GEN2: A Generative Prediction-Correction Framework for Long-time Emulations of Spatially-Resolved Climate Extremes

2025/08/21 by Mengze Wang, Wang, Mengze, Benedikt Barthel Sorensen +3 · 1 citation
Earth and Planetary Sciences · Engineering · Environmental Science · #Chaotic Dynamics (nlin.CD) #Climate variability and models #Computational Physics (physics.comp-ph) #Energy Load and Power Forecasting #FOS: Computer and information sciences #FOS: Physical sciences #Machine Learning (cs.LG) #Meteorological Phenomena and Simulations

paper · pdf · doi:10.48550/arxiv.2508.15196

openalex publication_date 2025/08/21 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Accurately quantifying the increased risks of climate extremes requires generating large ensembles of climate realization across a wide range of emissions scenarios, which is computationally challenging for conventional Earth System Models. We propose GEN2, a generative prediction-correction framework for an efficient and accurate forecast of the extreme event statistics. The prediction step is constructed as a conditional Gaussian emulator, followed by a non-Gaussian machine-learning (ML) correction step. The ML model is trained on pairs of the reference data and the emulated fields nudged towards the reference, to ensure the training is robust to chaos. We first validate the accuracy of our model on historical ERA5 data and then demonstrate the extrapolation capabilities on various future climate change scenarios. When trained on a single realization of one warming scenario, our model accurately predicts the statistics of extreme events in different scenarios, successfully extrapolating beyond the distribution of training data.

Citations

• Dynamical-generative downscaling of climate model ensembles
• The impact of internal variability on benchmarking deep learning climate emulators
• A probabilistic framework for learning non-intrusive corrections to long-time climate simulations from short-time training data
• A non-intrusive machine learning framework for debiasing long-time coarse resolution climate simulations and quantifying rare events statistics
• Learning skillful medium-range global weather forecasting
• ClimateSet: A Large-Scale Climate Model Dataset for Machine Learning
• ACE: A fast, skillful learned global atmospheric model for climate prediction
• Accurate medium-range global weather forecasting with 3D neural networks
• Debias Coarsely, Sample Conditionally: Statistical Downscaling through Optimal Transport and Probabilistic Diffusion Models
• Unpaired Downscaling of Fluid Flows with Diffusion Bridges
• Stochastic Interpolants: A Unifying Framework for Flows and Diffusions
• Stabilizing Machine Learning Prediction of Dynamics: Noise and Noise-inspired Regularization
• Flow Matching for Generative Modeling
• FourCastNet: A Global Data-driven High-resolution Weather Model using Adaptive Fourier Neural Operators
• Conditional Image Generation with Score-Based Diffusion Models
• Score-Based Generative Modeling through Stochastic Differential Equations
• The ERA5 global reanalysis
• Deep learning to represent subgrid processes in climate models
• Deep Residual Learning for Image Recognition
• U-Net: Convolutional Networks for Biomedical Image Segmentation

Cited by

• Score-based generative emulation of impact-relevant Earth system model outputs

Related

• A non-intrusive machine learning framework for debiasing long-time coarse resolution climate simulations and quantifying rare events statistics
• Loosely Conditioned Emulation of Global Climate Models With Generative Adversarial Networks
• Learning Compound Climate Extremes: Generative AI for Hot–Dry Event Risk
• Generative artificial intelligence improves projections of climate extremes
• Gaussian Framework and Optimal Projection of Weather Fields for Prediction of Extreme Events