---
id: ai-for-climate-science
title: "AI for Climate Science: Weather Prediction and Earth System Modeling"
schema_type: article
category: ai
language: en
confidence: medium
last_verified: "2026-05-28"
created_date: "2026-05-24"
generation_method: ai_structured
ai_models:
  - claude-4.5-sonnet
derived_from_human_seed: true
conflict_of_interest: none_declared
is_live_document: false
data_period: static
completeness: 0.85
atomic_facts:
  - id: af-ai-for-climate-science-1
    statement: >-
      The GraphCast paper presented a learned global weather forecasting system for medium-range
      forecasts.
    source_title: Learning skillful medium-range global weather forecasting
    source_url: https://www.science.org/doi/10.1126/science.adi2336
    confidence: medium
  - id: af-ai-for-climate-science-2
    statement: >-
      FourCastNet described a data-driven high-resolution global weather model based on adaptive
      Fourier neural operators.
    source_title: >-
      FourCastNet: A Global Data-driven High-resolution Weather Model using Adaptive Fourier Neural
      Operators
    source_url: https://arxiv.org/abs/2202.11214
    confidence: medium
  - id: af-ai-for-climate-science-3
    statement: >-
      The NeuralGCM paper described hybrid neural general circulation models for weather forecasting
      and climate simulation.
    source_title: Neural general circulation models for weather and climate
    source_url: https://www.nature.com/articles/s41586-024-07744-y
    confidence: medium
primary_sources:
  - id: ps-ai-for-climate-science-1
    title: Learning skillful medium-range global weather forecasting
    type: journal_article
    year: 2023
    authors:
      - Lam, Remi
      - Sanchez-Gonzalez, Alvaro
      - Willson, Matthew
      - et al.
    institution: Science
    doi: 10.1126/science.adi2336
    url: https://www.science.org/doi/10.1126/science.adi2336
  - id: ps-ai-for-climate-science-2
    title: >-
      FourCastNet: A Global Data-driven High-resolution Weather Model using Adaptive Fourier Neural
      Operators
    type: academic_paper
    year: 2022
    authors:
      - Pathak, Jaideep
      - Subramanian, Shashank
      - Harrington, Peter
      - et al.
    institution: arXiv
    url: https://arxiv.org/abs/2202.11214
  - id: ps-ai-for-climate-science-3
    title: Neural general circulation models for weather and climate
    type: journal_article
    year: 2024
    institution: Nature
    doi: 10.1038/s41586-024-07744-y
    url: https://www.nature.com/articles/s41586-024-07744-y
known_gaps:
  - AI-based climate projection for decadal timescales
  - Physical consistency of AI weather models under extreme events
disputed_statements: []
secondary_sources: []
updated: "2026-05-28"
---
## TL;DR
AI for climate and weather science is strongest when described as a set of specific learned forecasting and simulation methods. Public claims should distinguish weather forecasting, climate simulation, and operational climate-risk analysis.

## Core Explanation
Recent systems use machine learning to emulate parts of atmospheric prediction, learn from reanalysis data, or combine neural components with physical models. These tools can be fast and skillful in defined settings, but they do not replace the full scientific and operational workflow around climate observations, uncertainty, and model evaluation.

## Related Articles

- [AI for Weather Forecasting: Data-Driven Numerical Weather Prediction and Nowcasting](../ai-for-weather-forecasting.md)
- [AI for Disaster Prediction: Earthquake Forecasting, Flood Detection, and Early Warning Systems](../ai-disaster-prediction.md)
- [AI for Remote Sensing](../ai-for-remote-sensing.md)