dominated by physics-based numerical models. Recently, deep learning models have challenged this paradigm. However, the latter models may struggle when forecasting weather extremes. In this article, we argue for deep learning models specifically designed to handle extreme events, and we propose a foundational framework to develop such models.
Stefan Rahimi, Lei Huang, Jesse Norris, Alex Hall, Naomi Goldenson, Will Krantz, Benjamin Bass, Chad Thackeray, Henry Lin, Di Chen, Eli Dennis, Ethan Collins, Zachary J. Lebo, Emily Slinskey, Sara Graves, Surabhi Biyani, Bowen Wang, Stephen Cropper, and the UCLA Center for Climate Science Team
Geosci. Model Dev., 17, 2265–2286, https://doi.org/10.5194/gmd-17-2265-2024, https://doi.org/10.5194/gmd-17-2265-2024, 2024
Short summary
Short summary
Here, we project future climate across the western United States through the end of the 21st century using a regional climate model, embedded within 16 latest-generation global climate models, to provide the community with a high-resolution physically based ensemble of climate data for use at local scales. Strengths and weaknesses of the data are frankly discussed as we overview the downscaled dataset.
Romain Pilon and Daniela I. V. Domeisen
Geosci. Model Dev., 17, 2247–2264, https://doi.org/10.5194/gmd-17-2247-2024, https://doi.org/10.5194/gmd-17-2247-2024, 2024
Short summary
Short summary
This paper introduces a new method for detecting atmospheric cloud bands to identify long convective cloud bands that extend from the tropics to the midlatitudes. The algorithm allows for easy use and enables researchers to study the life cycle and climatology of cloud bands and associated rainfall. This method provides insights into the large-scale processes involved in cloud band formation and their connections between different regions, as well as differences across ocean basins.
Salvatore Larosa, Domenico Cimini, Donatello Gallucci, Saverio Teodosio Nilo, and Filomena Romano
Geosci. Model Dev., 17, 2053–2076, https://doi.org/10.5194/gmd-17-2053-2024, https://doi.org/10.5194/gmd-17-2053-2024, 2024
Short summary
Short summary
PyRTlib is an attractive educational tool because it provides a flexible and user-friendly way to broadly simulate how electromagnetic radiation travels through the atmosphere as it interacts with atmospheric constituents (such as gases, aerosols, and hydrometeors). PyRTlib is a so-called radiative transfer model; these are commonly used to simulate and understand remote sensing observations from ground-based, airborne, or satellite instruments.
Joffrey Dumont Le Brazidec, Pierre Vanderbecken, Alban Farchi, Grégoire Broquet, Gerrit Kuhlmann, and Marc Bocquet
Geosci. Model Dev., 17, 1995–2014, https://doi.org/10.5194/gmd-17-1995-2024, https://doi.org/10.5194/gmd-17-1995-2024, 2024
Short summary
Short summary
Our research presents an innovative approach to estimating power plant CO2 emissions from satellite images of the corresponding plumes such as those from the forthcoming CO2M satellite constellation. The exploitation of these images is challenging due to noise and meteorological uncertainties. To overcome these obstacles, we use a deep learning neural network trained on simulated CO2 images. Our method outperforms alternatives, providing a positive perspective for the analysis of CO2M images.
Kyoung-Min Kim, Si-Wan Kim, Seunghwan Seo, Donald R. Blake, Seogju Cho, James H. Crawford, Louisa K. Emmons, Alan Fried, Jay R. Herman, Jinkyu Hong, Jinsang Jung, Gabriele G. Pfister, Andrew J. Weinheimer, Jung-Hun Woo, and Qiang Zhang
Geosci. Model Dev., 17, 1931–1955, https://doi.org/10.5194/gmd-17-1931-2024, https://doi.org/10.5194/gmd-17-1931-2024, 2024
Short summary
Short summary
Three emission inventories were evaluated for East Asia using data acquired during a field campaign in 2016. The inventories successfully reproduced the daily variations of ozone and nitrogen dioxide. However, the spatial distributions of model ozone did not fully agree with the observations. Additionally, all simulations underestimated carbon monoxide and volatile organic compound (VOC) levels. Increasing VOC emissions over South Korea resulted in improved ozone simulations.
Sanam Noreen Vardag and Robert Maiwald
Geosci. Model Dev., 17, 1885–1902, https://doi.org/10.5194/gmd-17-1885-2024, https://doi.org/10.5194/gmd-17-1885-2024, 2024
Short summary
Short summary
We use the atmospheric transport model GRAMM/GRAL in a Bayesian inversion to estimate urban CO2 emissions on a neighbourhood scale. We analyse the effect of varying number, precision and location of CO2 sensors for CO2 flux estimation. We further test the inclusion of co-emitted species and correlation in the inversion. The study showcases the general usefulness of GRAMM/GRAL in measurement network design.
Abhishek Savita, Joakim Kjellsson, Robin Pilch Kedzierski, Mojib Latif, Tabea Rahm, Sebastian Wahl, and Wonsun Park
Geosci. Model Dev., 17, 1813–1829, https://doi.org/10.5194/gmd-17-1813-2024, https://doi.org/10.5194/gmd-17-1813-2024, 2024
Short summary
Short summary
The OpenIFS model is used to examine the impact of horizontal resolutions (HR) and model time steps. We find that the surface wind biases over the oceans, in particular the Southern Ocean, are sensitive to the model time step and HR, with the HR having the smallest biases. When using a coarse-resolution model with a shorter time step, a similar improvement is also found. Climate biases can be reduced in the OpenIFS model at a cheaper cost by reducing the time step rather than increasing the HR.
Ferdinand Briegel, Jonas Wehrle, Dirk Schindler, and Andreas Christen
Geosci. Model Dev., 17, 1667–1688, https://doi.org/10.5194/gmd-17-1667-2024, https://doi.org/10.5194/gmd-17-1667-2024, 2024
Short summary
Short summary
We present a new approach to model heat stress in cities using artificial intelligence (AI). We show that the AI model is fast in terms of prediction but accurate when evaluated with measurements. The fast-predictive AI model enables several new potential applications, including heat stress prediction and warning; downscaling of potential future climates; evaluation of adaptation effectiveness; and, more fundamentally, development of guidelines to support urban planning and policymaking.
Hauke Schmidt, Sebastian Rast, Jiawei Bao, Amrit Cassim, Shih-Wei Fang, Diego Jimenez-de la Cuesta, Paul Keil, Lukas Kluft, Clarissa Kroll, Theresa Lang, Ulrike Niemeier, Andrea Schneidereit, Andrew I. L. Williams, and Bjorn Stevens
Geosci. Model Dev., 17, 1563–1584, https://doi.org/10.5194/gmd-17-1563-2024, https://doi.org/10.5194/gmd-17-1563-2024, 2024
Short summary
Short summary
A recent development in numerical simulations of the global atmosphere is the increase in horizontal resolution to grid spacings of a few kilometers. However, the vertical grid spacing of these models has not been reduced at the same rate as the horizontal grid spacing. Here, we assess the effects of much finer vertical grid spacings, in particular the impacts on cloud quantities and the atmospheric energy balance.
Tao Zheng, Sha Feng, Jeffrey Steward, Xiaoxu Tian, David Baker, and Martin Baxter
Geosci. Model Dev., 17, 1543–1562, https://doi.org/10.5194/gmd-17-1543-2024, https://doi.org/10.5194/gmd-17-1543-2024, 2024
Short summary
Short summary
The tangent linear and adjoint models have been successfully implemented in the MPAS-CO2 system, which has undergone rigorous accuracy testing. This development lays the groundwork for a global carbon flux data assimilation system, which offers the flexibility of high-resolution focus on specific areas, while maintaining a coarser resolution elsewhere. This approach significantly reduces computational costs and is thus perfectly suited for future CO2 geostationery and imager satellites.
Kelvin H. Bates, Mathew J. Evans, Barron H. Henderson, and Daniel J. Jacob
Geosci. Model Dev., 17, 1511–1524, https://doi.org/10.5194/gmd-17-1511-2024, https://doi.org/10.5194/gmd-17-1511-2024, 2024
Short summary
Short summary
Accurate representation of rates and products of chemical reactions in atmospheric models is crucial for simulating concentrations of pollutants and climate forcers. We update the widely used GEOS-Chem atmospheric chemistry model with reaction parameters from recent compilations of experimental data and demonstrate the implications for key atmospheric chemical species. The updates decrease tropospheric CO mixing ratios and increase stratospheric nitrogen oxide mixing ratios, among other changes.
François Roberge, Alejandro Di Luca, René Laprise, Philippe Lucas-Picher, and Julie Thériault
Geosci. Model Dev., 17, 1497–1510, https://doi.org/10.5194/gmd-17-1497-2024, https://doi.org/10.5194/gmd-17-1497-2024, 2024
Short summary
Short summary
Our study addresses a challenge in dynamical downscaling using regional climate models, focusing on the lack of small-scale features near the boundaries. We introduce a method to identify this “spatial spin-up” in precipitation simulations. Results show spin-up distances up to 300 km, varying by season and driving variable. Double nesting with comprehensive variables (e.g. microphysical variables) offers advantages. Findings will help optimize simulations for better climate projections.
Eloisa Raluy-López, Juan Pedro Montávez, and Pedro Jiménez-Guerrero
Geosci. Model Dev., 17, 1469–1495, https://doi.org/10.5194/gmd-17-1469-2024, https://doi.org/10.5194/gmd-17-1469-2024, 2024
Short summary
Short summary
Atmospheric rivers (ARs) represent a significant source of water but are also related to extreme precipitation events. Here, we present a new regional-scale AR identification algorithm and apply it to three simulations that include aerosol interactions at different levels. The results show that aerosols modify the intensity and trajectory of ARs and redistribute the AR-related precipitation. Thus, the correct inclusion of aerosol effects is important in the simulation of AR behavior.
Sofía Gómez Maqueo Anaya, Dietrich Althausen, Matthias Faust, Holger Baars, Bernd Heinold, Julian Hofer, Ina Tegen, Albert Ansmann, Ronny Engelmann, Annett Skupin, Birgit Heese, and Kerstin Schepanski
Geosci. Model Dev., 17, 1271–1295, https://doi.org/10.5194/gmd-17-1271-2024, https://doi.org/10.5194/gmd-17-1271-2024, 2024
Short summary
Short summary
Mineral dust aerosol particles vary greatly in their composition depending on source region, which leads to different physicochemical properties. Most atmosphere–aerosol models consider mineral dust aerosols to be compositionally homogeneous, which ultimately increases model uncertainty. Here, we present an approach to explicitly consider the heterogeneity of the mineralogical composition for simulations of the Saharan atmospheric dust cycle with regard to dust transport towards the Atlantic.
Alexandros Milousis, Alexandra P. Tsimpidi, Holger Tost, Spyros N. Pandis, Athanasios Nenes, Astrid Kiendler-Scharr, and Vlassis A. Karydis
Geosci. Model Dev., 17, 1111–1131, https://doi.org/10.5194/gmd-17-1111-2024, https://doi.org/10.5194/gmd-17-1111-2024, 2024
Short summary
Short summary
This study aims to evaluate the newly developed ISORROPIA-lite aerosol thermodynamic module within the EMAC model and explore discrepancies in global atmospheric simulations of aerosol composition and acidity by utilizing different aerosol phase states. Even though local differences were found in regions where the RH ranged from 20 % to 60 %, on a global scale the results are similar. Therefore, ISORROPIA-lite can be a reliable and computationally effective alternative to ISORROPIA II in EMAC.
Marie-Adèle Magnaldo, Quentin Libois, Sébastien Riette, and Christine Lac
Geosci. Model Dev., 17, 1091–1109, https://doi.org/10.5194/gmd-17-1091-2024, https://doi.org/10.5194/gmd-17-1091-2024, 2024
Short summary
Short summary
With the worldwide development of the solar energy sector, the need for reliable solar radiation forecasts has significantly increased. However, meteorological models that predict, among others things, solar radiation have errors. Therefore, we wanted to know in which situtaions these errors are most significant. We found that errors mostly occur in cloudy situations, and different errors were highlighted depending on the cloud altitude. Several potential sources of errors were identified.
Dongqi Lin, Jiawei Zhang, Basit Khan, Marwan Katurji, and Laura E. Revell
Geosci. Model Dev., 17, 815–845, https://doi.org/10.5194/gmd-17-815-2024, https://doi.org/10.5194/gmd-17-815-2024, 2024
Short summary
Short summary
GEO4PALM is an open-source tool to generate static input for the Parallelized Large-Eddy Simulation (PALM) model system. Geospatial static input is essential for realistic PALM simulations. However, existing tools fail to generate PALM's geospatial static input for most regions. GEO4PALM is compatible with diverse geospatial data sources and provides access to free data sets. In addition, this paper presents two application examples, which show successful PALM simulations using GEO4PALM.
Piotr Zmijewski, Piotr Dziekan, and Hanna Pawlowska
Geosci. Model Dev., 17, 759–780, https://doi.org/10.5194/gmd-17-759-2024, https://doi.org/10.5194/gmd-17-759-2024, 2024
Short summary
Short summary
In computer simulations of clouds it is necessary to model the myriad of droplets that constitute a cloud. A popular method for this is to use so-called super-droplets (SDs), each representing many real droplets. It has remained a challenge to model collisions of SDs. We study how precipitation in a cumulus cloud depends on the number of SDs. Surprisingly, we do not find convergence in mean precipitation even for numbers of SDs much larger than typically used in simulations.
Roya Ghahreman, Wanmin Gong, Paul A. Makar, Alexandru Lupu, Amanda Cole, Kulbir Banwait, Colin Lee, and Ayodeji Akingunola
Geosci. Model Dev., 17, 685–707, https://doi.org/10.5194/gmd-17-685-2024, https://doi.org/10.5194/gmd-17-685-2024, 2024
Short summary
Short summary
The article explores the impact of different representations of below-cloud scavenging on model biases. A new scavenging scheme and precipitation-phase partitioning improve the model's performance, with better SO42- scavenging and wet deposition of NO3- and NH4+.
Daisuke Goto, Tatsuya Seiki, Kentaroh Suzuki, Hisashi Yashiro, and Toshihiko Takemura
Geosci. Model Dev., 17, 651–684, https://doi.org/10.5194/gmd-17-651-2024, https://doi.org/10.5194/gmd-17-651-2024, 2024
Short summary
Short summary
Global climate models with coarse grid sizes include uncertainties about the processes in aerosol–cloud–precipitation interactions. To reduce these uncertainties, here we performed numerical simulations using a new version of our global aerosol transport model with a finer grid size over a longer period than in our previous study. As a result, we found that the cloud microphysics module influences the aerosol distributions through both aerosol wet deposition and aerosol–cloud interactions.
Alexander de Meij, Cornelis Cuvelier, Philippe Thunis, Enrico Pisoni, and Bertrand Bessagnet
Geosci. Model Dev., 17, 587–606, https://doi.org/10.5194/gmd-17-587-2024, https://doi.org/10.5194/gmd-17-587-2024, 2024
Short summary
Short summary
In our study the robustness of the model responses to emission reductions in the EU is assessed when the emission data are changed. Our findings are particularly important to better understand the uncertainties associated to the emission inventories and how these uncertainties impact the level of accuracy of the resulting air quality modelling, which is a key for designing air quality plans. Also crucial is the choice of indicator to avoid misleading interpretations of the results.
Haiqin Li, Georg A. Grell, Ravan Ahmadov, Li Zhang, Shan Sun, Jordan Schnell, and Ning Wang
Geosci. Model Dev., 17, 607–619, https://doi.org/10.5194/gmd-17-607-2024, https://doi.org/10.5194/gmd-17-607-2024, 2024
Short summary
Short summary
We developed a simple and realistic method to provide aerosol emissions for aerosol-aware microphysics in a numerical weather forecast model. The cloud-radiation differences between the experimental (EXP) and control (CTL) experiments responded to the aerosol differences. The strong positive precipitation biases over North America and Europe from the CTL run were significantly reduced in the EXP run. This study shows that a realistic representation of aerosol emissions should be considered.
Nathan Patrick Arnold
Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2023-245, https://doi.org/10.5194/gmd-2023-245, 2024
Revised manuscript accepted for GMD
Short summary
Short summary
Earth System Models often represent the land surface at smaller scales than the atmosphere, but surface-atmosphere coupling uses only aggregated surface properties. This study presents a method to allow heterogeneous surface properties to modify boundary layer updrafts. The method is tested in single column experiments. Updraft properties are found to reasonably covary with surface conditions, and simulated boundary layer variability is enhanced over more heterogeneous land surfaces.
Giancarlo Ciarelli, Sara Tahvonen, Arineh Cholakian, Manuel Bettineschi, Bruno Vitali, Tuukka Petäjä, and Federico Bianchi
Geosci. Model Dev., 17, 545–565, https://doi.org/10.5194/gmd-17-545-2024, https://doi.org/10.5194/gmd-17-545-2024, 2024
Short summary
Short summary
The terrestrial ecosystem releases large quantities of biogenic gases in the Earth's Atmosphere. These gases can effectively be converted into so-called biogenic aerosol particles and, eventually, affect the Earth's climate. Climate prediction varies greatly depending on how these processes are represented in model simulations. In this study, we present a detailed model evaluation analysis aimed at understanding the main source of uncertainty in predicting the formation of biogenic aerosols.
Jiachen Liu, Eric Chen, and Shannon L. Capps
Geosci. Model Dev., 17, 567–585, https://doi.org/10.5194/gmd-17-567-2024, https://doi.org/10.5194/gmd-17-567-2024, 2024
Short summary
Short summary
Air pollution harms human life and ecosystems, but its sources are complex. Scientists and policy makers use air pollution models to advance knowledge and inform control strategies. We implemented a recently developed numeral system to relate any set of model inputs, like pollutant emissions from a given activity, to all model outputs, like concentrations of pollutants harming human health. This approach will be straightforward to update when scientists discover new processes in the atmosphere.
Cited articles
Abadi, M., Barham, P., Chen, J., Chen, Z., Davis, A., Dean, J., Devin, M., Ghemawat, S., Irving, G., Isard, M., Jia, Y., Jozefowicz, R., Kaiser, L., Kudlur, M., Levenberg, J., Mane, D., Monga, R., Moore, S., Murray, D., Olah, C., Schuster, M., Shlens, J., Steiner, B., Sutskever, I., Talwar, K., Tucker, P., Vanhoucke, V., Vasudevan, V., Viegas, F., Vinyals, O., Warden, P., Wattenberg, M., Wicke, M., Yu, Y., and Zheng, X.: Tensorflow: A system for large-scale machine learning, in: 12th USENIX Symposium on Operating Systems Design and Implementation (OSDI 16), Savannah, GA, US, 2–4 November 2016, 265–283, https://arxiv.org/abs/1603.04467 (last access: 6 September 2023), 2016. a
Alexeev, D.: PyTorch bindings for Fortran (v0.4), GitHub [code], https://github.com/alexeedm/pytorch-fortran (last access: 6 September 2023), 2023. a
Arnold, C., Sharma, S., and Weigel, T.: DKRZ-AIM/dkrz-hereon-icon-superdropnet: Integrating SuperdropNet (v0.1.0), Zenodo [code], https://doi.org/10.5281/zenodo.10069121, 2023a. a
Arnold, C., Sharma, S., and Weigel, T.: ICON Code v 2.6.5 including coupling schemes for integrating SuperdropNet, Zenodo [code], https://doi.org/10.5281/zenodo.8348256, 2023b. a
Arnold, C., Sharma, S., and Weigel, T.: Data set for: Efficient and Stable Coupling of the SuperdropNet Deep Learning-based Cloud Microphysics (v0.1.0) to the ICON Climate and Weather Model (v2.6.5), Zenodo [data set], https://doi.org/10.5281/zenodo.8348266, 2023c. a
Belochitski, A. and Krasnopolsky, V.: Robustness of neural network emulations of radiative transfer parameterizations in a state-of-the-art general circulation model, Geosci. Model Dev., 14, 7425–7437, https://doi.org/10.5194/gmd-14-7425-2021, 2021. a
Beucler, T., Pritchard, M., Rasp, S., Ott, J., Baldi, P., and Gentine, P.: Enforcing Analytic Constraints in Neural Networks Emulating Physical Systems, Phys. Rev. Lett., 126, 098302, https://doi.org/10.1103/PhysRevLett.126.098302, 2021. a
Bonanni, A., Hawkes, J., and Quintino, T.: infero: A lower-level API for Machine Learning inference in operations (version 0.2.0), GitHub [code], https://github.com/ecmwf/infero (last access: 6 September 2023), 2022. a
Brdar, S. and Seifert, A.: McSnow: A Monte-Carlo Particle Model for Riming and Aggregation of Ice Particles in a Multidimensional Microphysical Phase Space, J. Adv. Model. Earth Sy., 10, 187–206, https://doi.org/10.1002/2017MS001167, 2018. a, b, c, d
Brenowitz, N.: Call Python from Fortran (v0.2.1), Zenodo [code], https://doi.org/10.5281/zenodo.7779572, 2023. a
Brenowitz, N. D. and Bretherton, C. S.: Prognostic Validation of a Neural Network Unified Physics Parameterization, Geophys. Res. Lett., 45, 6289–6298, https://doi.org/10.1029/2018GL078510, 2018. a, b
Brenowitz, N. D. and Bretherton, C. S.: Spatially Extended Tests of a Neural Network Parametrization Trained by Coarse-Graining, J. Adv. Model. Earth Sy., 11, 2728–2744, https://doi.org/10.1029/2019MS001711, 2019. a, b
Brenowitz, N. D., Beucler, T., Pritchard, M., and Bretherton, C. S.: Interpreting and Stabilizing Machine-Learning Parametrizations of Convection, J. Atmos. Sci., 77, 4357–4375, https://doi.org/10.1175/JAS-D-20-0082.1, 2020a. a
Brenowitz, N. D., Henn, B., McGibbon, J., Clark, S. K., Kwa, A., Perkins, W. A., Watt-Meyer, O., and Bretherton, C. S.: Machine Learning Climate Model Dynamics: Offline versus Online Performance, arXiv [preprint], https://doi.org/10.48550/arXiv.2011.03081, 2020b. a
Brenowitz, N. D., Perkins, W. A., Nugent, J. M., Watt-Meyer, O., Clark, S. K., Kwa, A., Henn, B., McGibbon, J., and Bretherton, C. S.: Emulating Fast Processes in Climate Models, in: Machine Learning for the Physical Sciences, NEURIPS Workshop, arXiv [preprint], https://doi.org/10.48550/arXiv.2211.10774, 2022. a
Chantry, M., Hatfield, S., Dueben, P., Polichtchouk, I., and Palmer, T.: Machine learning emulation of gravity wave drag in numerical weather forecasting, J. Adv. Model. Earth Sy., 13, e2021MS002477, https://doi.org/10.1029/2021MS002477, 2021. a
Chevallier, F., Morcrette, J.-J., Chéruy, F., and Scott, N.: Use of a neural-network-based long-wave radiative-transfer scheme in the ECMWF atmospheric model, Q. J. Roy. Meteor. Soc., 126, 761–776, https://doi.org/10.1002/qj.49712656318, 2000. a
Chollet, F. et al.: Keras (v2.14.0), GitHub [code], https://github.com/fchollet/keras, last access: 6 September 2023. a
Christensen, H. and Zanna, L.: Parametrization in Weather and Climate Models, in: Oxford Research Encyclopedia of Climate Science, Oxford University Press, ISBN 978-0-19-022862-0, https://doi.org/10.1093/acrefore/9780190228620.013.826, 2022. a
Curcic, M.: A parallel Fortran framework for neural networks and deep learning, arXiv [preprint], https://doi.org/10.48550/arXiv.1902.06714, 2019. a
Dong, W., Fritts, D. C., Liu, A. Z., Lund, T. S., Liu, H.-L., and Snively, J.: Accelerating Atmospheric Gravity Wave Simulations Using Machine Learning: Kelvin-Helmholtz Instability and Mountain Wave Sources Driving Gravity Wave Breaking and Secondary Gravity Wave Generation, Geophys. Res. Lett., 50, e2023GL104668, https://doi.org/10.1029/2023GL104668, 2023. a
Dueben, P., Modigliani, U., Geer, A., Siemen, S., Pappenberger, F., Bauer, P., Brown, A., Palkovic, M., Raoult, B., Wedi, N., and Baousis, V.: Machine learning at ECMWF: A roadmap for the next 10 years, ECMWF Technical Memoranda, ECMWF, https://doi.org/10.21957/ge7ckgm, 2021. a
Dueben, P. D., Schultz, M. G., Chantry, M., Gagne, D. J., Hall, D. M., and McGovern, A.: Challenges and Benchmark Datasets for Machine Learning in the Atmospheric Sciences: Definition, Status, and Outlook, Artificial Intelligence for the Earth Systems, 1, e210002, https://doi.org/10.1175/AIES-D-21-0002.1, 2022. a
Elafrou, A., Orchard, D., and Cliffard, S.: fortran-pytorch-lib (commit: ffe833b66a6e1ce1c6cf023708d1f351a3a11f8b), GitHub [code], https://github.com/Cambridge-ICCS/fortran-pytorch-lib, last access: 6 September 2023. a
Gentine, P., Pritchard, M., Rasp, S., Reinaudi, G., and Yacalis, G.: Could Machine Learning Break the Convection Parameterization Deadlock?, Geophys. Res. Lett., 45, 5742–5751, https://doi.org/10.1029/2018GL078202, 2018. a
Grundner, A., Beucler, T., Gentine, P., Iglesias-Suarez, F., Giorgetta, M. A., and Eyring, V.: Deep Learning Based Cloud Cover Parameterization for ICON, J. Adv. Model. Earth Sy., 14, e2021MS002959, https://doi.org/10.1029/2021MS002959, 2022. a
Hanke, M., Redler, R., Holfeld, T., and Yastremsky, M.: YAC 1.2.0: new aspects for coupling software in Earth system modelling, Geosci. Model Dev., 9, 2755–2769, https://doi.org/10.5194/gmd-9-2755-2016, 2016. a
Hanke, M., Dreier, N.-A., and Redler, R.: YetAnotherCoupler (YAC) (version 2.6.1) [code], https://dkrz-sw.gitlab-pages.dkrz.de/yac/, last access: 6 September 2023. a
Irrgang, C., Boers, N., Sonnewald, M., Barnes, E. A., Kadow, C., Staneva, J., and Saynisch-Wagner, J.: Towards neural Earth system modelling by integrating artificial intelligence in Earth system science, Nature Machine Intelligence, 3, 667–674, https://doi.org/10.1038/s42256-021-00374-3, 2021. a
Krasnopolsky, V. M., Fox-Rabinovitz, M. S., and Chalikov, D. V.: New approach to calculation of atmospheric model physics: Accurate and fast neural network emulation of longwave radiation in a climate model, Mon. Weather Rev., 133, 1370–1383, https://doi.org/10.1175/MWR2923.1, 2005. a
McGovern, A., Lagerquist, R., Gagne, D. J., Jergensen, G. E., Elmore, K. L., Homeyer, C. R., and Smith, T.: Making the Black Box More Transparent: Understanding the Physical Implications of Machine Learning, B. Am. Meteorol. Soc., 100, 2175–2199, https://doi.org/10.1175/BAMS-D-18-0195.1, 2019. a
Meyer, D., Hogan, R. J., Dueben, P. D., and Mason, S. L.: Machine Learning Emulation of 3D Cloud Radiative Effects, J. Adv. Model. Earth Sy., 14, e2021MS002550, https://doi.org/10.1029/2021MS002550, 2022. a
Morrison, H., van Lier‐Walqui, M., Fridlind, A. M., Grabowski, W. W., Harrington, J. Y., Hoose, C., Korolev, A., Kumjian, M. R., Milbrandt, J. A., Pawlowska, H., Posselt, D. J., Prat, O. P., Reimel, K. J., Shima, S.-I., van Diedenhoven, B., and Xue, L.: Confronting the Challenge of Modeling Cloud and Precipitation Microphysics, J. Adv. Model. Earth Sy., 12, e2019MS001689, https://doi.org/10.1029/2019MS001689, 2020. a
Mu, B., Chen, L., Yuan, S., and Qin, B.: A radiative transfer deep learning model coupled into WRF with a generic fortran torch adaptor, Frontiers in Earth Science, 11, https://doi.org/10.3389/feart.2023.1149566, 2023. a
Nowack, P., Braesicke, P., Haigh, J., Abraham, N. L., Pyle, J., and Voulgarakis, A.: Using machine learning to build temperature-based ozone parameterizations for climate sensitivity simulations, Environ. Res. Lett., 13, 104016, https://doi.org/10.1088/1748-9326/aae2be, 2018. a
Ott, J., Pritchard, M., Best, N., Linstead, E., Curcic, M., and Baldi, P.: A Fortran-Keras Deep Learning Bridge for Scientific Computing, Scientific Programming, 2020, 8888811, https://doi.org/10.1155/2020/8888811, 2020. a
Partee, S., Ellis, M., Rigazzi, A., Shao, A. E., Bachman, S., Marques, G., and Robbins, B.: Using Machine Learning at scale in numerical simulations with SmartSim: An application to ocean climate modeling, J. Comput. Sci.-Neth., 62, 101707, https://doi.org/10.1016/j.jocs.2022.101707, 2022. a, b
Paszke, A., Gross, S., Massa, F., Lerer, A., Bradbury, J., Chanan, G., Killeen, T., Lin, Z., Gimelshein, N., Antiga, L., Desmaison, A., Kopf, A., Yang, E., DeVito, Z., Raison, M., Tejani, A., Chilamkurthy, S., Steiner, B., Fang, L., Bai, J., and Chintala, S.: PyTorch: An Imperative Style, High-Performance Deep Learning Library, in: Advances in Neural Information Processing Systems 32, Vancouver, Canada, 8–14 December 2019, Curran Associates, Inc., 8024–8035, http://papers.neurips.cc/paper/9015-pytorch-an-imperative-style-high-performance-deep-learning-library.pdf (last access: 6 September 2023), 2019. a
Qu, Y. and Shi, X.: Can a Machine Learning–Enabled Numerical Model Help Extend Effective Forecast Range through Consistently Trained Subgrid-Scale Models?, Artif. Intell. Earth Syst., 2, e220050, https://doi.org/10.1175/AIES-D-22-0050.1, 2023. a
Rasp, S.: Coupled online learning as a way to tackle instabilities and biases in neural network parameterizations: general algorithms and Lorenz 96 case study (v1.0), Geosci. Model Dev., 13, 2185–2196, https://doi.org/10.5194/gmd-13-2185-2020, 2020. a, b
Rasp, S., Pritchard, M. S., and Gentine, P.: Deep learning to represent subgrid processes in climate models, P. Natl. Acad. Sci. USA, 115, 9684–9689, https://doi.org/10.1073/pnas.1810286115, 2018. a
Rigo, A. and Fijalkowski, M.: C Foreign Function Interface for Python, CFFI [code], https://cffi.readthedocs.io/en/release-1.14/ (last access: 6 September 2023), 2018. a
Berkeley Lab: Inference Engine (v0.10.0), GitHub [code], https://github.com/BerkeleyLab/inference-engine/, 2023. a
Seifert, A. and Beheng, K.: A two-moment cloud microphysics parameterization for mixed-phase clouds. Part 1: Model description, Meteorol. Atmos. Phys., 92, 45–66, https://doi.org/10.1007/s00703-005-0112-4, 2006. a
Seifert, A. and Beheng, K. D.: A double-moment parameterization for simulating autoconversion, accretion and selfcollection, Atmos. Res., 59–60, 265–281, https://doi.org/10.1016/s0169-8095(01)00126-0, 2001. a, b
Seifert, A. and Rasp, S.: Potential and Limitations of Machine Learning for Modeling Warm-Rain Cloud Microphysical Processes, J. Adv. Model. Earth Sy., 12, e2020MS002301, https://doi.org/10.1029/2020MS002301, 2020. a, b
Sharma, S. and Greenberg, D.: SuperdropNet: a Stable and Accurate Machine Learning Proxy for Droplet-based Cloud Microphysics, arXiv [preprint], https://doi.org/10.48550/arXiv.2402.18354, 2024. a, b, c
Shima, S., Kusano, K., Kawano, A., Sugiyama, T., and Kawahara, S.: The super-droplet method for the numerical simulation of clouds and precipitation: a particle-based and probabilistic microphysics model coupled with a non-hydrostatic model, Q. J. Roy. Meteor. Soc., 135, 1307–1320, https://doi.org/10.1002/qj.441, 2009. a
Sonnewald, M., Lguensat, R., Jones, D. C., Dueben, P. D., Brajard, J., and Balaji, V.: Bridging observations, theory and numerical simulation of the ocean using machine learning, Environ. Res. Lett., 16, 073008, https://doi.org/10.1088/1748-9326/ac0eb0, 2021. a
Yuval, J. and O’Gorman, P. A.: Neural-Network Parameterization of Subgrid Momentum Transport in the Atmosphere, J. Adv. Model. Earth Sy., 15, e2023MS003606, https://doi.org/10.1029/2023MS003606, 2023. a
Yuval, J., O'Gorman, P. A., and Hill, C. N.: Use of Neural Networks for Stable, Accurate and Physically Consistent Parameterization of Subgrid Atmospheric Processes With Good Performance at Reduced Precision, Geophys. Res. Lett., 48, e2020GL091363, https://doi.org/10.1029/2020GL091363, 2021. a, b
Zhong, X., Ma, Z., Yao, Y., Xu, L., Wu, Y., and Wang, Z.: WRF–ML v1.0: a bridge between WRF v4.3 and machine learning parameterizations and its application to atmospheric radiative transfer, Geosci. Model Dev., 16, 199–209, https://doi.org/10.5194/gmd-16-199-2023, 2023. a
Short summary
In atmospheric models, rain formation is simplified to be computationally efficient. We trained a machine learning model, SuperdropNet, to emulate warm-rain formation based on super-droplet simulations. Here, we couple SuperdropNet with an atmospheric model in a warm-bubble experiment and find that the coupled simulation runs stable and produces reasonable results, making SuperdropNet a viable ML proxy for droplet simulations. We also present a comprehensive benchmark for coupling architectures.
In atmospheric models, rain formation is simplified to be computationally efficient. We trained...
