Butterfly Effect Confirmed in Global AI Weather Models: Evidence from Tropical Cyclone Forecasting
Abstract
A paradox recently emerged in artificial intelligence (AI) weather prediction research. While some claim AI weather models cannot simulate atmospheric butterfly effect, this conflicts with AI models' limited predictability and advances in AI ensemble forecasting. This study demonstrates via counterexamples that the butterfly effect does exist in AI weather predictions. For Super Typhoon Khanun, AI predictions are constrained by a double-attractor system. Minor initial perturbations confined to two regions trigger state transitions between two local attractors, causing a 1006-km difference in the predicted storm position on Day 7. This behavior is consistent with numerical weather prediction models and observed in ~12% of tropical cyclones in the past 5 years. These findings verify AI's ability to capture atmospheric chaos and provide the physical basis for AI ensemble forecasting.
Related
- Butterfly Effect and the Kinetic Energy Cascade in Probabilistic Machine Learning Weather Prediction Models
- Skillful Seasonal Tropical Cyclone Prediction with AI Weather--Climate Models Forced by Dynamically Predicted SSTs
- Tropical Cyclone Attributes in Sparse Autoencoder Features of an AI Weather Model
- SDECast: Probabilistic Weather Forecasting in Continuous Time with Neural SDEs
- Swift: an autoregressive consistency model for efficient weather forecasting
BibTeX
@misc{leung2026butterfly,
title = {{Butterfly Effect Confirmed in Global AI Weather Models: Evidence from Tropical Cyclone Forecasting}},
author = {Jeremy Cheuk-Hin Leung and Daosheng Xu and Weiye Yu and Shaojing Zhang and Xiaodong Zeng and Gaozhen Nie and Jie Feng and Jingchen Pu and Yi Li and Kaijun Ren and Qingcun Zeng and Banglin Zhang},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.39379},
eprint = {2609.39379},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.39379},
note = {CC BY-NC-SA 4.0},
}
RIS
TY - UNPB
TI - Butterfly Effect Confirmed in Global AI Weather Models: Evidence from Tropical Cyclone Forecasting
AU - Jeremy Cheuk-Hin Leung
AU - Daosheng Xu
AU - Weiye Yu
AU - Shaojing Zhang
AU - Xiaodong Zeng
AU - Gaozhen Nie
AU - Jie Feng
AU - Jingchen Pu
AU - Yi Li
AU - Kaijun Ren
AU - Qingcun Zeng
AU - Banglin Zhang
PY - 2026
DA - 2026/09/30
PB - arXiv
DO - 10.48550/arxiv.2609.39379
UR - https://arxiv.org/abs/2609.39379
AB - A paradox recently emerged in artificial intelligence (AI) weather prediction research. While some claim AI weather models cannot simulate atmospheric butterfly effect, this conflicts with AI models' limited predictability and advances in AI ensemble forecasting. This study demonstrates via counterexamples that the butterfly effect does exist in AI weather predictions. For Super Typhoon Khanun, AI predictions are constrained by a double-attractor system. Minor initial perturbations confined to two regions trigger state transitions between two local attractors, causing a 1006-km difference in the predicted storm position on Day 7. This behavior is consistent with numerical weather prediction models and observed in ~12% of tropical cyclones in the past 5 years. These findings verify AI's ability to capture atmospheric chaos and provide the physical basis for AI ensemble forecasting.
ER -