Climate storylines as a way of bridging the gap
between information and decision-making in
, Theodore G. ShepherdID
1Simulation and Data Lab Terrestrial Systems, Ju¨lich Supercomputing Centre, Forschungszentrum Ju¨lich,
Ju¨lich, Germany, 2Institute of Bio- and Geosciences: Agrosphere (IBG-3), Forschungszentrum Ju¨lich, Ju¨lich,
Germany, 3High Performance Computing in Applied Science and Engineering, Ju¨lich Supercomputing
Centre, Forschungszentrum Ju¨lich, Ju¨lich, Germany, 4Department of Meteorology, University of Reading,
Reading, United Kingdom
Physical climate storylines—physically self-consistent unfoldings of past events, or of plausible
future events or pathways —have recently emerged as a way of navigating the cascade of
uncertainty that arises when considering the impacts of climate change. The Working Group I
contribution to the Sixth Assessment Report of the Intergovernmental Panel on Climate
Change  has adopted this definition within its Glossary, and the concept is illustrated in
Fig 1. Given the high levels of uncertainty concerning the climate response of remote drivers
of regional change such as sea-surface temperature patterns, and of the dynamical conditions
leading to extreme events, any attempt to aggregate over that uncertainty inevitably leads to
general or weak statements . By conditioning on those uncertain aspects of the climate
response, storylines provide spatially and temporally coherent scenarios at the regional or local
scale. Storylines represent the overall uncertainty in a discrete manner, and are particularly
useful for exploring low-likelihood, high-impact outcomes .
The acid test of any science is generally understood to be successful prediction. For hydro-
logical risk, however, the combination of deep uncertainty in the climate response at the local
scale together with the non-stationarity of a changing climate challenges the kind of objective
probabilistic quantification that underpins any notion of predictability . But science also
rests on explanation, namely the attribution of an effect (whether observed or imagined) to a
set of meaningful causal factors . This is quite different from prediction, but relates directly
to decision-making, where the key concern is not uncertainty but rather the strength of evi-
dence behind various competing explanations —often including worst-case scenarios—and
the causality of those explanations is required to inform appropriate action. Due to its deter-
ministic representation of physical processes, physical modelling can provide explanations
together with deterministic, conditional quantification in the form of storylines.
Physical modelling has long been the cornerstone of explanation in physical climate sci-
ence, but as mentioned earlier, major systematic uncertainties remain. With the rapid growth
in the use of Artificial Intelligence/Machine Learning (AI/ML) tools across all areas of science,
there is a move away from physical modelling towards data-driven methods to assess climate
risk . At the same time, many climate scientists are pushing for km-scale physical modelling
to overcome the systematic model errors associated with the representation of atmospheric
convection . Although AI/ML has definite value in detecting patterns of change, it is inher-
ently based on statistical prediction of those patterns, rather than physically-based explanation.
PLOS Climate | https://doi.org/10.1371/journal.pclm.0000270 August 15, 2023 1 / 4
Citation: Caviedes-Voullième D, Shepherd TG
(2023) Climate storylines as a way of bridging the
gap between information and decision-making in
hydrological risk. PLOS Clim 2(8): e0000270.
Editor: Jamie Males, PLOS Climate, UNITED
Published: August 15, 2023
Copyright: ©2023 Caviedes-Voullième, Shepherd.
This is an open access article distributed under the
terms of the Creative Commons Attribution
License, which permits unrestricted use,
distribution, and reproduction in any medium,
provided the original author and source are
Funding: The authors received no specific funding
for this work.
Competing interests: The authors have declared
that no competing interests exist.
And while km-scale physical modelling would be transformative, systematic uncertainties will
surely remain and the simulated sample sizes will inevitably be small. Storylines could be useful
in serving as a bridge between these two divergent approaches, combining the strengths of
In hydrological science, the classical modelling approach has been based on highly parame-
terised models, often conceptual and process-based, but not really physical. Historically,
hydrological modelling has been concerned with quantitatively reproducing observable signa-
tures (e.g. hydrographs) in order to support the predictive power of the models . This how-
ever does not guarantee their explanatory power, fundamental for their reliability and
robustness in a changing environment. Moreover, the entire approach does not allow a fine-
grain process interrogation of the dynamics. Physical models are now becoming more widely
used, thanks to the evolution of computing capacity and remotely sensed spatial information
. These models provide explicitly resolved spatio-temporal information and causal explana-
tions. AI/ML tools have become prominent in hydrology too, e.g. to mine information to a
new level out of hydrological observations . As with physical climate science, storylines
can be used to bridge between these different sources of information.
Storylines can also be used to bridge between climate science and hydrological science for
understanding hydrological risk. IPCC Working Group I is now heavily using the concept of
Climatic Impact-Drivers, which are predictors of hydrological extremes such as floods .
Fig 1. Schematic of two types of physical climate storylines with a particular climate impact of concern (red). The storylines are defined by specified elements
(dark blue). Variable elements (light blue) are simulated conditional on the specified elements. The white elements are ‘blocked’ since their state does not need
to be known to determine the light blue elements. Other types of storylines could be defined by specifying other elements (e.g. storylines of different climate
sensitivities or different representative concentration pathways). (a) Event storyline, where the particular dynamical conditions during the event as well as the
regional warming are specified and control the hazard arising from the event. (b) Dynamical storyline, where the global warming level and remote drivers are
specified and control the long-term changes in atmospheric dynamics and regional warming. In both storylines, the impact is also conditioned on specified
exposure and vulnerability. From Box 10.2 of , adapted from .
PLOS Climate | https://doi.org/10.1371/journal.pclm.0000270 August 15, 2023 2 / 4
While these provide a useful first guess, storylines could be used to provide explainable and
hence actionable information from deterministic physically-based hydrological models, driven
with meaningful hydrometeorological events selected from counterfactual analysis, possibly
based on patterns identified via conceptual and data-based models. We argue that storylines
can provide a framework to adapt and prepare for extreme hydrological events, by supporting
the understanding of risk causality (explanatory power) including local conditions, and con-
textualising (into actionable information) the plausible risks triggered by extreme events not
well captured by probabilistic representations . Moreover, storylines incorporating physi-
cally-based simulation can enrich the local impact assessment of rare extreme events by assimi-
lating events which have occurred elsewhere, but for which the conditions are plausible in the
place of interest due to changing climate .
To make scientific information useful for decision-making means crossing the science-pol-
icy boundary. Cash et al.  suggested three requirements for this: salience, credibility, and
legitimacy of the information. They also emphasized that the difficulty primarily lies in the fact
that the actors on different sides of the boundary perceive and value these three attributes
quite differently. By providing conditional causal explanations of observed or imagined events
at a fine-grained scale, which can be directly connected to observations and impacts and can
be used to construct counter-factual events representing policy options, hydrological storylines
grounded in physically-based modelling have the potential to provide a ‘boundary object’ that
meets these requirements for both scientists and policy-makers . In so doing they help
make climate information meaningful at the local scale .
Conceptualization: Daniel Caviedes-Voullième, Theodore G. Shepherd.
Writing – original draft: Daniel Caviedes-Voullième, Theodore G. Shepherd.
1. Shepherd TG, Boyd E, Calel RA, Chapman SC, Dessai S, Dima-West IM, et al. Storylines: an alterna-
tive approach to representing uncertainty in physical aspects of climate change. Climatic Change.
2018; 151:555–71. https://doi.org/10.1007/s10584-018-2317-9 PMID: 30880852
2. IPCC (Intergovernmental Panel on Climate Change). Climate Change 2021: The Physical Science
Basis. Contribution of Working Group I to the Sixth Assessment Report of the Inter-governmental Panel
on Climate Change. Cambridge, UK and New York, NY, USA: 2021.
3. Shepherd TG. Storyline approach to the construction of regional climate change information. Proceed-
ings of the Royal Society A: Mathematical, Physical and Engineering Sciences. 2019; 475:20190013.
4. Beven K. Facets of uncertainty: epistemic uncertainty, non-stationarity, likelihood, hypothesis testing,
and communication. Hydrological Sciences Journal. 2016; 61:1652–65.
5. Halpern JY, Pearl J. Causes and explanations: a structural-model approach. Part II: Explanations. The
British Journal for the Philosophy of Science. 2005; 56:889–911.
6. Reichstein M, Camps-Valls G, Stevens B, Jung M, Denzler J, Carvalhais N, et al. Deep learning and
process understanding for data-driven Earth system science. Nature. 2019; 566:195–204. https://doi.
org/10.1038/s41586-019-0912-1 PMID: 30760912
7. Slingo J, Bates P, Bauer P, Belcher S, Palmer T, Stephens G, et al. Ambitious partnership needed for
reliable climate prediction. Nature Climate Change. 2022; 12:499–503.
8. Sivapalan M. From engineering hydrology to Earth system science: milestones in the transformation of
hydrologic science. Hydrology and Earth System Sciences. 2018; 22:1665–93.
9. Clark MP, Bierkens MFP, Samaniego L, Woods RA, Uijlenhoet R, Bennett KE, et al. The evolution of
process-based hydrologic models: historical challenges and the collective quest for physical realism.
Hydrology and Earth System Sciences. 2017; 21:3427–40. https://doi.org/10.5194/hess-21-3427-2017
PLOS Climate | https://doi.org/10.1371/journal.pclm.0000270 August 15, 2023 3 / 4
10. Nearing GS, Kratzert F, Keefe Sampson A, Pelissier CS, Klotz D, Frame JM, et al. What role does
hydrological science play in the age of machine learning? Water Resources Research. 2021; 57:
¨schl G, Bierkens MFP, Chambel A, Cudennec C, Destouni G, Fiori A, et al. Twenty-three unsolved
problems in hydrology (UPH)–a community perspective. Hydrological Sciences Journal. 2019;
12. Ludwig P, Ehmele F, Franca MJ, Mohr S, Caldas-Alvarez A, Daniell JE, et al. A multi-disciplinary analy-
sis of the exceptional flood event of July 2021 in central Europe–Part 2: Historical context and relation to
climate change. Natural Hazards and Earth System Sciences. 2023; 23:1287–1311.
13. Cash D, Clark WC, Alcock F, Dickson NM, Eckley N, Ja
¨ger J. Salience, credibility, legitimacy and
boundaries: Linking research, assessment and decision making (November 2002). Available at SSRN:
14. Shepherd TG, Lloyd EA. Meaningful climate science. Climatic Change. 2021; 169:17.
15. Rodrigues RR, Shepherd TG. Small is Beautiful: Climate-change science as if people mattered. PNAS
Nexus. 2022; 1:pgac009. https://doi.org/10.1093/pnasnexus/pgac009 PMID: 36712809
PLOS Climate | https://doi.org/10.1371/journal.pclm.0000270 August 15, 2023 4 / 4