Regional climate risk assessment from climate models using probabilistic machine learning

2024/12/11 by Zhong Yi Wan, Zhong Wan, Ignacio Lopez-Gomez +15 · 1 voice · 1 citation
Computer Science · Mathematics · Physics and Astronomy · #Bayesian Methods and Mixture Models #cs.LG #math.NA #physics.ao-ph

paper · pdf · doi:10.48550/arxiv.2412.08079

arxiv published 2024/12/11 · arxiv updated 2026/04/07

Abstract

Effective climate risk assessment is hindered by the resolution gap between coarse global climate models and the fine-scale information needed for regional decisions. We introduce GenFocal, an AI framework that generates statistically accurate, fine-scale weather from coarse climate projections, without requiring paired simulated and observed events during training. GenFocal synthesizes complex and long-lived hazards, such as heat waves and tropical cyclones, even when they are not well represented in the coarse climate projections. It also samples high-impact, rare events more accurately than leading methods. By translating large-scale climate projections into actionable, localized information, GenFocal provides a powerful new paradigm to improve climate adaptation and resilience strategies.

Cited by

• STLDM: Spatio-Temporal Latent Diffusion Model for Precipitation Nowcasting

Discussions

• How transportation of measures can be used for statistical downscaling, ie, from low-resolution to high-resolution while they are samples from two different manifolds? Our latest work arxiv.org/abs/ [bsky, 3 points, 0 comments]

Related

• Combining deep generative models with extreme value theory for synthetic hazard simulation: a multivariate and spatially coherent approach
• Learning Compound Climate Extremes: Generative AI for Hot–Dry Event Risk
• A Multi-Scale Deep Learning Framework for Projecting Weather Extremes
• Downscaling Extreme Rainfall Using Physical-Statistical Generative Adversarial Learning
• Accessing Convective Hazards Frequency Shift with Climate Change using Physics-Informed Machine Learning