Articles | Volume 17, issue 6
https://doi.org/10.5194/gmd-17-2347-2024
© Author(s) 2024. 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-17-2347-2024
© Author(s) 2024. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Advances and prospects of deep learning for medium-range extreme weather forecasting
Leonardo Olivetti
CORRESPONDING AUTHOR
Department of Earth Sciences, Uppsala University, 75236 Uppsala, Sweden
Centre of Natural Hazards and Disaster Science, Uppsala University, 75236 Uppsala, Sweden
Gabriele Messori
Department of Earth Sciences, Uppsala University, 75236 Uppsala, Sweden
Centre of Natural Hazards and Disaster Science, Uppsala University, 75236 Uppsala, Sweden
Department of Meteorology, Stockholm University, 10691 Stockholm, Sweden
Bolin Centre for Climate Research, Stockholm University, 10691 Stockholm, Sweden
Related authors
Leonardo Olivetti and Gabriele Messori
Geosci. Model Dev., 17, 7915–7962, https://doi.org/10.5194/gmd-17-7915-2024, https://doi.org/10.5194/gmd-17-7915-2024, 2024
Short summary
Short summary
Data-driven models are becoming a viable alternative to physics-based models for weather forecasting up to 15 d into the future. However, it is unclear whether they are as reliable as physics-based models when forecasting weather extremes. We evaluate their performance in forecasting near-surface cold, hot, and windy extremes globally. We find that data-driven models can compete with physics-based models and that the choice of the best model mainly depends on the region and type of extreme.
Gabriele Messori, Emily Boyd, Joakim Nivre, and Elena Raffetti
EGUsphere, https://doi.org/10.5194/egusphere-2025-3451, https://doi.org/10.5194/egusphere-2025-3451, 2025
This preprint is open for discussion and under review for Earth System Dynamics (ESD).
Short summary
Short summary
Understanding impacts of climate extremes is very important for society and the economy. We identify three challenges restricting this understanding: limited availability and quality of impact data, difficulties in understanding why given impacts occur and lack of reliable projections of future impacts. We also identify key opportunities, including newly released datasets, recent methodological and technical advances and interdisciplinary collaborations between the social and natural sciences.
Valerio Lembo, Gabriele Messori, Davide Faranda, Vera Melinda Galfi, Rune Grand Graversen, and Flavio Emanuele Pons
EGUsphere, https://doi.org/10.5194/egusphere-2025-2189, https://doi.org/10.5194/egusphere-2025-2189, 2025
Short summary
Short summary
Hemispheric heatwaves have fundamental implications for ecosystems and societies. They are studied together with the large-scale atmospheric dynamics, through the lens of the poleward heat transports by planetary-scale waves. Extremely weak transports of heat towards the Poles are found to be associated with hemispheric heatwaves in the Northern Hemisphere mid-latitudes. Therefore, we conclude that heat transports are a clear indicator, and possibly a precursor of hemispehric heatwaves.
Michael K. Schutte, Alice Portal, Simon H. Lee, and Gabriele Messori
Weather Clim. Dynam., 6, 521–548, https://doi.org/10.5194/wcd-6-521-2025, https://doi.org/10.5194/wcd-6-521-2025, 2025
Short summary
Short summary
Large-scale motions in the atmosphere, namely atmospheric waves, greatly impact the weather that we experience at the Earth's surface. Here we investigate how waves in the troposphere (the lower 10 km of the atmosphere) and the stratosphere (above the troposphere) interact to affect surface weather. We find that tropospheric waves that are reflected back down by the stratosphere change weather patterns and temperatures in North America. These changes can indirectly affect the weather in Europe.
Sara Lindersson and Gabriele Messori
Earth Syst. Sci. Data Discuss., https://doi.org/10.5194/essd-2025-128, https://doi.org/10.5194/essd-2025-128, 2025
Preprint under review for ESSD
Short summary
Short summary
The study of past temperature-related disasters requires information on socioeconomic impacts, hazard intensity and human exposure. This is often lacking in current disaster databases. SHEDIS-Temperature fills this gap by integrating impact records with information on disaster locations, high-resolution meteorological data, and population estimates. Covering 382 disasters in 71 countries (1979–2018), this dataset enables deeper analyses of heat-related risk and vulnerabilities.
Ferran Lopez-Marti, Mireia Ginesta, Davide Faranda, Anna Rutgersson, Pascal Yiou, Lichuan Wu, and Gabriele Messori
Earth Syst. Dynam., 16, 169–187, https://doi.org/10.5194/esd-16-169-2025, https://doi.org/10.5194/esd-16-169-2025, 2025
Short summary
Short summary
Explosive cyclones and atmospheric rivers are two main drivers of extreme weather in Europe. In this study, we investigate their joint changes in future climates over the North Atlantic. Our results show that both the concurrence of these events and the intensity of atmospheric rivers increase by the end of the century across different future scenarios. Furthermore, explosive cyclones associated with atmospheric rivers last longer and are deeper than those without atmospheric rivers.
Leonardo Olivetti and Gabriele Messori
Geosci. Model Dev., 17, 7915–7962, https://doi.org/10.5194/gmd-17-7915-2024, https://doi.org/10.5194/gmd-17-7915-2024, 2024
Short summary
Short summary
Data-driven models are becoming a viable alternative to physics-based models for weather forecasting up to 15 d into the future. However, it is unclear whether they are as reliable as physics-based models when forecasting weather extremes. We evaluate their performance in forecasting near-surface cold, hot, and windy extremes globally. We find that data-driven models can compete with physics-based models and that the choice of the best model mainly depends on the region and type of extreme.
Clare Marie Flynn, Julia Moemken, Joaquim G. Pinto, and Gabriele Messori
Earth Syst. Sci. Data Discuss., https://doi.org/10.5194/essd-2024-298, https://doi.org/10.5194/essd-2024-298, 2024
Revised manuscript accepted for ESSD
Short summary
Short summary
We created a new, publicly available database of the Top 50 most extreme European winter windstorms from each of four different meteorological input data sets covering the years 1995–2015. We found variability in all aspects of our database, from which storms were included in the Top 50 storms for each input to their spatial variability. We urge users of our database to consider the storms as identified from two or more input sources within our database, where possible.
Gabriele Messori, Antonio Segalini, and Alexandre M. Ramos
Earth Syst. Dynam., 15, 1207–1225, https://doi.org/10.5194/esd-15-1207-2024, https://doi.org/10.5194/esd-15-1207-2024, 2024
Short summary
Short summary
Simultaneous heatwaves or cold spells in remote geographical regions have potentially far-reaching impacts on society and the environment. Despite this, we have little knowledge of when and where these extreme events have occurred in the past decades. In this paper, we present a summary of past simultaneous heatwaves or cold spells and provide a computer program to enable other researchers to study them.
Marco Gaetani, Gabriele Messori, Francesco S. R. Pausata, Shivangi Tiwari, M. Carmen Alvarez Castro, and Qiong Zhang
Clim. Past, 20, 1735–1759, https://doi.org/10.5194/cp-20-1735-2024, https://doi.org/10.5194/cp-20-1735-2024, 2024
Short summary
Short summary
Palaeoclimate reconstructions suggest that, around 6000 years ago, a greening of the Sahara took place, accompanied by climate changes in the Northern Hemisphere at middle to high latitudes. In this study, a climate model is used to investigate how this drastic environmental change in the Sahara impacted remote regions. Specifically, climate simulations reveal significant modifications in atmospheric circulation over the North Atlantic, affecting North American and European climates.
Antonio Segalini, Jacopo Riboldi, Volkmar Wirth, and Gabriele Messori
Weather Clim. Dynam., 5, 997–1012, https://doi.org/10.5194/wcd-5-997-2024, https://doi.org/10.5194/wcd-5-997-2024, 2024
Short summary
Short summary
Planetary Rossby waves are created by topography and evolve in time. In this work, an analytical solution of this classical problem is proposed under the approximation of linear wave dynamics. The theory is able to describe reasonably well the evolution of the perturbation and compares well with full nonlinear simulations. Several relevant cases with single and double zonal jets are assessed with the theoretical framework
Davide Faranda, Gabriele Messori, Erika Coppola, Tommaso Alberti, Mathieu Vrac, Flavio Pons, Pascal Yiou, Marion Saint Lu, Andreia N. S. Hisi, Patrick Brockmann, Stavros Dafis, Gianmarco Mengaldo, and Robert Vautard
Weather Clim. Dynam., 5, 959–983, https://doi.org/10.5194/wcd-5-959-2024, https://doi.org/10.5194/wcd-5-959-2024, 2024
Short summary
Short summary
We introduce ClimaMeter, a tool offering real-time insights into extreme-weather events. Our tool unveils how climate change and natural variability affect these events, affecting communities worldwide. Our research equips policymakers and the public with essential knowledge, fostering informed decisions and enhancing climate resilience. We analysed two distinct events, showcasing ClimaMeter's global relevance.
Aleksa Stanković, Gabriele Messori, Joaquim G. Pinto, and Rodrigo Caballero
Weather Clim. Dynam., 5, 821–837, https://doi.org/10.5194/wcd-5-821-2024, https://doi.org/10.5194/wcd-5-821-2024, 2024
Short summary
Short summary
The article studies extreme winds near the surface over the North Atlantic Ocean. These winds are caused by storms that pass through this region. The strongest storms that have occurred in the winters from 1950–2020 are studied in detail and compared to weaker but still strong storms. The analysis shows that the storms associated with the strongest winds are preceded by another older storm that travelled through the same region and made the conditions suitable for development of extreme winds.
Derrick Muheki, Axel A. J. Deijns, Emanuele Bevacqua, Gabriele Messori, Jakob Zscheischler, and Wim Thiery
Earth Syst. Dynam., 15, 429–466, https://doi.org/10.5194/esd-15-429-2024, https://doi.org/10.5194/esd-15-429-2024, 2024
Short summary
Short summary
Climate change affects the interaction, dependence, and joint occurrence of climate extremes. Here we investigate the joint occurrence of pairs of river floods, droughts, heatwaves, crop failures, wildfires, and tropical cyclones in East Africa under past and future climate conditions. Our results show that, across all future warming scenarios, the frequency and spatial extent of these co-occurring extremes will increase in this region, particularly in areas close to the Nile and Congo rivers.
Sina Loriani, Yevgeny Aksenov, David Armstrong McKay, Govindasamy Bala, Andreas Born, Cristiano M. Chiessi, Henk Dijkstra, Jonathan F. Donges, Sybren Drijfhout, Matthew H. England, Alexey V. Fedorov, Laura Jackson, Kai Kornhuber, Gabriele Messori, Francesco Pausata, Stefanie Rynders, Jean-Baptiste Salée, Bablu Sinha, Steven Sherwood, Didier Swingedouw, and Thejna Tharammal
EGUsphere, https://doi.org/10.5194/egusphere-2023-2589, https://doi.org/10.5194/egusphere-2023-2589, 2023
Short summary
Short summary
In this work, we draw on paleoreords, observations and modelling studies to review tipping points in the ocean overturning circulations, monsoon systems and global atmospheric circulations. We find indications for tipping in the ocean overturning circulations and the West African monsoon, with potentially severe impacts on the Earth system and humans. Tipping in the other considered systems is considered conceivable but currently not sufficiently supported by evidence.
Emma Holmberg, Gabriele Messori, Rodrigo Caballero, and Davide Faranda
Earth Syst. Dynam., 14, 737–765, https://doi.org/10.5194/esd-14-737-2023, https://doi.org/10.5194/esd-14-737-2023, 2023
Short summary
Short summary
We analyse the duration of large-scale patterns of air movement in the atmosphere, referred to as persistence, and whether unusually persistent patterns favour warm-temperature extremes in Europe. We see no clear relationship between summertime heatwaves and unusually persistent patterns. This suggests that heatwaves do not necessarily require the continued flow of warm air over a region and that local effects could be important for their occurrence.
Axel Kleidon, Gabriele Messori, Somnath Baidya Roy, Ira Didenkulova, and Ning Zeng
Earth Syst. Dynam., 14, 241–242, https://doi.org/10.5194/esd-14-241-2023, https://doi.org/10.5194/esd-14-241-2023, 2023
Patrick Johannes Stoll, Rune Grand Graversen, and Gabriele Messori
Weather Clim. Dynam., 4, 1–17, https://doi.org/10.5194/wcd-4-1-2023, https://doi.org/10.5194/wcd-4-1-2023, 2023
Short summary
Short summary
The atmosphere is in motion and hereby transporting warm, cold, moist, and dry air to different climate zones. In this study, we investigate how this transport of energy organises in different manners. Outside the tropics, atmospheric waves of sizes between 2000 and 8000 km, which we perceive as cyclones from the surface, transport most of the energy and moisture poleward. In the winter, large-scale weather situations become very important for transporting energy into the polar regions.
Davide Faranda, Stella Bourdin, Mireia Ginesta, Meriem Krouma, Robin Noyelle, Flavio Pons, Pascal Yiou, and Gabriele Messori
Weather Clim. Dynam., 3, 1311–1340, https://doi.org/10.5194/wcd-3-1311-2022, https://doi.org/10.5194/wcd-3-1311-2022, 2022
Short summary
Short summary
We analyze the atmospheric circulation leading to impactful extreme events for the calendar year 2021 such as the Storm Filomena, Westphalia floods, Hurricane Ida and Medicane Apollo. For some of the events, we find that climate change has contributed to their occurrence or enhanced their intensity; for other events, we find that they are unprecedented. Our approach underscores the importance of considering changes in the atmospheric circulation when performing attribution studies.
Gabriele Messori, Marlene Kretschmer, Simon H. Lee, and Vivien Wendt
Weather Clim. Dynam., 3, 1215–1236, https://doi.org/10.5194/wcd-3-1215-2022, https://doi.org/10.5194/wcd-3-1215-2022, 2022
Short summary
Short summary
Over 10 km above the ground, there is a region of the atmosphere called the stratosphere. While there is very little air in the stratosphere itself, its interactions with the lower parts of the atmosphere – where we live – can affect the weather. Here we study a specific example of such an interaction, whereby processes occurring at the boundary of the stratosphere can lead to a continent-wide drop in temperatures in North America during winter.
Valerio Lembo, Federico Fabiano, Vera Melinda Galfi, Rune Grand Graversen, Valerio Lucarini, and Gabriele Messori
Weather Clim. Dynam., 3, 1037–1062, https://doi.org/10.5194/wcd-3-1037-2022, https://doi.org/10.5194/wcd-3-1037-2022, 2022
Short summary
Short summary
Eddies in mid-latitudes characterize the exchange of heat between the tropics and the poles. This exchange is largely uneven, with a few extreme events bearing most of the heat transported across latitudes in a season. It is thus important to understand what the dynamical mechanisms are behind these events. Here, we identify recurrent weather regime patterns associated with extreme transports, and we identify scales of mid-latitudinal eddies that are mostly responsible for the transport.
Assaf Hochman, Francesco Marra, Gabriele Messori, Joaquim G. Pinto, Shira Raveh-Rubin, Yizhak Yosef, and Georgios Zittis
Earth Syst. Dynam., 13, 749–777, https://doi.org/10.5194/esd-13-749-2022, https://doi.org/10.5194/esd-13-749-2022, 2022
Short summary
Short summary
Gaining a complete understanding of extreme weather, from its physical drivers to its impacts on society, is important in supporting future risk reduction and adaptation measures. Here, we provide a review of the available scientific literature, knowledge gaps and key open questions in the study of extreme weather events over the vulnerable eastern Mediterranean region.
Francesco S. R. Pausata, Gabriele Messori, Jayoung Yun, Chetankumar A. Jalihal, Massimo A. Bollasina, and Thomas M. Marchitto
Clim. Past, 17, 1243–1271, https://doi.org/10.5194/cp-17-1243-2021, https://doi.org/10.5194/cp-17-1243-2021, 2021
Short summary
Short summary
Far-afield changes in vegetation such as those that occurred over the Sahara during the middle Holocene and the consequent changes in dust emissions can affect the intensity of the South Asian Monsoon (SAM) rainfall and the lengthening of the monsoon season. This remote influence is mediated by anomalies in Indian Ocean sea surface temperatures and may have shaped the evolution of the SAM during the termination of the African Humid Period.
Gabriele Messori and Davide Faranda
Clim. Past, 17, 545–563, https://doi.org/10.5194/cp-17-545-2021, https://doi.org/10.5194/cp-17-545-2021, 2021
Short summary
Short summary
The palaeoclimate community must both analyse large amounts of model data and compare very different climates. Here, we present a seemingly very abstract analysis approach that may be fruitfully applied to palaeoclimate numerical simulations. This approach characterises the dynamics of a given climate through a small number of metrics and is thus suited to face the above challenges.
Gabriele Messori, Nili Harnik, Erica Madonna, Orli Lachmy, and Davide Faranda
Earth Syst. Dynam., 12, 233–251, https://doi.org/10.5194/esd-12-233-2021, https://doi.org/10.5194/esd-12-233-2021, 2021
Short summary
Short summary
Atmospheric jets are a key component of the climate system and of our everyday lives. Indeed, they affect human activities by influencing the weather in many mid-latitude regions. However, we still lack a complete understanding of their dynamical properties. In this study, we try to relate the understanding gained in idealized computer simulations of the jets to our knowledge from observations of the real atmosphere.
Assaf Hochman, Sebastian Scher, Julian Quinting, Joaquim G. Pinto, and Gabriele Messori
Earth Syst. Dynam., 12, 133–149, https://doi.org/10.5194/esd-12-133-2021, https://doi.org/10.5194/esd-12-133-2021, 2021
Short summary
Short summary
Skillful forecasts of extreme weather events have a major socioeconomic relevance. Here, we compare two approaches to diagnose the predictability of eastern Mediterranean heat waves: one based on recent developments in dynamical systems theory and one leveraging numerical ensemble weather forecasts. We conclude that the former can be a useful and cost-efficient complement to conventional numerical forecasts for understanding the dynamics of eastern Mediterranean heat waves.
Paolo De Luca, Gabriele Messori, Davide Faranda, Philip J. Ward, and Dim Coumou
Earth Syst. Dynam., 11, 793–805, https://doi.org/10.5194/esd-11-793-2020, https://doi.org/10.5194/esd-11-793-2020, 2020
Short summary
Short summary
In this paper we quantify Mediterranean compound temperature and precipitation dynamical extremes (CDEs) over the 1979–2018 period. The strength of the temperature–precipitation coupling during summer increased and is driven by surface warming. We also link the CDEs to compound hot–dry and cold–wet events during summer and winter respectively.
Cited articles
Bahdanau, D., Cho, K., and Bengio, Y.: Neural Machine Translation by Jointly Learning to Align and Translate, in: 3rd International Conference on Learning Representations, ICLR 2015, San Diego, CA, USA, 7–9 May 2015, Conference Track Proceedings, edited by: Bengio, Y. and LeCun, Y., https://doi.org/10.48550/arXiv.1409.0473, 2015. a
Balkema, A. A. and De Haan, L.: Residual Life Time at Great Age, Ann. Probab., 2, 792–804, https://doi.org/10.1214/aop/1176996548, 1974. a, b
Barnes, A. P., McCullen, N., and Kjeldsen, T. R.: Forecasting seasonal to sub-seasonal rainfall in Great Britain using convolutional-neural networks, Theor. Appl. Climatol., 151, 421–432, https://doi.org/10.1007/s00704-022-04242-x, 2023. a
Battaglia, P. W., Hamrick, J. B., Bapst, V., Sanchez-Gonzalez, A., Zambaldi, V., Malinowski, M., Tacchetti, A., Raposo, D., Santoro, A., Faulkner, R., Gulcehre, C., Song, F., Ballard, A., Gilmer, J., Dahl, G., Vaswani, A., Allen, K., Nash, C., Langston, V., Dyer, C., Heess, N., Wierstra, D., Kohli, P., Botvinick, M., Vinyals, O., Li, Y., and Pascanu, R.: Relational inductive biases, deep learning, and graph networks, arXiv, https://doi.org/10.48550/arXiv.1806.01261, 2018. a, b, c
Bauer, P., Dueben, P., Chantry, M., Doblas-Reyes, F., Hoefler, T., McGovern, A., and Stevens, B.: Deep learning and a changing economy in weather and climate prediction, Nat. Rev. Earth Environ., 4, 507–509, https://doi.org/10.1038/s43017-023-00468-z, 2023. a
Ben-Bouallegue, Z., Clare, M. C. A., Magnusson, L., Gascon, E., Maier-Gerber, M., Janousek, M., Rodwell, M., Pinault, F., Dramsch, J. S., Lang, S. T. K., Raoult, B., Rabier, F., Chevallier, M., Sandu, I., Dueben, P., Chantry, M., and Pappenberger, F.: The rise of data-driven weather forecasting, arXiv, https://doi.org/10.48550/arXiv.2307.10128, 2023. a, b
Bengio, Y. and Gingras, F.: Recurrent Neural Networks for Missing or Asynchronous Data, in: Advances in Neural Information Processing Systems, vol. 8, MIT Press, https://papers.nips.cc/paper_files/paper/1995/hash/ffeed84c7cb1ae7bf4ec4bd78275bb98-Abstract.html (last access: 18 March 2024), 1995. a
Bengio, Y., Simard, P., and Frasconi, P.: Learning long-term dependencies with gradient descent is difficult, IEEE transactions on neural networks / a publication of the IEEE Neural Networks Council, 5, 157–66, https://doi.org/10.1109/72.279181, 1994. a
Beucler, T., Pritchard, M., Gentine, P., and Rasp, S.: Towards Physically-Consistent, Data-Driven Models of Convection, in: IGARSS 2020 – 2020 IEEE International Geoscience and Remote Sensing Symposium, 3987–3990, https://doi.org/10.1109/IGARSS39084.2020.9324569, 2020. a
Bi, K., Xie, L., Zhang, H., Chen, X., Gu, X., and Tian, Q.: Pangu-Weather: A 3D High-Resolution Model for Fast and Accurate Global Weather Forecast, arXiv, https://doi.org/10.48550/arXiv.2211.02556, 2022. a, b
Boomsma, W. and Frellsen, J.: Spherical convolutions and their application in molecular modelling, in: Advances in Neural Information Processing Systems, 30, Curran Associates, Inc., https://papers.nips.cc/paper_files/paper/2017/hash/1113d7a76ffceca1bb350bfe145467c6-Abstract.html (last access: 18 March 2024), 2017. a
Carreau, J. and Bengio, Y.: A Hybrid Pareto Model for Conditional Density Estimation of Asymmetric Fat-Tail Data, in: Proceedings of the Eleventh International Conference on Artificial Intelligence and Statistics, 51–58, PMLR, https://proceedings.mlr.press/v2/carreau07a.html (last access: 18 March 2024), 2007. a
Chkeir, S., Anesiadou, A., Mascitelli, A., and Biondi, R.: Nowcasting extreme rain and extreme wind speed with machine learning techniques applied to different input datasets, Atmos. Res., 282, 106548, https://doi.org/10.1016/j.atmosres.2022.106548, 2023. a
Cho, K., van Merriënboer, B., Bahdanau, D., and Bengio, Y.: On the Properties of Neural Machine Translation: Encoder–Decoder Approaches, in: Proceedings of SSST-8, Eighth Workshop on Syntax, Semantics and Structure in Statistical Translation, 103–111, Association for Computational Linguistics, Doha, Qatar, https://doi.org/10.3115/v1/W14-4012, 2014. a
Cisneros, D., Richards, J., Dahal, A., Lombardo, L., and Huser, R.: Deep graphical regression for jointly moderate and extreme Australian wildfires, arXiv, https://doi.org/10.48550/arXiv.2308.14547, 2023. a
Civitarese, D. S., Szwarcman, D., Zadrozny, B., and Watson, C.: Extreme Precipitation Seasonal Forecast Using a Transformer Neural Network, arXiv, https://doi.org/10.48550/arXiv.2107.06846, 2021. a
Clare, M. C., Jamil, O., and Morcrette, C. J.: Combining distribution-based neural networks to predict weather forecast probabilities, Q. J. Roy. Meteor. Soc., 147, 4337–4357, https://doi.org/10.1002/qj.4180, 2021. a, b
Cohen, T. S., Geiger, M., Koehler, J., and Welling, M.: Spherical CNNs, arXiv, https://doi.org/10.48550/arXiv.1801.10130, 2018. a
de Burgh-Day, C. O. and Leeuwenburg, T.: Machine learning for numerical weather and climate modelling: a review, Geosci. Model Dev., 16, 6433–6477, https://doi.org/10.5194/gmd-16-6433-2023, 2023. a
Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., Uszkoreit, J., and Houlsby, N.: An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale, in: International Conference on Learning Representations, https://openreview.net/forum?id=YicbFdNTTy (last access: 18 March 2024), 2020. a, b
Dueben, P. D. and Bauer, P.: Challenges and design choices for global weather and climate models based on machine learning, Geosci. Model Dev., 11, 3999–4009, https://doi.org/10.5194/gmd-11-3999-2018, 2018. a
ECMWF: Machine Learning model data, https://www.ecmwf.int/en/forecasts/dataset/machine-learning-model-data (last access: 18 March 2024), 2023. a
Espeholt, L., Agrawal, S., Sønderby, C., Kumar, M., Heek, J., Bromberg, C., Gazen, C., Hickey, J., Bell, A., and Kalchbrenner, N.: Skillful Twelve Hour Precipitation Forecasts using Large Context Neural Networks, arxiv, https://doi.org/10.48550/arXiv.2111.07470, 2021. a
Grönquist, P., Yao, C., Ben-Nun, T., Dryden, N., Dueben, P., Li, S., and Hoefler, T.: Deep learning for post-processing ensemble weather forecasts, Philos. T. Roy. Soc. A, 379, 20200092, https://doi.org/10.1098/rsta.2020.0092, 2021. a
Guastavino, S., Piana, M., Tizzi, M., Cassola, F., Iengo, A., Sacchetti, D., Solazzo, E., and Benvenuto, F.: Prediction of severe thunderstorm events with ensemble deep learning and radar data, Sci. Rep., 12, 20049, https://doi.org/10.1038/s41598-022-23306-6, 2022. a
Gutzwiller, K. J. and Serno, K. M.: Using the risk of spatial extrapolation by machine-learning models to assess the reliability of model predictions for conservation, Landscape Ecol., 38, 1363–1372, https://doi.org/10.1007/s10980-023-01651-9, 2023. a
Haidar, A. and Verma, B.: Monthly Rainfall Forecasting Using One-Dimensional Deep Convolutional Neural Network, IEEE Access, 6, 69053–69063, https://doi.org/10.1109/ACCESS.2018.2880044, 2018. a, b
Hall, T., Brooks, H. E., and Doswell, C. A.: Precipitation Forecasting Using a Neural Network, Weather Forecast., 14, 338–345, https://doi.org/10.1175/1520-0434(1999)014<0338:PFUANN>2.0.CO;2, 1999. a, b
Hersbach, H., Bell, B., Berrisford, P., Hirahara, S., Horányi, A., Muñoz-Sabater, J., Nicolas, J., Peubey, C., Radu, R., Schepers, D., Simmons, A., Soci, C., Abdalla, S., Abellan, X., Balsamo, G., Bechtold, P., Biavati, G., Bidlot, J., Bonavita, M., De Chiara, G., Dahlgren, P., Dee, D., Diamantakis, M., Dragani, R., Flemming, J., Forbes, R., Fuentes, M., Geer, A., Haimberger, L., Healy, S., Hogan, R. J., Hólm, E., Janisková, M., Keeley, S., Laloyaux, P., Lopez, P., Lupu, C., Radnoti, G., de Rosnay, P., Rozum, I., Vamborg, F., Villaume, S., and Thépaut, J.-N.: The ERA5 global reanalysis, Q. J. Roy. Meteor. Soc., 146, 1999–2049, https://doi.org/10.1002/qj.3803, 2020. a, b
Hersbach, H., Bell, B., Berrisford, P., Biavati, G., Horányi, A., Muñoz Sabater, J., Nicolas, J., Peubey, C., Radu, R., Rozum, I., Schepers, D., Simmons, A., Soci, C., Dee, D., and Thépaut, J.-N.: ERA5 hourly data on single levels from 1940 to present, Copernicus Climate Change Service (C3S) Climate Data Store (CDS) [data set], https://doi.org/10.24381/cds.adbb2d47, 2023a. a
Hersbach, H., Bell, B., Berrisford, P., Biavati, G., Horányi, A., Muñoz Sabater, J., Nicolas, J., Peubey, C., Radu, R., Rozum, I., Schepers, D., Simmons, A., Soci, C., Dee, D., and Thépaut, J.-N.: ERA5 hourly data on pressure levels from 1940 to present, Copernicus Climate Change Service (C3S) Climate Data Store (CDS) [data set], https://doi.org/10.24381/cds.bd0915c6, 2023b. a
Hochreiter, S. and Schmidhuber, J.: Long Short-term Memory, Neural Comput., 9, 1735–80, https://doi.org/10.1162/neco.1997.9.8.1735, 1997. a
Hodson, T. O.: Root-mean-square error (RMSE) or mean absolute error (MAE): when to use them or not, Geosci. Model Dev., 15, 5481–5487, https://doi.org/10.5194/gmd-15-5481-2022, 2022. a
Hu, Y., Chen, L., Wang, Z., and Li, H.: SwinVRNN: A Data-Driven Ensemble Forecasting Model via Learned Distribution Perturbation, J. Adv. Model. Earth Sy., 15, e2022MS003211, https://doi.org/10.1029/2022MS003211, 2023. a
Ivakhnenko, A. G. and Lapa, V. G.: Cybernetic Predicting Devices, Joint Publications Research Service, available from the Clearinghouse for Federal Scientific and Technical Information, 1965. a
Kashinath, K., Mustafa, M., Albert, A., Wu, J.-L., Jiang, C., Esmaeilzadeh, S., Azizzadenesheli, K., Wang, R., Chattopadhyay, A., Singh, A., Manepalli, A., Chirila, D., Yu, R., Walters, R., White, B., Xiao, H., Tchelepi, H. A., Marcus, P., Anandkumar, A., Hassanzadeh, P., and Prabhat, n.: Physics-informed machine learning: case studies for weather and climate modelling, Philos. T. Roy. Soc. A, 379, 20200093, https://doi.org/10.1098/rsta.2020.0093, 2021. a, b
Klein, B., Wolf, L., and Afek, Y.: A Dynamic Convolutional Layer for short rangeweather prediction, in: 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 4840–4848, https://doi.org/10.1109/CVPR.2015.7299117, 2015. a, b
Koenker, R. and Bassett, G.: Regression Quantiles, Econometrica, 46, 33–50, https://doi.org/10.2307/1913643, 1978. a
Kramer, M. A.: Nonlinear principal component analysis using autoassociative neural networks, AIChE Journal, 37, 233–243, https://doi.org/10.1002/aic.690370209, 1991. a
Kron, W., Löw, P., and Kundzewicz, Z. W.: Changes in risk of extreme weather events in Europe, Environ. Sci. Policy, 100, 74–83, https://doi.org/10.1016/j.envsci.2019.06.007, 2019. a
Lam, R., Sanchez-Gonzalez, A., Willson, M., Wirnsberger, P., Fortunato, M., Pritzel, A., Ravuri, S., Ewalds, T., Alet, F., Eaton-Rosen, Z., Hu, W., Merose, A., Hoyer, S., Holland, G., Stott, J., Vinyals, O., Mohamed, S., and Battaglia, P.: GraphCast: Learning skillful medium-range global weather forecasting, arXiv, https://doi.org/10.48550/arXiv.2212.12794, 2022. a
Lam, R., Sanchez-Gonzalez, A., Willson, M., Wirnsberger, P., Fortunato, M., Alet, F., Ravuri, S., Ewalds, T., Eaton-Rosen, Z., Hu, W., Merose, A., Hoyer, S., Holland, G., Vinyals, O., Stott, J., Pritzel, A., Mohamed, S., and Battaglia, P.: Learning skillful medium-range global weather forecasting, Science, 382, 1416–1421, https://doi.org/10.1126/science.adi2336, 2023. a, b, c, d, e, f, g, h, i
LeCun, Y. and Bengio, Y.: Convolutional networks for images, speech, and time series, The handbook of brain theory and neural networks, Citeseer, 3361, 1995. a
LeNail, A.: NN-SVG: Publication-Ready Neural Network Architecture Schematics, J. Open Source Softw., 4, 747, https://doi.org/10.21105/joss.00747, 2019. a
Li, X., Du, Z., and Song, G.: A Method of Rainfall Runoff Forecasting Based on Deep Convolution Neural Networks, in: 2018 Sixth International Conference on Advanced Cloud and Big Data (CBD), Lanzhou, China, 304–310, https://doi.org/10.1109/CBD.2018.00061, 2018. a, b
Merz, B., Kuhlicke, C., Kunz, M., Pittore, M., Babeyko, A., Bresch, D. N., Domeisen, D. I. V., Feser, F., Koszalka, I., Kreibich, H., Pantillon, F., Parolai, S., Pinto, J. G., Punge, H. J., Rivalta, E., Schröter, K., Strehlow, K., Weisse, R., and Wurpts, A.: Impact Forecasting to Support Emergency Management of Natural Hazards, Rev. Geophys., 58, RG000704, https://doi.org/10.1029/2020RG000704, 2020. a, b, c
Molina, M. J., O'Brien, T. A., Anderson, G., Ashfaq, M., Bennett, K. E., Collins, W. D., Dagon, K., Restrepo, J. M., and Ullrich, P. A.: A Review of Recent and Emerging Machine Learning Applications for Climate Variability and Weather Phenomena, Artificial Intelligence for the Earth Systems, 2, e220086, https://doi.org/10.1175/AIES-D-22-0086.1, 2023. a
Nguyen, T., Brandstetter, J., Kapoor, A., Gupta, J. K., and Grover, A.: ClimaX: A foundation model for weather and climate, arXiv, https://doi.org/10.48550/arXiv.2301.10343, 2023. a, b, c, d
Pasche, O. C. and Engelke, S.: Neural Networks for Extreme Quantile Regression with an Application to Forecasting of Flood Risk, arXiv, https://doi.org/10.48550/arXiv.2208.07590, 2023. a, b
Pathak, J., Subramanian, S., Harrington, P., Raja, S., Chattopadhyay, A., Mardani, M., Kurth, T., Hall, D., Li, Z., Azizzadenesheli, K., Hassanzadeh, P., Kashinath, K., and Anandkumar, A.: FourCastNet: A Global Data-driven High-resolution Weather Model using Adaptive Fourier Neural Operators, arXiv, https://doi.org/10.48550/arXiv.2202.11214, 2022. a, b, c, d, e, f, g
Qiu, M., Zhao, P., Zhang, K., Huang, J., Shi, X., Wang, X., and Chu, W.: A Short-Term Rainfall Prediction Model Using Multi-task Convolutional Neural Networks, in: 2017 IEEE International Conference on Data Mining (ICDM), New Orleans, LA, USA, 395–404, https://doi.org/10.1109/ICDM.2017.49, 2017. a, b
Rasp, S., Dueben, P. D., Scher, S., Weyn, J. A., Mouatadid, S., and Thuerey, N.: WeatherBench: A Benchmark Data Set for Data-Driven Weather Forecasting, J. Adv. Model. Earth Sy., 12, e2020MS002203, https://doi.org/10.1029/2020MS002203, 2020. a
Rasp, S., Hoyer, S., Merose, A., Langmore, I., Battaglia, P., Russel, T., Sanchez-Gonzalez, A., Yang, V., Carver, R., Agrawal, S., Chantry, M., Bouallegue, Z. B., Dueben, P., Bromberg, C., Sisk, J., Barrington, L., Bell, A., and Sha, F.: WeatherBench 2: A benchmark for the next generation of data-driven global weather models, arXiv, https://doi.org/10.48550/arXiv.2308.15560, 2024. a
Ravuri, S., Lenc, K., Willson, M., Kangin, D., Lam, R., Mirowski, P., Fitzsimons, M., Athanassiadou, M., Kashem, S., Madge, S., Prudden, R., Mandhane, A., Clark, A., Brock, A., Simonyan, K., Hadsell, R., Robinson, N., Clancy, E., Arribas, A., and Mohamed, S.: Skilful precipitation nowcasting using deep generative models of radar, Nature, 597, 672–677, https://doi.org/10.1038/s41586-021-03854-z, 2021. a
Ren, X., Li, X., Ren, K., Song, J., Xu, Z., Deng, K., and Wang, X.: Deep Learning-Based Weather Prediction: A Survey, Big Data Research, 23, 100178, https://doi.org/10.1016/j.bdr.2020.100178, 2021. a
Rumelhart, D. E., Hinton, G. E., and Williams, R. J.: Learning representations by back-propagating errors, Nature, 323, 533–536, https://doi.org/10.1038/323533a0, 1986. a
Scarselli, F., Gori, M., Tsoi, A. C., Hagenbuchner, M., and Monfardini, G.: The Graph Neural Network Model, IEEE T. Neural Networ., 20, 61–80, https://doi.org/10.1109/TNN.2008.2005605, 2009. a
Scher, S. and Messori, G.: Generalization properties of feed-forward neural networks trained on Lorenz systems, Nonlin. Processes Geophys., 26, 381–399, https://doi.org/10.5194/npg-26-381-2019, 2019a. a
Scher, S. and Messori, G.: Weather and climate forecasting with neural networks: using general circulation models (GCMs) with different complexity as a study ground, Geosci. Model Dev., 12, 2797–2809, https://doi.org/10.5194/gmd-12-2797-2019, 2019b. a
Scher, S. and Messori, G.: Ensemble Methods for Neural Network-Based Weather Forecasts, J. Adv. Model. Earth Sy., 13, MS002331, https://doi.org/10.1029/2020MS002331, 2021. a
Scher, S. and Messori, G.: Spherical convolution and other forms of informed machine learning for deep neural network based weather forecasts, arXiv, https://doi.org/10.48550/arXiv.2008.13524, 2023. a
Schultz, M. G., Betancourt, C., Gong, B., Kleinert, F., Langguth, M., Leufen, L. H., Mozaffari, A., and Stadtler, S.: Can deep learning beat numerical weather prediction?, Philos. T. Roy. Soc. A, 379, 20200097, https://doi.org/10.1098/rsta.2020.0097, 2021. a, b, c
Silini, R., Lerch, S., Mastrantonas, N., Kantz, H., Barreiro, M., and Masoller, C.: Improving the prediction of the Madden–Julian Oscillation of the ECMWF model by post-processing, Earth Syst. Dynam., 13, 1157–1165, https://doi.org/10.5194/esd-13-1157-2022, 2022. a
Taylor, J. W.: A quantile regression neural network approach to estimating the conditional density of multiperiod returns, J. Forecast., 19, 299–311, https://doi.org/10.1002/1099-131X(200007)19:4<299::AID-FOR775>3.0.CO;2-V, 2000. a
Thuemmel, J., Karlbauer, M., Otte, S., Zarfl, C., Martius, G., Ludwig, N., Scholten, T., Friedrich, U., Wulfmeyer, V., Goswami, B., and Butz, M. V.: Inductive biases in deep learning models for weather prediction, arXiv, https://doi.org/10.48550/arXiv.2304.04664, 2023. a
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, Å., and Polosukhin, I.: Attention is All you Need, in: Advances in Neural Information Processing Systems, edited by Guyon, I., Luxburg, U. V., Bengio, S., Wallach, H., Fergus, R., Vishwanathan, S., and Garnett, R., vol. 30, Curran Associates, Inc., https://proceedings.neurips.cc/paper_files/paper/2017/file/3f5ee243547dee91fbd053c1c4a845aa-Paper.pdf (last access: 18 March 2024), 2017. a
Weyn, J. A., Durran, D. R., and Caruana, R.: Can Machines Learn to Predict Weather? Using Deep Learning to Predict Gridded 500-hPa Geopotential Height From Historical Weather Data, J. Adv. Model. Earth Sy., 11, 2680–2693, https://doi.org/10.1029/2019MS001705, 2019. a
Zhang, Y., Long, M., Chen, K., Xing, L., Jin, R., Jordan, M. I., and Wang, J.: Skilful nowcasting of extreme precipitation with NowcastNet, Nature, 619, 526–532, https://doi.org/10.1038/s41586-023-06184-4, 2023. a
Zhong, X., Chen, L., Liu, J., Lin, C., Qi, Y., and Li, H.: FuXi-Extreme: Improving extreme rainfall and wind forecasts with diffusion model, arXiv, https://doi.org/10.48550/arXiv.2310.19822, 2023. a
Zhu, A., Li, X., Mo, Z., and Wu, R.: Wind power prediction based on a convolutional neural network, in: 2017 International Conference on Circuits, Devices and Systems (ICCDS), Chengdu, China, 131–135, https://doi.org/10.1109/ICCDS.2017.8120465, 2017. a, b
Executive editor
This article provides a concise and well-written review of the current state of numerical weather prediction using machine learning models. Given how quickly this field is evolving, it's difficult for the traditional peer review process to capture all developments in this space, but this manuscript provides an excellent snapshot of the current state of the art.
This article provides a concise and well-written review of the current state of numerical...
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.
In the last decades, weather forecasting up to 15 d into the future has been dominated by...