A brand new paper by Frauke Feser and Ted Shepherd has just been published:
“The concepts of spectrally nudged storylines for extreme event attribution” (Nature Communications Earth & Environment, 2025)
I really enjoyed reading it. The paper is clearly explaining a concept that can often feel quite technical. It covers:
- What spectrally nudged storylines (SN storylines) actually are
- Their purpose and scientific foundation
- The difference between storylines and scenarios
- How SN storylines relate to other conditional attribution methods
- The way forward, how SN storylines can be used in the future
All of this is presented with scientific depth, but in a succinct and accessible way.
Why it matters
Extreme event attribution (EEA) is about understanding the role of climate change in today’s extreme weather events. Different methods exist, each with their strengths and limitations. Spectrally nudged storylines are particularly powerful because they allow us to explore how the same event would unfold in different climate states.
The method works by constraining atmospheric dynamics (using spectral nudging) while letting thermodynamics evolve freely. In practice, this means:
- You can isolate the thermodynamic effects of climate change (e.g. warming, moisture changes).
- You can keep the meteorological “story” of the event intact, which makes results easy to interpret and communicate.
- It is especially suited for studying extremes like heatwaves, droughts, heavy rainfall, and tropical storms.
The Way Forward
What excites me most is not just what spectrally nudged storylines can already do, but where the field is heading. The concepts outlined by Feser & Shepherd fit into a much bigger picture of how attribution science is evolving, here a small selection:
From hazard to impact attribution.
Until now, most attribution work has focused on the meteorological hazard (e.g. heat, rainfall). The next step is linking directly to impacts — how those hazards translate into damages, health effects, or ecosystem responses. Storyline methods, because of their clarity, are well suited for this bridge.
Bridging storylines and probabilistic approaches
The attribution community has sometimes been divided between conditional storylines and unconditioned probabilistic frameworks. In reality, these methods are complementary. By applying statistical analysis on a collection of events in the SN storylines, a bridge between the two methods can be build.
The role of AI
New machine learning and AI tools can support attribution by uncovering nonlinear relationships and hidden drivers of extreme weather events. Projects such as CLINT and XAIDA are already exploring how AI can enhance climate storylines, improve uncertainty quantification, and accelerate analysis.
Spectrally nudged storylines are therefore not just another method in the toolbox — they are a key stepping stone toward a more integrated and collaborative future in attribution science.
Linda van Garderen 