Articles | Volume 16, issue 11
https://doi.org/10.5194/gmd-16-3241-2023
© Author(s) 2023. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
https://doi.org/10.5194/gmd-16-3241-2023
© Author(s) 2023. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Implementation of a machine-learned gas optics parameterization in the ECMWF Integrated Forecasting System: RRTMGP-NN 2.0
Weather Research, Danish Meteorological Institute, Copenhagen, Denmark
Robin J. Hogan
Research Department, European Centre for Medium-Range Weather Forecasts, Reading, UK
Department of Meteorology, University of Reading, Reading, UK
Related authors
No articles found.
Kaah P. Menang, Stefan A. Buehler, Lukas Kluft, Robin J. Hogan, and Florian E. Roemer
Atmos. Chem. Phys., 25, 11689–11701, https://doi.org/10.5194/acp-25-11689-2025, https://doi.org/10.5194/acp-25-11689-2025, 2025
Short summary
Short summary
We investigated the impact of the shortwave water vapour continuum absorption on clear-sky shortwave radiative feedback. For current temperatures, the impact is modest (<2%). In a warmer world, continuum-induced uncertainty in estimated feedback would be up to ~5%. Representing continuum absorption with the widely used semi-empirical model in radiative transfer calculations leads to an underestimation of this feedback. Constraining the shortwave continuum will help reduce these discrepancies.
Paolo Andreozzi, Mark D. Fielding, Robin J. Hogan, Richard M. Forbes, Samuel Rémy, Birger Bohn, and Ulrich Löhnert
EGUsphere, https://doi.org/10.5194/egusphere-2025-3790, https://doi.org/10.5194/egusphere-2025-3790, 2025
This preprint is open for discussion and under review for Atmospheric Chemistry and Physics (ACP).
Short summary
Short summary
Aerosols significantly contribute to the Earth’s climate, but models still struggle at representing them. Here we use satellite observations of clouds to improve aerosols in our weather and air-quality model. We show that African wildfires induce too bright simulated clouds and that our model removes too much aerosol from ice-containing clouds. This showcases how our approach effectively targets poorly observed aerosol processes, potentially informing weather forecasting and climate models.
Howard W. Barker, Jason N. S. Cole, Najda Villefranque, Zhipeng Qu, Almudena Velázquez Blázquez, Carlos Domenech, Shannon L. Mason, and Robin J. Hogan
Atmos. Meas. Tech., 18, 3095–3107, https://doi.org/10.5194/amt-18-3095-2025, https://doi.org/10.5194/amt-18-3095-2025, 2025
Short summary
Short summary
Measurements made by three instruments aboard EarthCARE are used to retrieve estimates of cloud and aerosol properties. A radiative closure assessment of these retrievals is performed by the ACMB-DF processor. Radiative transfer models acting on retrieved information produce broadband radiances commensurate with measurements made by EarthCARE’s broadband radiometer. Measured and modelled radiances for small domains are compared, and the likelihood of them differing by 10 W m2 defines the closure.
Jean-François Grailet, Robin J. Hogan, Nicolas Ghilain, David Bolsée, Xavier Fettweis, and Marilaure Grégoire
Geosci. Model Dev., 18, 1965–1988, https://doi.org/10.5194/gmd-18-1965-2025, https://doi.org/10.5194/gmd-18-1965-2025, 2025
Short summary
Short summary
The MAR (Modèle Régional Atmosphérique) is a regional climate model used for weather forecasting and studying the climate over various regions. This paper presents an update of MAR thanks to which it can precisely decompose solar radiation, in particular in the UV (ultraviolet) and photosynthesis ranges, both being critical to human health and ecosystems. As a first application of this new capability, this paper presents a method for predicting UV indices with MAR.
Robert Schoetter, Robin James Hogan, Cyril Caliot, and Valéry Masson
Geosci. Model Dev., 18, 405–431, https://doi.org/10.5194/gmd-18-405-2025, https://doi.org/10.5194/gmd-18-405-2025, 2025
Short summary
Short summary
Radiation is relevant to the atmospheric impact on people and infrastructure in cities as it can influence the urban heat island, building energy consumption, and human thermal comfort. A new urban radiation model, assuming a more realistic form of urban morphology, is coupled to the urban climate model Town Energy Balance (TEB). The new TEB is evaluated with a reference radiation model for a variety of urban morphologies, and an improvement in the simulated radiative observables is found.
Johannes Röttenbacher, André Ehrlich, Hanno Müller, Florian Ewald, Anna E. Luebke, Benjamin Kirbus, Robin J. Hogan, and Manfred Wendisch
Atmos. Chem. Phys., 24, 8085–8104, https://doi.org/10.5194/acp-24-8085-2024, https://doi.org/10.5194/acp-24-8085-2024, 2024
Short summary
Short summary
Weather prediction models simplify the physical processes related to light scattering by clouds consisting of complex ice crystals. Whether these simplifications are the cause for uncertainties in their prediction can be evaluated by comparing them with measurement data. Here we do this for Arctic ice clouds over sea ice using airborne measurements from two case studies. The model performs well for thick ice clouds but not so well for thin ones. This work can be used to improve the model.
Robin J. Hogan, Anthony J. Illingworth, Pavlos Kollias, Hajime Okamoto, and Ulla Wandinger
Atmos. Meas. Tech., 17, 3081–3083, https://doi.org/10.5194/amt-17-3081-2024, https://doi.org/10.5194/amt-17-3081-2024, 2024
Hanno Müller, André Ehrlich, Evelyn Jäkel, Johannes Röttenbacher, Benjamin Kirbus, Michael Schäfer, Robin J. Hogan, and Manfred Wendisch
Atmos. Chem. Phys., 24, 4157–4175, https://doi.org/10.5194/acp-24-4157-2024, https://doi.org/10.5194/acp-24-4157-2024, 2024
Short summary
Short summary
A weather model is used to compare solar radiation with measurements from an aircraft campaign in the Arctic. Model and observations agree on the downward radiation but show differences in the radiation reflected by the surface and the clouds, which in the model is too low above sea ice and too high above open ocean. The model–observation bias is reduced above open ocean by a realistic fraction of clouds and less cloud liquid water and above sea ice by less dark sea ice and more cloud droplets.
Shannon L. Mason, Howard W. Barker, Jason N. S. Cole, Nicole Docter, David P. Donovan, Robin J. Hogan, Anja Hünerbein, Pavlos Kollias, Bernat Puigdomènech Treserras, Zhipeng Qu, Ulla Wandinger, and Gerd-Jan van Zadelhoff
Atmos. Meas. Tech., 17, 875–898, https://doi.org/10.5194/amt-17-875-2024, https://doi.org/10.5194/amt-17-875-2024, 2024
Short summary
Short summary
When the EarthCARE mission enters its operational phase, many retrieval data products will be available, which will overlap both in terms of the measurements they use and the geophysical quantities they report. In this pre-launch study, we use simulated EarthCARE scenes to compare the coverage and performance of many data products from the European Space Agency production model, with the intention of better understanding the relation between products and providing a compact guide to users.
Megan A. Stretton, William Morrison, Robin J. Hogan, and Sue Grimmond
Geosci. Model Dev., 16, 5931–5947, https://doi.org/10.5194/gmd-16-5931-2023, https://doi.org/10.5194/gmd-16-5931-2023, 2023
Short summary
Short summary
Cities' materials and forms impact radiative fluxes. We evaluate the SPARTACUS-Urban multi-layer approach to modelling longwave radiation, describing realistic 3D geometry statistically using the explicit DART (Discrete Anisotropic Radiative Transfer) model. The temperature configurations used are derived from thermal camera observations. SPARTACUS-Urban accurately predicts longwave fluxes, with a low computational time (cf. DART), but has larger errors with sunlit/shaded surface temperatures.
Shannon L. Mason, Robin J. Hogan, Alessio Bozzo, and Nicola L. Pounder
Atmos. Meas. Tech., 16, 3459–3486, https://doi.org/10.5194/amt-16-3459-2023, https://doi.org/10.5194/amt-16-3459-2023, 2023
Short summary
Short summary
We present a method for accurately estimating the contents and properties of clouds, snow, rain, and aerosols through the atmosphere, using the combined measurements of the radar, lidar, and radiometer instruments aboard the upcoming EarthCARE satellite, and evaluate the performance of the retrieval, using test scenes simulated from a numerical forecast model. When EarthCARE is in operation, these quantities and their estimated uncertainties will be distributed in a data product called ACM-CAP.
Abdanour Irbah, Julien Delanoë, Gerd-Jan van Zadelhoff, David P. Donovan, Pavlos Kollias, Bernat Puigdomènech Treserras, Shannon Mason, Robin J. Hogan, and Aleksandra Tatarevic
Atmos. Meas. Tech., 16, 2795–2820, https://doi.org/10.5194/amt-16-2795-2023, https://doi.org/10.5194/amt-16-2795-2023, 2023
Short summary
Short summary
The Cloud Profiling Radar (CPR) and ATmospheric LIDar (ATLID) aboard the EarthCARE satellite are used to probe the Earth's atmosphere by measuring cloud and aerosol profiles. ATLID is sensitive to aerosols and small cloud particles and CPR to large ice particles, snowflakes and raindrops. It is the synergy of the measurements of these two instruments that allows a better classification of the atmospheric targets and the description of the associated products, which are the subject of this paper.
Beatriz M. Monge-Sanz, Alessio Bozzo, Nicholas Byrne, Martyn P. Chipperfield, Michail Diamantakis, Johannes Flemming, Lesley J. Gray, Robin J. Hogan, Luke Jones, Linus Magnusson, Inna Polichtchouk, Theodore G. Shepherd, Nils Wedi, and Antje Weisheimer
Atmos. Chem. Phys., 22, 4277–4302, https://doi.org/10.5194/acp-22-4277-2022, https://doi.org/10.5194/acp-22-4277-2022, 2022
Short summary
Short summary
The stratosphere is emerging as one of the keys to improve tropospheric weather and climate predictions. This study provides evidence of the role the stratospheric ozone layer plays in improving weather predictions at different timescales. Using a new ozone modelling approach suitable for high-resolution global models that provide operational forecasts from days to seasons, we find significant improvements in stratospheric meteorological fields and stratosphere–troposphere coupling.
David Meyer, Thomas Nagler, and Robin J. Hogan
Geosci. Model Dev., 14, 5205–5215, https://doi.org/10.5194/gmd-14-5205-2021, https://doi.org/10.5194/gmd-14-5205-2021, 2021
Short summary
Short summary
A major limitation in training machine-learning emulators is often caused by the lack of data. This paper presents a cheap way to increase the size of training datasets using statistical techniques and thereby improve the performance of machine-learning emulators.
Robin J. Hogan and Marco Matricardi
Geosci. Model Dev., 13, 6501–6521, https://doi.org/10.5194/gmd-13-6501-2020, https://doi.org/10.5194/gmd-13-6501-2020, 2020
Short summary
Short summary
A key component of computer models used to predict weather and climate is the radiation scheme, which calculates how solar and infrared radiation heats and cools the atmosphere and surface, including the important role of greenhouse gases. This paper describes the experimental protocol and large datasets for a new project, CKDMIP, to evaluate and improve the accuracy of the treatment of atmospheric gases in the radiation schemes used worldwide, as well as their computational speed.
Cited articles
Bradbury, J., Frostig, R., Hawkins, P., Johnson, M. J., Leary, C., Maclaurin,
D., Necula, G., Paszke, A., VanderPlas, J., Wanderman-Milne, S., and
Zhang, Q.: JAX: composable transformations of Python+NumPy programs, http://github.com/google/jax (last access: 8 June 2023), 2018. 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
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, 2020. a
Chevallier, F., Chéruy, F., Scott, N., and Chédin, A.: A neural network
approach for a fast and accurate computation of a longwave radiative budget,
J. Appl. Meteorol., 37, 1385–1397, 1998. a
Cotronei, A. and Slawig, T.: Single-precision arithmetic in ECHAM radiation reduces runtime and energy consumption, Geosci. Model Dev., 13, 2783–2804, https://doi.org/10.5194/gmd-13-2783-2020, 2020. a, b
Curcic, M.: A parallel Fortran framework for neural networks and deep learning, in: ACM SIGPLAN Fortran Forum, vol. 38, ACM New York, NY, USA, 4–21, https://doi.org/10.1145/3323057.3323059, 2019. a
Garand, L., Turner, D., Larocque, M., Bates, J., Boukabara, S., Brunel, P.,
Chevallier, F., Deblonde, G., Engelen, R., Hollingshead, M., et al.: Radiance
and Jacobian intercomparison of radiative transfer models applied to HIRS and
AMSU channels, J. Geophys. Res.-Atmos., 106,
24017–24031, 2001. 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
Goody, R., West, R., Chen, L., and Crisp, D.: The correlated-k method for
radiation calculations in nonhomogeneous atmospheres, J. Quant.
Spectrosc. Ra., 42, 539–550, 1989. a
Hogan, R. J. and Matricardi, M.: A Tool for Generating Fast k-Distribution
Gas-Optics Models for Weather and Climate Applications, J. Adv. Model. Earth Sy., 14, e2022MS003033,
https://doi.org/10.1029/2022MS003033, 2022. a, b, c, d
Hogan, R. J., Schäfer, S. A., Klinger, C., Chiu, J. C., and Mayer, B.:
Representing 3-D cloud radiation effects in two-stream schemes: 2. Matrix
formulation and broadband evaluation, J. Geophys. Res.-Atmos., 121, 8583–8599, 2016. a
Hogan, R. J., Ahlgrimm, M., Balsamo, G., Beljaars, A., Berrisford, P., Bozzo,
A., Di Giuseppe, F., Forbs, R. M., Haiden, T., Lang, S., Mayer, M.,
Polichtchouk, I., Sandu, I., Vitart, V., and Wedi, N.: Radiation in numerical
weather prediction, Tech. Memo. 816, ECMWF, https://www.ecmwf.int/en/elibrary/80347-radiation-numerical-weather-prediction (last access: 8 June 2023), 2017.
a
Iacono, M. J., Mlawer, E. J., Clough, S. A., and Morcrette, J.-J.: Impact of an
improved longwave radiation model, RRTM, on the energy budget and
thermodynamic properties of the NCAR community climate model, CCM3, J.
Geophys. Res.-Atmos., 105, 14873–14890, 2000. a
Inness, A., Ades, M., Agustí-Panareda, A., Barré, J., Benedictow, A., Blechschmidt, A.-M., Dominguez, J. J., Engelen, R., Eskes, H., Flemming, J., Huijnen, V., Jones, L., Kipling, Z., Massart, S., Parrington, M., Peuch, V.-H., Razinger, M., Remy, S., Schulz, M., and Suttie, M.: The CAMS reanalysis of atmospheric composition, Atmos. Chem. Phys., 19, 3515–3556, https://doi.org/10.5194/acp-19-3515-2019, 2019. a
Krasnopolsky, V., Fox-Rabinovitz, M., Hou, Y., Lord, S., and Belochitski, A.:
Accurate and fast neural network emulations of model radiation for the NCEP
coupled climate forecast system: climate simulations and seasonal
predictions, Mon. Weather Rev., 138, 1822–1842, 2010. a
Krasnopolsky, V. M., Fox-Rabinovitz, M. S., and Belochitski, A. A.: Decadal
climate simulations using accurate and fast neural network emulation of full,
longwave and shortwave, radiation, Mon. Weather Rev., 136, 3683–3695,
2008. a
Lagerquist, R., Turner, D., Ebert-Uphoff, I., Stewart, J., and Hagerty, V.:
Using Deep Learning to Emulate and Accelerate a Radiative Transfer Model,
J. Atmos. Ocean. Tech., 38, 1673–1696, 2021. a
Liu, Y., Caballero, R., and Monteiro, J. M.: RadNet 1.0: exploring deep learning architectures for longwave radiative transfer, Geosci. Model Dev., 13, 4399–4412, https://doi.org/10.5194/gmd-13-4399-2020, 2020. a, b
Meinshausen, M., Vogel, E., Nauels, A., Lorbacher, K., Meinshausen, N., Etheridge, D. M., Fraser, P. J., Montzka, S. A., Rayner, P. J., Trudinger, C. M., Krummel, P. B., Beyerle, U., Canadell, J. G., Daniel, J. S., Enting, I. G., Law, R. M., Lunder, C. R., O'Doherty, S., Prinn, R. G., Reimann, S., Rubino, M., Velders, G. J. M., Vollmer, M. K., Wang, R. H. J., and Weiss, R.: Historical greenhouse gas concentrations for climate modelling (CMIP6), Geosci. Model Dev., 10, 2057–2116, https://doi.org/10.5194/gmd-10-2057-2017, 2017. a
Mlawer, E. J., Taubman, S. J., Brown, P. D., Iacono, M. J., and Clough, S. A.:
Radiative transfer for inhomogeneous atmospheres: RRTM, a validated
correlated-k model for the longwave, J. Geophys. Res.-Atmos., 102, 16663–16682, 1997. a
Pal, A., Mahajan, S., and Norman, M. R.: Using Deep Neural Networks as
Cost-Effective Surrogate Models for Super-Parameterized E3SM Radiative
Transfer, Geophys. Res. Lett., 46, 6069–6079,
https://doi.org/10.1029/2018GL081646, 2019. a
Pincus, R., Forster, P. M., and Stevens, B.: The Radiative Forcing Model Intercomparison Project (RFMIP): experimental protocol for CMIP6, Geosci. Model Dev., 9, 3447–3460, https://doi.org/10.5194/gmd-9-3447-2016, 2016. a
Pincus, R., Mlawer, E. J., and Delamere, J. S.: Balancing accuracy, efficiency,
and flexibility in radiation calculations for dynamical models, J.
Adv. Model. Earth Sy., 11, 3074–3089,
https://doi.org/10.1029/2019MS001621, 2019. a, b, c, d
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
Roh, S. and Song, H.-J.: Evaluation of neural network emulations for radiation
parameterization in cloud resolving model, Geophys. Res. Lett., 47,
e2020GL089444, https://doi.org/10.1029/2020GL089444, 2020. a
Shonk, J. K. and Hogan, R. J.: Tripleclouds: An efficient method for
representing horizontal cloud inhomogeneity in 1D radiation schemes by using
three regions at each height, J. Climate, 21, 2352–2370, 2008. a
Song, H.-J. and Roh, S.: Improved weather forecasting using neural network
emulation for radiation parameterization, J. Adv. Model.
Earth Sy., 13, e2021MS002609,
https://doi.org/10.1029/2021MS002609, 2021. a, b, c
Ukkonen, P.: Exploring pathways to more accurate machine learning emulation of atmospheric radiative transfer, J. Adv. Model. Earth Sy., 14, e2021MS002875, https://doi.org/10.1029/2021MS002875, 2022a. a, b, c, d
Ukkonen, P.: Improving the trade-off between accuracy and efficiency of
atmospheric radiative transfer computations by using machine learning and
code optimization, PhD thesis, School of The Faculty of Science, University
of Copenhagen, ResearchGate, https://doi.org/10.13140/RG.2.2.27880.03, 2022b. a
Ukkonen, P.: peterukk/rte-rrtmgp-nn: 2.0, Zenodo [code], https://doi.org/10.5281/zenodo.7413935, 2022c. a
Ukkonen, P.: Code and extensive data for training neural networks for radiation, used in “Implementation of a machine-learned gas optics parameterization in the ECMWF Integrated Forecasting System: RRTMGP-NN 2.0”, Zenodo [code and data set], https://doi.org/10.5281/zenodo.6576680, 2022d.
a
Ukkonen, P.: Optimized version of the ecRad radiation scheme with new RRTMGP-NN gas optics, Zenodo [code], https://doi.org/10.5281/zenodo.7148329, 2022e. a
Ukkonen, P., Pincus, R., Hogan, R. J., Nielsen, K. P., and Kaas, E.:
Accelerating radiation computations for dynamical models with targeted
machine learning and code optimization, J. Adv. Model. Earth
Sy., 12, e2020MS002226, https://doi.org/10.1029/2020MS002226,
2020. a, b, c, d, e, f, g, h, i, j, k, l, m, n, o, p, q, r
Veerman, M. A., Pincus, R., Stoffer, R., Van Leeuwen, C. M., Podareanu, D., and Van Heerwaarden, C. C.: Predicting atmospheric optical properties for
radiative transfer computations using neural networks, Philos. T. R. Soc. A, 379, 20200095, https://doi.org/10.1098/rsta.2020.0095, 2021. a, b
Wang, X., Han, Y., Xue, W., Yang, G., and Zhang, G. J.: Stable climate simulations using a realistic general circulation model with neural network parameterizations for atmospheric moist physics and radiation processes, Geosci. Model Dev., 15, 3923–3940, https://doi.org/10.5194/gmd-15-3923-2022, 2022. 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
Short summary
Climate and weather models suffer from uncertainties resulting from approximated processes. Solar and thermal radiation is one example, as it is computationally too costly to simulate precisely. This has led to attempts to replace radiation codes based on physical equations with neural networks (NNs) that are faster but uncertain. In this paper we use global weather simulations to demonstrate that a middle-ground approach of using NNs only to predict optical properties is accurate and reliable.
Climate and weather models suffer from uncertainties resulting from approximated processes....