Challenges and design choices for global weather and climate models based on machine learning

By Peter Dueben, Peter Bauer. 2018

372 citations recorded by the source corpus.

Abstract. Can models that are based on deep learning and trained on atmospheric data compete with weather and climate models that are based on physical principles and the basic equations of motion? This question has been asked often recently due to the boom in deep-learning techniques. The question is valid given the huge amount of data that are available, the computational efficiency of deep-learning techniques and the limitations of today's weather and climate models in particular with respect to resolution and complexity.In this paper, the question will be discussed in the context of global weather forecasts. A toy model for global weather predictions will be presented and used to identify challenges and fundamental design choices for a forecast system based on neural networks.

Subfields: Atmospheric Science, Environmental Engineering, Global and Planetary Change

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Atmospheric Science

In Environmental Engineering

In Global and Planetary Change

Citations in the corpus

Stratospheric Aerosol Injection as a Deep Reinforcement Learning Problem (2019, background): “Recently, the advent of deep learning has led to a number of successful emulation attempts of full GCMs used for weather prediction (Düben, 2018), as well as for sub-grid scale processes (Brenowitz, 2018; Rasp, 2018), including precipitation (O’Gorman, 2018).”
Can Machines Learn to Predict Weather? Using Deep Learning to Predict Gridded 500‐hPa Geopotential Height From Historical Weather Data (2019, methodology): “We improve upon the prior work of Dueben and Bauer (2018) by using more advanced NN architectures, more years of reanalysis data, and higher resolution (albeit over the northern hemisphere only), while in contrast to Scher (2018) and Scher and Messori (2019), we predict observed weather instead of…”
ClimaX: A foundation model for weather and climate (2023, neutral): “…a steady rise in data-driven approaches for forecasting of atmospheric variables, especially for meteorological applications (Grover et al., 2015; Dueben & Bauer, 2018; Weber et al., 2020; Scher & Messori, 2019; Scher, 2018; Kashinath et al., 2021; Schultz et al., 2021; Reichstein et al., 2019;…”
Improving Data‐Driven Global Weather Prediction Using Deep Convolutional Neural Networks on a Cubed Sphere (2020, methodology): “This is remarkable considering the difficulties that have been previously encountered in creating stable models of atmospheric flows from purely data-driven techniques [11] or with data-driven algorithms inserted as parameterizations in GCMs [e.”
FuXi: a cascade machine learning forecasting system for 15-day global weather forecast (2023, methodology): “This approach is similar to the time-stepping methods used in conventional NWP models [18].”
Forecast combinations: An over 50-year review (2022, background): “These techniques lead to improved computational efficiency in creating ensemble forecasts with much larger ensemble sizes (Dueben and Bauer, 2018; Scher, 2018; Rasp and Thuerey, 2021; Scher and Messori, 2021).”
WeatherBench 2: A Benchmark for the Next Generation of Data‐Driven Global Weather Models (2023, background): “A number of initial attempts at building data‐driven medium‐range NWP models (Dueben & Bauer, 2018; Scher, 2018; Weyn et al., 2019) led to the creation of WeatherBench (Rasp et al., 2020, from here on called WB1), a benchmark for global, medium‐range weather prediction.”
Process‐Guided Deep Learning Predictions of Lake Water Temperature (2019, background): “As evidence from another domain that the pursuit into forecasting should consider hybrid model architectures, Dueben and Bauer (2018) evaluated the potential for making simpli ﬁ ed forecasts of global weather with DL, concluding that while many bene ﬁ ts of DL are clear (especially for short ‐ term…”
Can deep learning beat numerical weather prediction? (2021, neutral): “There have also been some attempts to produce data-driven weather forecasts, for example by Dueben & Bauer [86], who used a multi-layer perceptron, or Grover et al. [16], who constructed a three-stage model consisting of boosted decision trees, a dynamic Gaussian Process model, and a deep belief net”
WeatherBench: A Benchmark Data Set for Data‐Driven Weather Forecasting (2020, background): “are from the Climate Forecast System (CFS) Reanalysis from 1979–2010 with 2.5-degree horizontal resolution and cropped to the northern hemisphere .”
Physics-informed machine learning: case studies for weather and climate modelling (2021, methodology): “Given the computational efficiency of ML techniques and the limitations of today’s numerical weather prediction and climate models, in particular, with respect to resolution and complexity, ML offers attractive alternatives for weather and climate forecasting [120].”