Skip to content
Open access

Regional climate risk assessment from climate models using probabilistic machine learning

Dec 2024 · Nature Machine Intelligence · Vol 8, pp. 1560 - 1568 · 17 citations · ⚡ 1 influential · 52 references
Computer Science Mathematics Physics

TL;DR

GenFocal is introduced, an AI framework that generates statistically accurate, fine-scale weather from coarse climate projections without requiring paired simulated and observed events during training, and provides a powerful paradigm to improve climate adaptation and resilience strategies.

Abstract

Effective climate risk assessment is hindered by the resolution gap between coarse global climate models and the fine-scale information needed for regional decision making. Existing downscaling methods struggle to bridge this gap: physics-based methods are too computationally expensive for large ensemble sampling of extreme events, while traditional statistical methods fail to capture multivariate spatiotemporal dependencies crucial to compound risk estimation. Furthermore, current machine learning approaches rely on temporally aligned training pairs, which are unavailable for free-running climate projections. Here 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 heatwaves 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 paradigm to improve climate adaptation and resilience strategies. A generative AI framework called GenFocal is introduced for climate downscaling, producing realistic fine-scale weather from coarse projections and improving regional risk estimates of compound extremes such as heatwaves and tropical cyclones.

Read PDF

Similar papers

Open access Oct 2026

Harnessing massive climate data for enhanced risk prediction: a GA-optimized LSTM and XGBoost hybrid approach

Prediction of climate risk has become an essential challenge owing to the heightened occurrence of extreme weather conditions caused by global climate change. Conventional methods tend to miss the spatiotemporal dependencies of multi-dimensional climate data, especially when using static models alone or shallow learnin...

Bahjat Fakieh · 0 citations
Open access Sep 2026

Climate-stratified assessment of PM2.5 using machine learning: Geographic controls dominate meteorological factors

A climate-stratified machine-learning framework using 49,997 monthly MERRA-2 satellite-assimilated samples across five latitude-band climate zones, evaluated through Leave-One-Climate-Out (LOCO) cross-validation with nested hyperparameter tuning to prevent data leakage is developed.

Someyo Kamal Utsho, M. R. Fahim, R. Samsudin · 0 citations

Learning Hierarchical Causal Representations of the Effects of Forcings on Temperature in Climate Models

A hierarchical causal representation learning framework applied to sea surface temperature fields from a state-of-the-art global climate model is developed that accurately predicts the long-term global mean and regional temperature evolution and shows physically realistic responses to perturbations in greenhouse gas an...

Shan Zhao, Ilija Trajković, Julia Kaltenborn et al. · 0 citations
Review Open access Sep 2026

Machine Learning–Enabled Climate Risk Scenarios for Local Government Decision-Making

Local governments face increasingly complex and compounding climate risks, yet the analytical outputs of climate risk assessment often remain weakly connected to municipal decision-making routines. Although machine learning (ML) has advanced hazard forecasting and impact estimation by exploiting high-dimensional data a...

Özden Şentürk · 0 citations
Oct 2026

Assessment of climate change impacts on hydropower potential in intermittent river basins using bootstrap aggregating machine learning method

Devastating environmental effects of relying too much on fossil fuels include air pollution, acid rain and global warming. One renewable energy source that is crucial for slowing climate change is hydropower. However, climate change affects its potential and makes hydropower generation unpredictable. The bootstrap ag...

Turan Alakbar, H. İ. Burgan · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.