Opportunities and challenges for machine learning in weather and climate modelling: hard, medium and soft AI

Chantry, Matthew; Christensen, Hilary; Dueben, Peter; Palmer, Tim

doi:10.1098/rsta.2020.0083

Cited by 92 publications

(59 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The recent boom in both hardware and software developments around machine learning has caused many fields to examine the possible boons that machine learning can bring. In the field of weather and climate forecasting, researchers are examining the applicability of machine learning techniques to a spectrum of problems (Chantry et al., 2021 ), covering changes from the seismic to the incremental. Seismic changes include investigations into whether machine learning can replace the whole forecasting system, either by learning from observational data (Sønderby et al., 2020 ) or atmospheric reanalysis (Rasp et al., 2020 ).…”

Section: Introductionmentioning

confidence: 99%

Machine Learning Emulation of Gravity Wave Drag in Numerical Weather Forecasting

Chantry

Hatfield

Dueben

et al. 2021

J Adv Model Earth Syst

Self Cite

View full text Add to dashboard Cite

We assess the value of machine learning as an accelerator for the parameterisation schemes of operational weather forecasting systems, specifically the parameterisation of non-orographic gravity wave drag. Emulators of this scheme can be trained to produce stable and accurate results up to seasonal forecasting timescales. Generally, more complex networks produce more accurate emulators. By training on an increased complexity version of the existing parameterisation scheme we build emulators that produce more accurate forecasts. For medium range forecasting we find evidence our emulators are more accurate than the version of the parametrisation scheme that is used for operational predictions.Using the current operational CPU hardware our emulators have a similar computational cost to the existing scheme, but are heavily limited by data movement. On GPU hardware our emulators perform ten times faster than the existing scheme on a CPU.

show abstract

Section: Introductionmentioning

confidence: 99%

Machine Learning Emulation of Gravity Wave Drag in Numerical Weather Forecasting

Chantry

Hatfield

Dueben

et al. 2021

J Adv Model Earth Syst

Self Cite

View full text Add to dashboard Cite

show abstract

“…There is now increasing interest around developing data-driven Deep Learning (DL) models for weather forecasting owing to their orders of magnitude lower computational cost as compared to state-of-the-art NWP models [Schultz et al, 2021, Balaji, 2021, Irrgang et al, 2021, Reichstein et al, 2019. Many studies have attempted to build data-driven models for forecasting the large-scale circulation of the atmosphere, either trained on climate model outputs, general circulation models (GCM) [Scher and Messori, 2018, 2019, Chattopadhyay et al, 2020a, reanalysis products [Weyn et al, 2019, 2021, Rasp and Thuerey, 2021a, Arcomano et al, 2020, Chantry et al, 2021, Grönquist et al, 2021, or a blend of climate model outputs and reanalysis products [Rasp and Thuerey, 2021a].…”

Section: Introductionmentioning

confidence: 99%

FourCastNet: A Global Data-driven High-resolution Weather Model using Adaptive Fourier Neural Operators

Pathak¹,

Subramanian²,

Harrington³

et al. 2022

Preprint

135

158

View full text Add to dashboard Cite

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 • 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 small-scale variables, 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.

show abstract

“…In the area of weather and climate forecasting, supervised learning has been used for various purposes (see e.g., Reichstein et al, 2019;Chantry et al, 2021;Düben et al, 2021, and references therein). These include: (i) to emulate the full dynamics of a system (Pathak et al, 2017(Pathak et al, , 2018Fablet et al, 2018;Nguyen et al, 2019;Brajard et al, 2020;Patel et al, 2021;Schultz et al, 2021;Sonnewald et al, 2021), (ii) to improve a physics-based model with data-driven correction or parameterisation (O'Gorman and Dwyer, 2018;Rasp et al, 2018;Bolton and Zanna, 2019;Rasp, 2020;Bonavita and Laloyaux, 2020;Nguyen et al, 2021;Gottwald and Reich, 2021;.…”

Section: Introductionmentioning

confidence: 99%

Supervised machine learning to estimate instabilities in chaotic systems: estimation of local Lyapunov exponents

Ayers¹,

Lau²,

Amezcua³

et al. 2022

Preprint

View full text Add to dashboard Cite

In chaotic dynamical systems such as the weather, prediction errors grow faster in some situations than in others.Real-time knowledge about the error growth could enable strategies to adjust the modelling and forecasting infrastructure on-the-fly to increase accuracy and/or reduce computation time. For example one could change the spatiotemporal resolution of the numerical model, locally increase the data availability, etc. Local Lyapunov exponents are known indicators of the rate at which very small prediction errors grow over a finite time interval. However, their computation is very expensive: it requires maintaining and evolving a tangent linear model, orthogonalisation algorithms and storing large matrices. In this study, we investigate the capability of supervised machine learning to estimate the imminent local Lyapunov exponents, from input of current and recent time steps of the system trajectory, as an alternative to the classical method. Thus machine learning is not used here to emulate a physical model or some of its components, but "non intrusively" as a complementary tool. Specifically, we test the accuracy of four popular supervised learning algorithms: regression trees, multilayer perceptrons, convolutional neural networks and long short-term memory networks. Experiments are conducted 1

show abstract

Opportunities and challenges for machine learning in weather and climate modelling: hard, medium and soft AI

Cited by 92 publications

References 36 publications

Machine Learning Emulation of Gravity Wave Drag in Numerical Weather Forecasting

Machine Learning Emulation of Gravity Wave Drag in Numerical Weather Forecasting

FourCastNet: A Global Data-driven High-resolution Weather Model using Adaptive Fourier Neural Operators

Supervised machine learning to estimate instabilities in chaotic systems: estimation of local Lyapunov exponents

Contact Info

Product

Resources

About