Improving Data‐Driven Global Weather Prediction Using Deep Convolutional Neural Networks on a Cubed Sphere

By Jonathan A. Weyn, Dale R. Durran, R. Caruana. Journal of Advances in Modeling Earth Systems, 2020

340 citations recorded by the source corpus.

We present a significantly improved data‐driven global weather forecasting framework using a deep convolutional neural network (CNN) to forecast several basic atmospheric variables on a global grid. New developments in this framework include an off‐line volume‐conservative mapping to a cubed‐sphere grid, improvements to the CNN architecture and the minimization of the loss function over multiple steps in a prediction sequence. The cubed‐sphere remapping minimizes the distortion on the cube faces on which convolution operations are performed and provides natural boundary conditions for padding in the CNN. Our improved model produces weather forecasts that are indefinitely stable and produce realistic weather patterns at lead times of several weeks and longer. For short‐ to medium‐range forecasting, our model significantly outperforms persistence, climatology, and a coarse‐resolution dynamical numerical weather prediction (NWP) model. Unsurprisingly, our forecasts are worse than those from a high‐resolution state‐of‐the‐art operational NWP system. Our data‐driven model is able to learn to forecast complex surface temperature patterns from few input atmospheric state variables. On annual time scales, our model produces a realistic seasonal cycle driven solely by the prescribed variation in top‐of‐atmosphere solar forcing. Although it currently does not compete with operational weather forecasting models, our data‐driven CNN executes much faster than those models, suggesting that machine learning could prove to be a valuable tool for large‐ensemble forecasting.

Subfields: Atmospheric Science, Global and Planetary Change, Statistical and Nonlinear Physics

View Paper: https://www.semanticscholar.org/paper/294f64f1890376795003e217ee339708d5347b1d

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 Global and Planetary Change

In Statistical and Nonlinear Physics

Citations in the corpus

What's Behind the Couch? Directed Ray Distance Functions (DRDF) for 3D Scene Reconstruction (2021, neutral): “In many cases, the expected value is precisely what is desired like in object detection [59, 71] or in ARIMA models [44], weather prediction [66] but in others it leads to poor outcomes.”
Spherical Fourier Neural Operators: Learning Stable Dynamics on the Sphere (2023, background): “Weyn et al. (2020; 2021) noticed the limitations of this representation when treating the poles and proposed an extension based on the equiangular gnomonic cubed sphere.”
ClimaX: A foundation model for weather and climate (2023, neutral): “…models in many scenarios, such as nowcasting of precipitation (Ravuri et al., 2021; Sønderby et al., 2020) and medium-range forecasting of variables like temperature, wind and humidity (Weyn et al., 2020; Rasp & Thuerey, 2021; Keisler, 2022; Pathak et al., 2022; Bi et al., 2022; Lam et al., 2022).”
Neural general circulation models for weather and climate (2023, background): “…janniyuval@google.com; shoyer@google.com although there has been some success in using machine-learning approaches on longer timescales 17,18 , these models have not demonstrated the ability to outperform existing GCMs. Hybrid models that combine GCMs with machine learning are…”
WeatherBench 2: A Benchmark for the Next Generation of Data‐Driven Global Weather Models (2023, methodology): “Weyn et al. (2020) used a cubed sphere projection in combination with a UNet architecture to iteratively predict the atmospheric state at 2° resolution.”
FourCastNet: A Global Data-driven High-resolution Weather Model using Adaptive Fourier Neural Operators (2022, methodology): “The authors work with a coarser resolution of 2 ◦ and forecast variables relating to geopotential heights, geopotential thickness, and 2-m temperature (see [Weyn et al., 2020] for further details).”
FuXi: a cascade machine learning forecasting system for 15-day global weather forecast (2023, background): “[9] proposed a multi-time-step loss function to minimize errors over multiple iterated time steps.”