Weather forecasting: Difference between revisions
Content added Content deleted
Aaron.spring (talk | contribs) (→Relevant Groups and Organizations: add s2s-ai-challenge and context) |
No edit summary |
||
| Line 1: | Line 1: | ||
''This page is about how machine learning (ML) can be used in the context of weather forecasting. For an overview of weather forecasting more generally, please see the [https://en.wikipedia.org/wiki/Weather_forecasting Wikipedia page] on this topic.'' |
|||
{{Stub}} |
|||
{{Disclaimer}} |
|||
ML is a suitable tool for making short-term weather predictions<ref>{{Cite web|url=https://www.ecmwf.int/en/elibrary/19877-machine-learning-ecmwf-roadmap-next-10-years|title=Machine learning at ECMWF: A roadmap for the next 10 years|last=Düben|first=Peter|last2=Modigliani|first2=Umberto|date=2021|website=www.ecmwf.int|access-date=2021-01-25|last3=Geer|first3=Alan|last4=Siemen|first4=Stephan|last5=Pappenberger|first5=Florian|last6=Bauer|first6=Peter|last7=Brown|first7=Andy|last8=Palkovic|first8=Martin|last9=Raoult|first9=Baudouin}}</ref> based on the observed initial conditions, and post-processing the output from weather models<ref>{{Cite journal|last=Rasp|first=Stephan|last2=Lerch|first2=Sebastian|date=2018-11-01|title=Neural Networks for Postprocessing Ensemble Weather Forecasts|url=https://journals.ametsoc.org/view/journals/mwre/146/11/mwr-d-18-0187.1.xml|journal=Monthly Weather Review|language=EN|volume=146|issue=11|pages=3885–3900|doi=10.1175/MWR-D-18-0187.1|issn=1520-0493}}</ref>. |
ML is a suitable tool for making short-term weather predictions<ref>{{Cite web|url=https://www.ecmwf.int/en/elibrary/19877-machine-learning-ecmwf-roadmap-next-10-years|title=Machine learning at ECMWF: A roadmap for the next 10 years|last=Düben|first=Peter|last2=Modigliani|first2=Umberto|date=2021|website=www.ecmwf.int|access-date=2021-01-25|last3=Geer|first3=Alan|last4=Siemen|first4=Stephan|last5=Pappenberger|first5=Florian|last6=Bauer|first6=Peter|last7=Brown|first7=Andy|last8=Palkovic|first8=Martin|last9=Raoult|first9=Baudouin}}</ref> based on the observed initial conditions, and post-processing the output from weather models<ref>{{Cite journal|last=Rasp|first=Stephan|last2=Lerch|first2=Sebastian|date=2018-11-01|title=Neural Networks for Postprocessing Ensemble Weather Forecasts|url=https://journals.ametsoc.org/view/journals/mwre/146/11/mwr-d-18-0187.1.xml|journal=Monthly Weather Review|language=EN|volume=146|issue=11|pages=3885–3900|doi=10.1175/MWR-D-18-0187.1|issn=1520-0493}}</ref>. |
||
Latest revision as of 13:41, 26 August 2021
This page is about how machine learning (ML) can be used in the context of weather forecasting. For an overview of weather forecasting more generally, please see the Wikipedia page on this topic.
ML is a suitable tool for making short-term weather predictions[1] based on the observed initial conditions, and post-processing the output from weather models[2].
ML Application Areas[edit | edit source]
- Storm tracking: While climate models can forecast long-term changes in the climate system, separate systems are required to detect specific extreme weather phenomena, like cyclones, atmospheric rivers, and tornadoes. Identifying extreme events in climate model outputs can inform scientific understanding of where and when these events may occur. ML can help classify, detect, and track climate-related extreme events such as hurricanes in climate model outputs.
- Dust storm Prediction: Dust storms affect people, their properties, and their activities. For this reason, it is crucial to adopt automatic systems by using machine learning to predict or at least enable early detection of dust storms to reduce their deleterious impacts.
- Postprocessing of the output of weather/climate models: The ML model gets climate/weather model as inputs and learn patterns to improve these predictions.
- Forward prediction model: The ML model integrates initial conditions into the future. Iterate the ML prediction to predict further into the future.
Background Readings[edit | edit source]
Conferences, Journals, and Professional Organizations[edit | edit source]
- Opportunities and challenges for machine learning in weather and climate modelling: hard, medium and soft AI[3]
Libraries and Tools[edit | edit source]
- forecast verification metrics for gridded climate data
- forecast verification for gridded climate data
- Deep learning models for global weather prediction on a cubed sphere
- post-processing experiments with neural networks
- A benchmark dataset for data-driven weather forecasting
Data[edit | edit source]
Future Directions[edit | edit source]
Relevant Groups and Organizations[edit | edit source]
References[edit | edit source]
- ↑ Düben, Peter; Modigliani, Umberto; Geer, Alan; Siemen, Stephan; Pappenberger, Florian; Bauer, Peter; Brown, Andy; Palkovic, Martin; Raoult, Baudouin (2021). "Machine learning at ECMWF: A roadmap for the next 10 years". www.ecmwf.int. Retrieved 2021-01-25.
- ↑ Rasp, Stephan; Lerch, Sebastian (2018-11-01). "Neural Networks for Postprocessing Ensemble Weather Forecasts". Monthly Weather Review. 146 (11): 3885–3900. doi:10.1175/MWR-D-18-0187.1. ISSN 1520-0493.
- ↑ Chantry, Matthew; Christensen, Hannah; Dueben, Peter; Palmer, Tim (2021). "Opportunities and challenges for machine learning in weather and climate modelling: hard, medium and soft AI".