Neural General Circulation Models for Weather and Climate

Abstract

General circulation models (GCMs) are the foundation of weather and climate prediction. GCMs are physics-based simulators which combine a numerical solver for large-scale dynamics with tuned representations for small-scale processes such as cloud formation. Recently, machine learning (ML) models trained on reanalysis data achieved comparable or better skill than GCMs for deterministic weather forecasting. However, these models have not demonstrated improved ensemble forecasts, or shown sufficient stability for long-term weather and climate simulations. Here we present the first GCM that combines a differentiable solver for atmospheric dynamics with ML components, and show that it can generate forecasts of deterministic weather, ensemble weather and climate on par with the best ML and physics-based methods. NeuralGCM is competitive with ML models for 1-10 day forecasts, and with the European Centre for Medium-Range Weather Forecasts ensemble prediction for 1-15 day forecasts. With prescribed sea surface temperature, NeuralGCM can accurately track climate metrics such as global mean temperature for multiple decades, and climate forecasts with 140 km resolution exhibit 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. 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.

arXiv comment92 pages, 54 figures. Nature (2024)

References in corpus (10)

• Neural networks for post-processing ensemble weather forecasts
• Earth System Modeling 2.0: A Blueprint for Models That Learn From Observations and Targeted High-Resolution Simulations
• Stable machine-learning parameterization of subgrid processes for climate modeling at a range of resolutions
• Improving data-driven global weather prediction using deep convolutional neural networks on a cubed sphere
• Spatially Extended Tests of a Neural Network Parametrization Trained by Coarse-graining
• Are General Circulation Models obsolete?
• Non-local parameterization of atmospheric subgrid processes with neural networks
• WeatherBench 2: A benchmark for the next generation of data-driven global weather models
• ACE: A fast, skillful learned global atmospheric model for climate prediction
• Differentiable Programming for Earth System Modeling

Cited by in corpus (26)

• FengWu: Pushing the Skillful Global Medium-range Weather Forecast beyond 10 Days Lead
• Can AI weather models predict out-of-distribution gray swan tropical cyclones?
• Validating Deep Learning Weather Forecast Models on Recent High-Impact Extreme Events
• A Practical Probabilistic Benchmark for AI Weather Models
• FuXi Weather: A data-to-forecast machine learning system for global weather
• Samudra: An AI Global Ocean Emulator for Climate
• Global spatio-temporal downscaling of ERA5 precipitation through generative AI
• Automatic Differentiation-based Full Waveform Inversion with Flexible Workflows
• Reconstructing Historical Climate Fields With Deep Learning
• Generalization capabilities and robustness of hybrid models grounded in physics compared to purely deep learning models
• ClimateIQA: A New Dataset and Benchmark to Advance Vision-Language Models in Meteorology Anomalies Analysis
• Physics-Guided Learning of Meteorological Dynamics for Weather Downscaling and Forecasting
• Differentiable Cosmological Hydrodynamics for Field-Level Inference and High Dimensional Parameter Constraints
• Physically Consistent Global Atmospheric Data Assimilation with Machine Learning in Latent Space
• A Spatiotemporal Radar-Based Precipitation Model for Water Level Prediction and Flood Forecasting
• Seasonal forecasting using the GenCast probabilistic machine learning model
• Ensemble data assimilation to diagnose AI-based weather prediction model: A case with ClimaX version 0.3.1
• A Mechanism-Learning Deeply Coupled Model for Remote Sensing Retrieval of Global Land Surface Temperature
• Constructing Extreme Heatwave Storylines with Differentiable Climate Models
• Electron neural closure for turbulent magnetosheath simulations: energy channels
• FM-LLM: A frequency-enhanced mixture-of-experts framework for adapting LLMs to time series forecasting
• Physics-constrained generative machine learning-based high-resolution downscaling of Greenland's surface mass balance and surface temperature
• Projecting U.S. coastal storm surge risks and impacts with deep learning
• A robust and stable hybrid neural network/finite element method for 2D flows that generalizes to different geometries
• Data-driven Seasonal Climate Predictions via Variational Inference and Transformers
• Efficient fine-tuning of 37-level GraphCast with the Canadian global deterministic analysis