Probabilistic storyline attribution using machine learning
Frieder Loer, Maybritt Schillinger, Sebastian Sippel · 2026
Read Probabilistic storyline attribution using machine learning by Frieder Loer, Maybritt Schillinger, Sebastian Sippel (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2606-02550.txt.
Description
Loer and co-authors attribute heat extremes to climate change with a learning-based storyline method. Storyline attribution holds the air flow pattern fixed. It asks how the same event plays out under other climate states. Climate models cannot move such counterfactuals across forcing states directly. The team uses distributional autoencoders instead.
The autoencoders model full distributions of European temperature fields. Inputs are the circulation state and the global warming level. Conditional probability ratios follow from the distributions. Training uses coupled climate model simulations. Tests span factual and counterfactual simulation sets. A case study revisits the 2003 European heatwave. With 2003 air flow held fixed, peak intensity rises in the 2028 and 2053 versions. Probability ratios reach 2.1 and 3.2 against 2003. The method gives storyline studies a statistical engine.
Catalog record
| CREATOR | Frieder Loer, Maybritt Schillinger, Sebastian Sippel |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 62 KB |
| LICENSE | http://creativecommons.org/licenses/by-sa/4.0/ |
| RIGHTS | CC BY SA |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2606.02550