FourCastNet, short for Fourier Forecasting Neural Network, is a global data-driven weather forecasting model that provides accurate short to medium-range global predictions at \$0.25ˆ\\textbackslashcirc\\$ resolution. FourCastNet accurately forecasts high-resolution, fast-timescale variables such as the surface wind speed, precipitation, and atmospheric water vapor. It has important implications for planning wind energy resources, predicting extreme weather events such as tropical cyclones, extra-tropical cyclones, and atmospheric rivers. FourCastNet matches the forecasting accuracy of the ECMWF Integrated Forecasting System (IFS), a state-of-the-art Numerical Weather Prediction (NWP) model, at short lead times for large-scale variables, while outperforming IFS for variables with complex fine-scale structure, including precipitation. FourCastNet generates a week-long forecast in less than 2 seconds, orders of magnitude faster than IFS. The speed of FourCastNet enables the creation of rapid and inexpensive large-ensemble forecasts with thousands of ensemble-members for improving probabilistic forecasting. We discuss how data-driven deep learning models such as FourCastNet are a valuable addition to the meteorology toolkit to aid and augment NWP models.
arXiv Fulltext PDF:/Users/pascal/Zotero/storage/8IFX2TBW/Pathak et al. - 2022 - FourCastNet A Global Data-driven High-resolution .pdf:application/pdf;arXiv.org Snapshot:/Users/pascal/Zotero/storage/KTIF8CTD/2202.html:text/html
%0 Generic
%1 pathak_fourcastnet_2022
%A Pathak, Jaideep
%A Subramanian, Shashank
%A Harrington, Peter
%A Raja, Sanjeev
%A Chattopadhyay, Ashesh
%A Mardani, Morteza
%A Kurth, Thorsten
%A Hall, David
%A Li, Zongyi
%A Azizzadenesheli, Kamyar
%A Hassanzadeh, Pedram
%A Kashinath, Karthik
%A Anandkumar, Animashree
%D 2022
%I arXiv
%K climate deeplearning idea:big_data_geo_2 prediction todo:read weather
%R 10.48550/arXiv.2202.11214
%T FourCastNet: A Global Data-driven High-resolution Weather Model using Adaptive Fourier Neural Operators
%U http://arxiv.org/abs/2202.11214
%X FourCastNet, short for Fourier Forecasting Neural Network, is a global data-driven weather forecasting model that provides accurate short to medium-range global predictions at \$0.25ˆ\\textbackslashcirc\\$ resolution. FourCastNet accurately forecasts high-resolution, fast-timescale variables such as the surface wind speed, precipitation, and atmospheric water vapor. It has important implications for planning wind energy resources, predicting extreme weather events such as tropical cyclones, extra-tropical cyclones, and atmospheric rivers. FourCastNet matches the forecasting accuracy of the ECMWF Integrated Forecasting System (IFS), a state-of-the-art Numerical Weather Prediction (NWP) model, at short lead times for large-scale variables, while outperforming IFS for variables with complex fine-scale structure, including precipitation. FourCastNet generates a week-long forecast in less than 2 seconds, orders of magnitude faster than IFS. The speed of FourCastNet enables the creation of rapid and inexpensive large-ensemble forecasts with thousands of ensemble-members for improving probabilistic forecasting. We discuss how data-driven deep learning models such as FourCastNet are a valuable addition to the meteorology toolkit to aid and augment NWP models.
@misc{pathak_fourcastnet_2022,
abstract = {FourCastNet, short for Fourier Forecasting Neural Network, is a global data-driven weather forecasting model that provides accurate short to medium-range global predictions at \$0.25{\textasciicircum}\{{\textbackslash}circ\}\$ resolution. FourCastNet accurately forecasts high-resolution, fast-timescale variables such as the surface wind speed, precipitation, and atmospheric water vapor. It has important implications for planning wind energy resources, predicting extreme weather events such as tropical cyclones, extra-tropical cyclones, and atmospheric rivers. FourCastNet matches the forecasting accuracy of the ECMWF Integrated Forecasting System (IFS), a state-of-the-art Numerical Weather Prediction (NWP) model, at short lead times for large-scale variables, while outperforming IFS for variables with complex fine-scale structure, including precipitation. FourCastNet generates a week-long forecast in less than 2 seconds, orders of magnitude faster than IFS. The speed of FourCastNet enables the creation of rapid and inexpensive large-ensemble forecasts with thousands of ensemble-members for improving probabilistic forecasting. We discuss how data-driven deep learning models such as FourCastNet are a valuable addition to the meteorology toolkit to aid and augment NWP models.},
added-at = {2023-08-11T14:24:26.000+0200},
author = {Pathak, Jaideep and Subramanian, Shashank and Harrington, Peter and Raja, Sanjeev and Chattopadhyay, Ashesh and Mardani, Morteza and Kurth, Thorsten and Hall, David and Li, Zongyi and Azizzadenesheli, Kamyar and Hassanzadeh, Pedram and Kashinath, Karthik and Anandkumar, Animashree},
biburl = {https://www.bibsonomy.org/bibtex/27c725e2908978d066179e178c3bc9a61/annakrause},
doi = {10.48550/arXiv.2202.11214},
file = {arXiv Fulltext PDF:/Users/pascal/Zotero/storage/8IFX2TBW/Pathak et al. - 2022 - FourCastNet A Global Data-driven High-resolution .pdf:application/pdf;arXiv.org Snapshot:/Users/pascal/Zotero/storage/KTIF8CTD/2202.html:text/html},
interhash = {9c23ece78ff16fa3f6c14e3f5275cd0e},
intrahash = {7c725e2908978d066179e178c3bc9a61},
keywords = {climate deeplearning idea:big_data_geo_2 prediction todo:read weather},
month = feb,
note = {arXiv:2202.11214 [physics]},
publisher = {arXiv},
shorttitle = {{FourCastNet}},
timestamp = {2023-08-11T14:25:33.000+0200},
title = {{FourCastNet}: {A} {Global} {Data}-driven {High}-resolution {Weather} {Model} using {Adaptive} {Fourier} {Neural} {Operators}},
url = {http://arxiv.org/abs/2202.11214},
urldate = {2023-07-10},
year = 2022
}