General circulation models (GCMs) are the foundation of weather and climate prediction^1,2. GCMs are physics-based simulators that combine a numerical solver for large-scale dynamics with tuned representations for small-scale processes such as cloud formation. Recently, machine-learning models trained on reanalysis data have achieved comparable or better skill than GCMs for deterministic weather forecasting^3,4. However, these models have not demonstrated improved ensemble forecasts, or shown sufficient stability for long-term weather and climate simulations. Here we present a GCM that combines a differentiable solver for atmospheric dynamics with machine-learning components and show that it can generate forecasts of deterministic weather, ensemble weather and climate on par with the best machine-learning and physics-based methods. NeuralGCM is competitive with machine-learning models for one- to ten-day forecasts, and with the European Centre for Medium-Range Weather Forecasts ensemble prediction for one- to fifteen-day forecasts. With prescribed sea surface temperature, NeuralGCM can accurately track climate metrics for multiple decades, and climate forecasts with 140-kilometre resolution show emergent phenomena such as realistic frequency and trajectories of tropical cyclones. For both weather and climate, our approach offers orders of magnitude computational savings over conventional GCMs, although our model does not extrapolate to substantially different future climates. Our results show that end-to-end deep learning is compatible with tasks performed by conventional GCMs and can enhance the large-scale physical simulations that are essential for understanding and predicting the Earth system.

A review of the experimental protocol and motivation for DYAMOND, the first intercomparison project of global storm-resolving models, is presented. Nine models submitted simulation output for a 40-day (1 August–10 September 2016) intercomparison period. Eight of these employed a tiling of the sphere that was uniformly less than 5 km. By resolving the transient dynamics of convective storms in the tropics, global storm-resolving models remove the need to parameterize tropical deep convection, providing a fundamentally more sound representation of the climate system and a more natural link to commensurately high-resolution data from satellite-borne sensors. The models and some basic characteristics of their output are described in more detail, as is the availability and planned use of this output for future scientific study. Tropically and zonally averaged energy budgets, precipitable water distributions, and precipitation from the model ensemble are evaluated, as is their representation of tropical cyclones and the predictability of column water vapor, the latter being important for tropical weather.

Data‐driven approaches, most prominently deep learning, have become powerful tools for prediction in many domains. A natural question to ask is whether data‐driven methods could also be used to predict global weather patterns days in advance. First studies show promise but the lack of a common data set and evaluation metrics make intercomparison between studies difficult. Here we present a benchmark data set for data‐driven medium‐range weather forecasting (specifically 3–5 days), a topic of high scientific interest for atmospheric and computer scientists alike. We provide data derived from the ERA5 archive that has been processed to facilitate the use in machine learning models. We propose simple and clear evaluation metrics which will enable a direct comparison between different methods. Further, we provide baseline scores from simple linear regression techniques, deep learning models, as well as purely physical forecasting models. The data set is publicly available at https://github.com/pangeo‐data/WeatherBench and the companion code is reproducible with tutorials for getting started. We hope that this data set will accelerate research in data‐driven weather forecasting.

WeatherBench 2 is an update to the global, medium‐range (1–14 days) weather forecasting benchmark proposed by (Rasp et al., 2020, https://doi.org/10.1029/2020ms002203), designed with the aim to accelerate progress in data‐driven weather modeling. WeatherBench 2 consists of an open‐source evaluation framework, publicly available training, ground truth and baseline data as well as a continuously updated website with the latest metrics and state‐of‐the‐art models: https://sites.research.google/weatherbench. This paper describes the design principles of the evaluation framework and presents results for current state‐of‐the‐art physical and data‐driven weather models. The metrics are based on established practices for evaluating weather forecasts at leading operational weather centers. We define a set of headline scores to provide an overview of model performance. In addition, we also discuss caveats in the current evaluation setup and challenges for the future of data‐driven weather forecasting.

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.

Purpose of Review Global cloud-resolving models (GCRMs) are a new type of atmospheric model which resolve nonhydrostatic accelerations globally with kilometer-scale resolution. This review explains what distinguishes GCRMs from other types of models, the problems they solve, and the questions their more commonplace use is raising. Recent Findings GCRMs require high-resolution discretization over the sphere but can differ in many other respects. They are beginning to be used as a main stream research tool. The first GCRM intercomparison studies are being coordinated, raising new questions as to how best to exploit their advantages. Summary GCRMs are designed to resolve the multiscale nature of moist convection in the global dynamics context, without using cumulus parameterization. Clouds are simulated with cloud microphysical schemes in ways more comparable to observations. Because they do not suffer from ambiguity arising from cumulus parameterization, as computational resources increase, GCRMs are the promise of a new generation of global weather and climate simulations.