Generative artificial intelligence improves projections of climate extremes

Climate change is amplifying extreme events, posing escalating risks to biodiversity, human health, and food security. GCMs are essential for projecting future climate, yet their coarse resolution and high computational costs constrain their ability to represent extremes. Here, we introduce FuXi-CMIPAlign, a generative deep learning framework for downscaling CMIP outputs. The model integrates Flow Matching for generative modeling with domain adaptation via MMD loss to align feature distributions between training data and inference data, thereby mitigating input discrepancies and improving accuracy, stability, and generalization across emission scenarios. FuXi-CMIPAlign performs spatial, temporal, and multivariate downscaling, enabling more realistic simulation of compound extremes such as TCs.

• Publication year

  2025

• Venue

  Unknown venue

• Publication date

  2025-08-22

• Fields of study

  Physics, Computer Science, Environmental Science

• Identifiers
  arXiv 2508.16396
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Extreme Cold Days and Spells in Northern Europe at 0.5°C–2.0°C Global Warming Levels

  2025cited by this paper
• Statistical downscaling differences strongly alter projected climate damages

  2025cited by this paper
• Large reductions in tropical bird abundance attributable to heat extreme intensification

  2025cited by this paper
• Record-breaking extremes in a warming climate

  2025cited by this paper
• Global emergence of unprecedented lifetime exposure to climate extremes

  2025cited by this paper
• Precipitation extremes in 2023

  2024cited by this paper
• Global supply chains amplify economic costs of future extreme heat risk

  2024cited by this paper
• Climate threats to coastal infrastructure and sustainable development outcomes

  2024cited by this paper
• Deep learning in statistical downscaling for deriving high spatial resolution gridded meteorological data: A systematic review

  2024cited by this paper
• Europe faces up to tenfold increase in extreme fires in a warming climate

  2024cited by this paper
• Fast, scale-adaptive and uncertainty-aware downscaling of Earth system model fields with generative machine learning

  2024influential reference
• Droughts and floods in a changing climate and implications for multi-hazard urban planning: A review

  2024cited by this paper
• The ERA5 global reanalysis from 1940 to 2022

  2024cited by this paper
• Quantifying cascading power outages during climate extremes considering renewable energy integration

  2024cited by this paper
• Climate change impacts on plant pathogens, food security and paths forward

  2023cited by this paper
• Evaluation of statistical downscaling techniques and projection of climate extremes in central Texas, USA

  2023cited by this paper
• The 2023 report of the Lancet Countdown on health and climate change: the imperative for a health-centred response in a world facing irreversible harms

  2023cited by this paper
• Climate change will accelerate the high-end risk of compound drought and heatwave events

  2023cited by this paper
• The global costs of extreme weather that are attributable to climate change

  2023cited by this paper
• Downscaling multi-model climate projection ensembles with deep learning (DeepESD): contribution to CORDEX EUR-44

  2022cited by this paper
• The effect of rainfall changes on economic production

  2022cited by this paper
• Global transportation infrastructure exposure to the change of precipitation in a warmer world

  2022cited by this paper
• Flow Matching for Generative Modeling

  2022cited by this paper
• Can low-resolution CMIP6 ScenarioMIP models provide insight into future European post-tropical-cyclone risk?

  2022cited by this paper
• TempestExtremes v2.1: A Community Framework for Feature Detection, Tracking and Analysis in Large Datasets

  2021cited by this paper
• Evaluation of extreme precipitation over Asia in CMIP6 models

  2021cited by this paper
• The EC-Earth3 Earth System Model for the Climate Model Intercomparison Project 6

  2021cited by this paper
• The emergence of heat and humidity too severe for human tolerance

  2020cited by this paper
• Evaluation of the CMIP6 multi-model ensemble for climate extreme indices

  2020cited by this paper
• Projected Future Changes in Tropical Cyclones Using the CMIP6 HighResMIP Multimodel Ensemble

  2020cited by this paper
• The ERA5 global reanalysis

  2020cited by this paper
• The Nippon Foundation-GEBCO Seabed 2030 Project : The Quest to See the World's Oceans Completely Mapped by 2030

  2018cited by this paper
• Learning Blind Video Temporal Consistency

  2018cited by this paper
• Dynamical downscaling of regional climate: A review of methods and limitations

  2018cited by this paper
• DeepSD: Generating High Resolution Climate Change Projections through Single Image Super-Resolution

  2017cited by this paper
• High Resolution Model Intercomparison Project (HighResMIP v1.0) for CMIP6

  2016cited by this paper
• The Scenario Model Intercomparison Project (ScenarioMIP) for CMIP6

  2016cited by this paper
• Bias Correction, Quantile Mapping, and Downscaling: Revisiting the Inflation Issue

  2013cited by this paper
• The Statistical DownScaling Model: insights from one decade of application

  2013cited by this paper
• A Kernel Method for the Two-Sample-Problem

  2006cited by this paper
• The Coupled Model Intercomparison Project (CMIP)

  2000cited by this paper
• A COMPARISON OF DELTA CHANGE AND DOWNSCALED GCM SCENARIOS FOR THREE MOUNTAINOUS BASINS IN THE UNITED STATES 1

  2000cited by this paper
• Clivar/GCOS/WMO Workshop on Indices and Indicators for Climate Extremes Workshop Summary

  1999cited by this paper
• Climate downscaling: techniques and application

  1996cited by this paper

• Zero-Shot Statistical Downscaling via Diffusion Posterior Sampling

  2026cites this paper