ArticlePDF Available

Climate storylines as a way of bridging the gap between information and decision-making in hydrological risk

DOI:10.1371/journal.pclm.0000270
Authors:
Daniel Caviedes-Voullième at Forschungszentrum Jülich
Daniel Caviedes-Voullième
Ted Shepherd at University of Reading
Ted Shepherd
Download full-text PDF
Read full-text
Download citation

Figures

Figures - available via license: Creative Commons Attribution 4.0 International
Content may be subject to copyright.
Available via license: CC BY 4.0
Content may be subject to copyright.
OPINION
Climate storylines as a way of bridging the gap
between information and decision-making in
hydrological risk
Daniel Caviedes-VoullièmeID
1,2
, Theodore G. ShepherdID
3,4
*
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
*t.shepherd@fz-juelich.de
Physical climate storylines—physically self-consistent unfoldings of past events, or of plausible
future events or pathways [1]—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 [2] 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 [3]. 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 [2].
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 [4]. But science also
rests on explanation, namely the attribution of an effect (whether observed or imagined) to a
set of meaningful causal factors [5]. 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 [4]—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 [6]. 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 [7]. 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
PLOS Climate | https://doi.org/10.1371/journal.pclm.0000270 August 15, 2023 1 / 4
a1111111111
a1111111111
a1111111111
a1111111111
a1111111111
OPEN ACCESS
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.
https://doi.org/10.1371/journal.pclm.0000270
Editor: Jamie Males, PLOS Climate, UNITED
KINGDOM
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
credited.
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
each.
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 [8]. 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
[9]. 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 [10]. 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 [2].
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 [2], adapted from [3].
https://doi.org/10.1371/journal.pclm.0000270.g001
PLOS CLIMATE
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 [11]. 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 [12].
To make scientific information useful for decision-making means crossing the science-pol-
icy boundary. Cash et al. [13] 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 [14]. In so doing they help
make climate information meaningful at the local scale [15].
Author Contributions
Conceptualization: Daniel Caviedes-Voullième, Theodore G. Shepherd.
Writing – original draft: Daniel Caviedes-Voullième, Theodore G. Shepherd.
References
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
PMID: 32747855
PLOS CLIMATE
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:
e2020WR028091.
11. Blo
¨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;
64:1141–58.
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:
http://dx.doi.org/10.2139/ssrn.372280.
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
PLOS Climate | https://doi.org/10.1371/journal.pclm.0000270 August 15, 2023 4 / 4
ResearchGate has not been able to resolve any citations for this publication.
A multi-disciplinary analysis of the exceptional flood event of July 2021 in central Europe – Part 2: Historical context and relation to climate change
Article
Full-text available
Heavy precipitation over western Germany and neighboring countries in July 2021 led to widespread floods, with the Ahr and Erft river catchments being particularly affected. Following the event characterization and process analysis in Part 1, here we put the 2021 event in the historical context regarding precipitation and discharge records and in terms of the temporal transformation of the valley morphology. Furthermore, we evaluated the role of ongoing and future climate change on the modification of rainfall totals and the associated flood hazard, as well as implications for flood management. The event was among the five heaviest precipitation events of the past 70 years in Germany. However, consideration of the large LAERTES-EU regional climate model (RCM) ensemble revealed a substantial underestimation of both return levels and periods based on extreme value statistics using only observations. An analysis of homogeneous hydrological data of the last 70 years demonstrated that the event discharges exceeded by far the statistical 100-year return levels. Nevertheless, the flood peaks at the Ahr river were comparable to the reconstructed major historical events of 1804 and 1910, which were not included in the flood risk assessment so far. A comparison between the 2021 and past events showed differences in terms of the observed hydro-morphodynamic processes which enhanced the flood risk due to changes in the landscape organization and occupation. The role of climate change and how the 2021 event would unfold under warmer or colder conditions (within a −2 to +3 K range) was considered based on both a pseudo global warming (PGW) model experiments and the analysis of an RCM ensemble. The PGW experiments showed that the spatial mean precipitation scales with the theoretical Clausius–Clapeyron (CC) relation, predicting a 7 % to 9 % increase per degree of warming. Using the PGW rainfall simulations as input to a hydrological model of the Ahr river basin revealed a strong and non-linear effect on flood peaks: for the +2 K scenario, the 18 % increase in areal rainfall led to a 39 % increase of the flood peak at gauge Altenahr. The analysis of the high-resolution convection-permitting KIT-KLIWA RCM ensemble confirmed the CC scaling for moderate spatial mean precipitation but showed a super CC scaling of up to 10 % for higher intensities. Moreover, the spatial extent of such precipitation events is also expected to increase.
View
Small is beautiful: climate-change science as if people mattered
Article
Full-text available
  • Mar 2022
There is a widely accepted gap between the production and use of climate information. It is also widely accepted that at least part of the reason for this situation lies in the challenge of bridging between what may be characterized as ‘‘top-down’’ approaches to climate information on the global scale, and local decision contexts, which necessarily take a ‘‘bottom-up’’ perspective, in which climate change is just one factor among many to consider. We here reflect on the insights provided in a different context—that of economics—by E.F. Schumacher in his celebrated book Small is Beautiful (1973), to see what light they might shed on this challenge, with a focus on climate-change science for adaptation. Schumacher asked how economics might look if it was structured “as if people mattered”. We ask the same question of climate-change science, and find many parallels. One is the need to grapple with the complexity of local situations, which can be addressed by expressing climate knowledge in a conditional form. A second is the importance of simplicity when dealing with deep uncertainty, which can be addressed through the use of physical climate storylines. A third is the need to empower local communities to make sense of their own situation, which can be addressed by developing ‘‘intermediate technologies’’ that build trust and transparency. Much of climate-change science is necessarily big science. We argue that in order to make climate information useable for adaptation, it is also necessary to discover the beauty of smallness.
View
Meaningful climate science
Article
Full-text available
Within the climate science community, useable climate science has been understood as quantitative, usually as a best estimate together with a quantified uncertainty. Physical scientists are trained to produce numbers and to draw general, abstract conclusions. In general, however, people relate much better to stories and to events they have experienced, which are inevitably contingent and particular. Sheila Jasanoff has argued elsewhere that the process of abstraction in climate science “detaches knowledge from meaning”. Perhaps useable climate science is, then, meaningful climate science. We argue here that the development of meaningful climate science can be achieved by adopting a storyline approach to climate variability and change. By ‘storyline’ we mean a physically self-consistent unfolding of past events or of plausible future events or pathways. Storylines represent a combination of qualitative and quantitative information, where the qualitative element represents a packaging or contextualization of the quantitative aspects, which ensures that data can be meaningfully interpreted. Viewed from this perspective, we show that physical climate storylines can be aligned with several well-established vehicles for translation of knowledge between diverse communities: narratives, boundary objects, and data journeys. They can therefore be used as a ‘pidgin language’ to enrich the set of tools available to climate scientists to bring meaning to climate knowledge. “And what is the use of a book”, thought Alice, “without pictures or conversations?” (Lewis Carroll, Alice’s Adventures in Wonderland).
View
What Role Does Hydrological Science Play in the Age of Machine Learning?
Article
Full-text available
This paper is derived from a keynote talk given at the Google's 2020 Flood Forecasting Meets Machine Learning Workshop. Recent experiments applying deep learning to rainfall‐runoff simulation indicate that there is significantly more information in large‐scale hydrological data sets than hydrologists have been able to translate into theory or models. While there is a growing interest in machine learning in the hydrological sciences community, in many ways, our community still holds deeply subjective and nonevidence‐based preferences for models based on a certain type of “process understanding” that has historically not translated into accurate theory, models, or predictions. This commentary is a call to action for the hydrology community to focus on developing a quantitative understanding of where and when hydrological process understanding is valuable in a modeling discipline increasingly dominated by machine learning. We offer some potential perspectives and preliminary examples about how this might be accomplished.
View
Twenty-three Unsolved Problems in Hydrology (UPH) – a community perspective
Article
Full-text available
  • Jun 2019
This paper is the outcome of a community initiative to identify major unsolved scientific problems in hydrology motivated by a need for stronger harmonisation of research efforts. The procedure involved a public consultation through on-line media, followed by two workshops through which a large number of potential science questions were collated, prioritised, and synthesised. In spite of the diversity of the participants (230 scientists in total), the process revealed much about community priorities and the state of our science: a preference for continuity in research questions rather than radical departures or redirections from past and current work. Questions remain focussed on process-based understanding of hydrological variability and causality at all space and time scales. Increased attention to environmental change drives a new emphasis on understanding how change propagates across interfaces within the hydrological system and across disciplinary boundaries. In particular, the expansion of the human footprint raises a new set of questions related to human interactions with nature and water cycle feedbacks in the context of complex water management problems. We hope that this reflection and synthesis of the 23 unsolved problems in hydrology will help guide research efforts for some years to come.
View
Storyline approach to the construction of regional climate change information
Article
Full-text available
  • May 2019
Climate science seeks to make statements of confidence about what has happened, and what will happen (conditional on scenario). The approach is effective for the global, thermodynamic aspects of climate change, but is ineffective when it comes to aspects of climate change related to atmospheric circulation, which are highly uncertain. Yet, atmospheric circulation strongly mediates climate impacts at the regional scale. In this way, the confidence framework, which focuses on avoiding type 1 errors (false alarms), raises the prospect of committing type 2 errors (missed warnings). This has ethical implications. At the regional scale, however, where information on climate change has to be combined with many other factors affecting vulnerability and exposure-most of which are highly uncertain-the societally relevant question is not 'What will happen?' but rather 'What is the impact of particular actions under an uncertain regional climate change?' This reframing of the question can cut the Gordian knot of regional climate change information, provided one distinguishes between epistemic and aleatoric uncertainties-something that is generally not done in climate projections. It is argued that the storyline approach to climate change-the identification of physically self-consistent, plausible pathways-has the potential to accomplish precisely this.
View
Storylines: an alternative approach to representing uncertainty in physical aspects of climate change
Article
Full-text available
As climate change research becomes increasingly applied, the need for actionable information is growing rapidly. A key aspect of this requirement is the representation of uncertainties. The conventional approach to representing uncertainty in physical aspects of climate change is probabilistic, based on ensembles of climate model simulations. In the face of deep uncertainties, the known limitations of this approach are becoming increasingly apparent. An alternative is thus emerging which may be called a ‘storyline’ approach. We define a storyline as a physically self-consistent unfolding of past events, or of plausible future events or pathways. No a priori probability of the storyline is assessed; emphasis is placed instead on understanding the driving factors involved, and the plausibility of those factors. We introduce a typology of four reasons for using storylines to represent uncertainty in physical aspects of climate change: (i) improving risk awareness by framing risk in an event-oriented rather than a probabilistic manner, which corresponds more directly to how people perceive and respond to risk; (ii) strengthening decision-making by allowing one to work backward from a particular vulnerability or decision point, combining climate change information with other relevant factors to address compound risk and develop appropriate stress tests; (iii) providing a physical basis for partitioning uncertainty, thereby allowing the use of more credible regional models in a conditioned manner and (iv) exploring the boundaries of plausibility, thereby guarding against false precision and surprise. Storylines also offer a powerful way of linking physical with human aspects of climate change.
View
From engineering hydrology to Earth system science: Milestones in the transformation of hydrologic science
Article
Full-text available
Hydrology has undergone almost transformative changes over the past 50 years. Huge strides have been made in the transition from early empirical approaches to rigorous approaches based on the fluid mechanics of water movement on and below the land surface. However, progress has been hampered by problems posed by the presence of heterogeneity, including subsurface heterogeneity present at all scales. The inability to measure or map the heterogeneity everywhere prevented the development of balance equations and associated closure relations at the scales of interest, and has led to the virtual impasse we are presently in, in terms of development of physically based models needed for hydrologic predictions. An alternative to the mapping of heterogeneity everywhere is a new Earth system science view, which sees the heterogeneity as the end result of co-evolutionary hydrological, geomorphological, ecological, and pedological processes, each operating at a different rate, which help to shape the landscapes that we find in nature, including the heterogeneity that we do not readily see. The expectation is that instead of specifying exact details of the heterogeneity in our models, we can replace it (without loss of information) with the ecosystem function that they perform. Guided by this new Earth system science perspective, development of hydrologic science is now addressing new questions using novel holistic co-evolutionary approaches as opposed to the physical, fluid mechanics based reductionist approaches that we inherited from the recent past. In the emergent Anthropocene, the co-evolutionary view has expanded further to involve interactions and feedbacks with human-social processes as well. In this paper, I present my own perspective of key milestones in the transformation of hydrologic science from engineering hydrology to Earth system science, drawn from the work of several students and colleagues of mine, and discuss their implication for hydrologic observations, theory development, and predictions.
View
Ambitious partnership needed for reliable climate prediction
Article
  • Jun 2022
Current global climate models struggle to represent precipitation and related extreme events, with serious implications for the physical evidence base to support climate actions. A leap to kilometre-scale models could overcome this shortcoming but requires collaboration on an unprecedented scale.
View
Deep learning and process understanding for data-driven Earth system science
Article
Machine learning approaches are increasingly used to extract patterns and insights from the ever-increasing stream of geospatial data, but current approaches may not be optimal when system behaviour is dominated by spatial or temporal context. Here, rather than amending classical machine learning, we argue that these contextual cues should be used as part of deep learning (an approach that is able to extract spatio-temporal features automatically) to gain further process understanding of Earth system science problems, improving the predictive ability of seasonal forecasting and modelling of long-range spatial connections across multiple timescales, for example. The next step will be a hybrid modelling approach, coupling physical process models with the versatility of data-driven machine learning.
View