# Missing the Butterfly and Predicting the Past: Features or Bugs of Accurate AI Weather Models?

arXiv: https://arxiv.org/abs/2608.25835 · HTML: https://mlsanity.com/paper/2608.25835/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 25 of its 108 references that are in the map.
- First posted on arXiv 26 Aug 2026.

## Builds on (most cited first, 25 of 25)

- [Scalable Diffusion Models with Transformers](https://mlsanity.com/paper/2212.09748/) (2022)
- [Optuna: A Next-generation Hyperparameter Optimization Framework](https://mlsanity.com/paper/1907.10902/) (2019)
- [Reconciling modern machine learning practice and the bias-variance trade-off](https://mlsanity.com/paper/1812.11118/) (2018)
- [FourCastNet: A Global Data-driven High-resolution Weather Model using Adaptive Fourier Neural Operators](https://mlsanity.com/paper/2202.11214/) (2022)
- [Hyperband: A Novel Bandit-Based Approach to Hyperparameter Optimization](https://mlsanity.com/paper/1603.06560/) (2016)
- [Physics-Informed Neural Operator for Learning Partial Differential Equations](https://mlsanity.com/paper/2111.03794/) (2021)
- [Neural General Circulation Models for Weather and Climate](https://mlsanity.com/paper/2311.07222/) (2023)
- [WeatherBench 2: A benchmark for the next generation of data-driven global weather models](https://mlsanity.com/paper/2308.15560/) (2023)
- [Spherical Fourier Neural Operators: Learning Stable Dynamics on the Sphere](https://mlsanity.com/paper/2306.03838/) (2023)
- [Why and When Can Deep -- but Not Shallow -- Networks Avoid the Curse of Dimensionality: a Review](https://mlsanity.com/paper/1611.00740/) (2016)
- [ClimODE: Climate and Weather Forecasting with Physics-informed Neural ODEs](https://mlsanity.com/paper/2404.10024/) (2024)
- [ACE2: Accurately learning subseasonal to decadal atmospheric variability and forced responses](https://mlsanity.com/paper/2411.11268/) (2024)
- [Challenges of learning multi-scale dynamics with AI weather models: Implications for stability and one solution](https://mlsanity.com/paper/2304.07029/) (2023)
- [Can AI weather models predict out-of-distribution gray swan tropical cyclones?](https://mlsanity.com/paper/2410.14932/) (2024)
- [Dynamical Tests of a Deep-Learning Weather Prediction Model](https://mlsanity.com/paper/2309.10867/) (2023)
- [Scaling Laws of Global Weather Models](https://mlsanity.com/paper/2602.22962/) (2026)
- [Probing forced responses and causality in data-driven climate emulators: conceptual limitations and the role of reduced-order models](https://mlsanity.com/paper/2506.22552/) (2025)
- [Atmospheric Predictability Beyond 30 Days with Machine Learning](https://mlsanity.com/paper/2504.20238/) (2025)
- [Using local dynamics to explain analog forecasting of chaotic systems](https://mlsanity.com/paper/2007.14216/) (2020)
- [AI-boosted rare event sampling to characterize extreme weather](https://mlsanity.com/paper/2510.27066/) (2025)
- [Decision-oriented benchmarking to transform AI weather forecast access: Application to the Indian monsoon](https://mlsanity.com/paper/2602.03767/) (2026)
- [Rigorous uncertainty quantification of probabilistic AI weather forecasts with conformal prediction](https://mlsanity.com/paper/2606.19642/) (2026)
- [Designing probabilistic AI monsoon forecasts to inform agricultural decision-making](https://mlsanity.com/paper/2603.07893/) (2026)
- [On Neural Scaling Laws for Weather Emulation through Continual Training](https://mlsanity.com/paper/2603.25687/) (2026)
- [The physics of AI weather models](https://mlsanity.com/paper/2605.23778/) (2026)

## Abstract (arXiv)

AI weather prediction (AIWP) models rival physics-based models, yet the sources of their unexpected forecast accuracy and the degree of their physical fidelity remain unclear. Here, across a hierarchy spanning observation-based reanalysis, a general circulation model, and the multi-scale Lorenz system, we show that AI models can be trained to skillfully predict the past (backcast), though backcasts are systematically less accurate than forecasts. However, skillful backcasting appears to violate the second law of thermodynamics, and all these forecasting and backcasting models miss the butterfly effect. We trace the surprising forecast accuracy, missing butterfly, and skillful backcasting to a single cause: inevitable coarse-graining of training data, which removes fast, small scales and/or some variables. From the Lorenz system to official Pangu-Weather models, reducing coarse-graining makes AI predictions more physics-like (arrow of time and butterfly-like effects emerge), but forecast accuracy declines. Results offer an explanation for AIWP models' forecast skill: unlike physics-based models, they implicitly learn how fast, small scales affect large scales without inheriting their rapid error growth. Broader implications are that AI models' proliferation calls for revisiting predictability theories and long-term climate emulation strategies, and backcasting offers a useful, new lens for such analyses.