Quantitative Storytelling: Science, Narratives, and Uncertainty in Nexus Innovations

Quantitative Storytelling: Science, narratives and uncertainty in nexus innovations Louisa Jane Di Felice a; Violeta Cabello b; Maddalena Ripa c d; Cristina Madrid a a Institute of Environmental Science and Technology, Universitat Autònoma de Barcelona, 08193 Bellaterra, Spain b Science Collab, Basque Center for Climate Change c Autonomous University of Barcelona, Barcelona, Spain d Italian Ministry of Education, University and Research (MIUR), Italy Abstract Innovations are central instruments of sustainability policies. They project future visions onto technological solutions, and enable win-win framings of complex sustainability issues. Yet, they also create new problems by interconnecting different resources such as water, food and energy, what is known as the “WEF nexus.” In this paper, we apply a new approach called Quantitative Storytelling (QST) to the assessment of four innovations with a strong nexus component in EU policy: biofuels, shale gas, electric vehicles, and alternative water resources. Recognizing irreducible pluralism and uncertainties, QST inspects the relationships between the narratives used to frame sustainability issues and the evidence on those issues. Our experiences outlined two rationales for implementing QST. First, QST can be used to question dominant narratives that promote certain innovations despite evidence against their effectiveness. Second, QST can offer avenues for pluralistic processes of cocreation of alternative narratives and imaginaries. We reflect on the implementation of QST and on the role played by different uncertainties throughout these processes. Our experiences suggest that while the role of nexus assessments using both numbers and narratives may not be instrumental in directly inducing policy change, they are valuable means to open discussions on innovations outside of dominant nexus imaginaries. Keywords: narratives, innovations, WEF nexus, complexity, uncertainty Published version in Science, Technology and Human Values: https://doi.org/10.1177/ 1 Introduction The analysis of sustainability challenges in complex social-ecological systems calls for the integration of multiple analytical methods, frames, and perspectives. The EU Horizon 2020 project MAGIC, short for Moving Towards Adaptive Governance in Complexity: Informing Nexus Security, investigates the relationships between nexus elements of social-ecological systems, such as water, energy and food, and the narratives that are used to frame those relationships and systems (Ripa and Giampietro 2017). To do so, MAGIC combines quantitative analyses, qualitative analyses, and the engagement of different actors, through a transdisciplinary approach called Quantitative Storytelling (QST). QST was proposed to address gaps in the science for governance literature with respect to complexity and uncertainty as ground principles for knowledge co-production (Saltelli and Giampietro 2017). By recognizing the irreducible plurality of complex reality and the inevitable existence of uncertainties, QST aims to check the coherence between the narratives used to frame sustainability issues tied to nexus dimensions and the evidence on those issues, where the evidence can be already existing, or newly produced by QST researchers. In this paper, we reflect on our shared experiences of using the QST approach in assessing the narratives underpinning innovations for sustainability, focusing on four innovations that have a strong nexus component: alternative water resources, biofuels, shale gas and electric vehicles. The concept of the water–energy–food nexus was popularized following the Bonn 2011 Nexus Conference, in response to climate change and social changes including population growth, globalization, economic growth and urbanization (Hoff 2011). A nexus lens emphasizes the complex relationships between water, energy, and food (Harwood 2018). Grounded in a holistic, systems perspective, critical nexus assessments can highlight the inherent trade-offs between sustainability dimensions, showing how decisions taken in one sector of the nexus may have direct or indirect consequences on other sectors (van Gevelt 2020). However, discourses of future pathways can also take 2 the form of nexus imaginaries. In this case, a specific solution, often in the form of innovation, is expected to simultaneously solve problems across multiple nexus dimensions (Cairns and Krzywoszynska 2016; Pfotenhauer and Jasanoff 2017). These win-win framings assume that consequences of complex policy decisions can be mapped onto binary outcomes, generating either winners or losers (Jasanoff 2007). Focusing on the science-policy-society interface, (Urbinatti et al. 2020) introduce the “nexus of humility” framework. Humility is associated with irreducible uncertainty and epistemic pluralism. In this view, the nexus shifts from being a problem to be solved, to being a way of viewing problems. Following this view, in this paper and in the QST process developed within the MAGIC project we open our nexus focus from an object-oriented one (i.e., the nexus as a set of material relationships between water, energy and food) to a relational one (i.e., the nexus as an entangled set of relationships between different observers and complex reality). This view includes different actors, values, and imaginaries within nexus relationships. In this paper, we focus on nexus innovations in the European policy context. Since innovations are future-oriented, actors cannot predict-then-act from an evidence base (Borup et al. 2006). Faced with complex issues and incomplete knowledge, they must rely on narratives, expectations, and imaginaries to avoid paralysis. Our definition of narrative is conceptually grounded in complexity (Zellmer, Allen, and Kesseboehmer 2006) and in postnormal science (Funtowicz and Ravetz 1993). Complexity arises from systemobserver relationships: representations of complex situations cannot be compressed into a single, self-consistent view. This entails impredicativity, with events unfolding and simultaneously affecting one another at different scales. As such, there is no unique direction of causality in complex systems (Mitchell 2002). For example, innovations affect and are affected by behaviors, like in the chicken-egg paradox. To allow for decision-making in impredicative situations, narratives are used to simplify reality by fitting it within a chosen (causal) storyline. When narratives are projected onto the future, they become expectations, i.e., “future-oriented abstractions,” smoothing uncertainties to provide a shared vision that different actors can act upon (Borup et al. 2006). Narratives and expectations are constructed within sociotechnical imaginaries, i.e., collectively held social beliefs and values defining what is 3 desirable (Jasanoff and Kim 2015). Impredicativity is also present in the narrative-imaginary relationship: on one hand, narratives are grounded in imaginaries; on the other, they can reinforce those imaginaries, or create and perform new ones. As with the chicken-egg paradox, it is impossible to say whether a narrative or an imaginary came first, as they co-create one another across different timescales. Imaginaries help explain why, out of the universe of possibilities, some visions of techno-scientific and social order win support over others, reducing the indeterminacy of the future. Actors may choose to view complex sustainability challenges using different narratives, grounded in their own imaginaries and expectations. This generates a plurality of voices and values (Jasanoff 2007; Allen and Giampietro 2006). To handle this plurality, scientific knowledge must actively deal with the uncertainty associated with the existence of diverse narratives. By recognizing multiple (and often contradictory) legitimate perspectives, the attention of QST is placed on how innovations are framed, what evidence is used to support them, and how evidence and narratives interact. Throughout the process, the role played by different types of uncertainties in science-policy relations is observed. In the rest of this paper, we present a meta-analysis of four case studies carried out within the MAGIC project, looking at the narratives behind alternative water resources, electric vehicles, biofuels, and shale gas within European contexts. Each case study was developed by applying the QST approach to discuss the role played by those innovations in solving complex sustainability problems. Details on individual case studies can be found in Cabello et al. (2020, 2021) for alternative water resources; Di Felice et al. (2020) and Di Felice, Renner, and Giampietro (2021) for electric vehicles; Cadillo-Benalcazar et al. (2020) and Ripa, Cadillo-Benalcazar, and Giampietro (2021) for biofuels and Madrid Lopez (2020) for shale gas. By exploring the commonalities and differences of the case studies, we re-frame QST as a tool to navigate and inspect science-policy relations, using our experiences to reflect on the interactions between scientific assessments, plural narratives, and uncertainty in the governance of nexus innovations. 4 2 Quantitative Storytelling, Narratives & Uncertainty QST was developed within MAGIC as a common methodological ground for transdisciplinary nexus research (Matthews et al. 2017). It builds and expands on the accounting scheme known as MuSIASEM (Multi-Scale Analysis of Societal and Ecological Metabolism), developed prior to QST for quantitative accounting of the nexus across scales (Giampietro, Mayumi, and Ramos-Martin 2009; Giampietro et al. 2014). MuSIASEM is grounded in bio-economics, complexity, and hierarchy theory. It contributes to system perspectives of sustainability issues, by describing social-ecological systems across hierarchical levels, geographical and temporal scales. MuSIASEM is characterized by a focus on the relations among nexus variables -- how water, food, energy, land, GHG emissions, labor, etc. are interlinked in social-ecological processes at different analytical levels. Complexity in system-observer relationships means that situations can be viewed differently depending on how a problem is framed and analyzed, and by whom (Rosen 1991; Kovacic and Giampietro 2015). The way actors choose to frame and address reality builds on their narratives, which cannot be true or false. In this context, MuSIASEM does not aim to generate quantitative assessments that are true or correct, but ones which are coherent at a given point in time, by linking numbers and variables across scales and levels (what is also known as the “Sudoku effect” (Giampietro and Bukkens 2015)). To do so, MuSIASEM accounts for the inputs and outputs associated with processes at a chosen scale of analysis (e.g., the amount of water consumed, and greenhouse gases emitted by a power plant) and aggregates those processes for higherscale, contextualized descriptions (e.g., the amount of water consumed, and greenhouse gases emitted by the electricity sector of a country, in a given year). For more details on MuSIASEM, see (Giampietro et al. 2014). QST broadens MuSIASEM to include coherence not only between numbers at different levels and scales, but also between numbers and the narratives they build on. This leads to a shift from MuSIASEM’s goal of generating coherent quantitative assessments, to a wider goal in QST of generating assessments that are fit for purpose (i.e., adequate in addressing desired concerns) (Funtowicz and Ravetz 1990). For example, if the concern is ensuring security of energy supply in a country, MuSIASEM builds a coherent analysis by breaking down the country into its different regions, sectors into sub-sectors, and linking energy production and consumption patterns with other nexus dimensions. With a broader scope, 5 QST would add phases of qualitative analysis and/or public interaction to answer questions such as: Why is security important, and to whom is it important? What untold stories are associated with official security narratives? Initial narratives Frame the analysis Design & run MuSIASEM Feedback on narratives ? The QST cycle Figure 1. Initial QST cycle Figure 1 shows the initial conceptualization of QST as an iterative cycle, with a quantitative part and a qualitative one (characterized by the analysis of narratives). At the start of the cycle, narratives relevant to a given innovation are identified. From those narratives, nexus problems to be explored are framed. A MuSIASEM analysis is then designed and executed. The information produced by the quantitative analysis is then interpreted, feeding back into the pool of narratives on the topic. QST allows for each step to be developed in different ways, depending on the purpose of the cycle. This is due both to the openness of the approach, and to the fact that it is still new, and thus not rigorously formalized. Engagement activities can be used at any stage of the QST process: to identify narratives, to design the quantitative assessment or to interpret and discuss results. Crucially, actors are involved in the final stage of sharing and validating analytical feedback. If the narratives are taken from interviews with policymakers, for example, the QST cycle ideally closes by presenting those policymakers with information, leading to the emergence of a different set of narratives in relation to the problem at hand. 6 To understand concerns around innovations and the problems they are supposed to address, we distinguish between three types of interrelated policy narratives to start the cycle, which can be identified either through direct interviews with relevant actors, or through text analyses of policy documents (or both): justification narratives, normative narratives, and explanation narratives (Giampietro 2018). Normative narratives answer the question: “What should be done?” Justification narratives operate at a higher level of abstraction, answering the question: “Why should it be done?” They identify relevant concerns to be addressed and relate them with solutions and normative goals. In doing so, they create expectations for the performance of those solutions. Explanation narratives operate at a lower abstraction level, answering the question: “How should it be done?” They specify the conditions under which those goals should be pursued. This taxonomy is useful in mapping relations among policy narratives, but it does not carry information about who the narrative is generated by, and for what purpose. To include this positionality dimension, we follow the distinction between dominant and alternative imaginaries used by (Longhurst and Chilvers 2019), translating it into dominant and alternative narratives. In this logic, dominant narratives are those generated within centers of power. These centers could be government institutions, academic ones, lobbies, etc. There can be contrasting dominant narratives at different governance levels, e.g., EU narratives versus those of a local government. Alternative narratives are decentralized ones, generated outside of these centers of power. This distinction, which is dependent on levels of governance, is useful when discussing the possible impacts of scientific assessments. The original goal of QST was one of falsification of dominant innovation narratives, using multiscale nexus assessments to show how these narratives are not necessarily grounded in biophysical reality. In this context, the aim was to substitute dominant innovation narratives with a set of alternative ones which are more coherent across environmental and socio-technical dimensions and fit for the concerns they are meant to address. This goal, however, had important limitations as it viewed innovations as purely material, rather than political, objects. We will reflect on this in Sections 4 and 5. As mentioned in the Introduction, QST also recognizes that uncertainty is inevitable in the processes of knowledge production and decision-making. Following the work of (Knight 1921), (Wynne 1992), and (Stirling 2003), 7 uncertainty can be further split into the concepts of risk, strict uncertainty, ignorance, indeterminacy and ambiguity. With risk, the odds are known; in strict uncertainty, unknowns are known; in ignorance, unknowns are unknown. Indeterminacy and ambiguity are irreducible properties of complexity: the former refers to the fact that complex systems are characterized by open causal chains, while the latter refers to the irreducibility of plural perspectives (Kovacic and Di Felice 2019). The functionality of ambiguity in decision-making processes has been long acknowledged. In the policy realm, (Matland 1995) first made the distinction between ambiguity of means (plural explanation narratives) and ambiguity of goals (plural normative narratives), each being functional in order to reach policy decisions. There is also ambiguity in the problem framing (plural justification narratives) when actors have different understandings and underlying assumptions on what the problem is, and why it is a problem (Brugnach and Ingram 2012). Henry Kissinger referred to “constructive ambiguity” as the “deliberate use of ambiguous language in a sensitive issue in order to advance some political purpose” (Berridge and James 2003, 51). However, recognizing our grounding in complexity, it is important here to make the distinction between vagueness and ambiguity (Kovacic and Di Felice 2019). The former refers to lack of clarity when using a term (so constructive ambiguity, to us, falls under vagueness) and is a political choice. For example, the statement that oil imports to the EU are from politically unstable regions (EC 2013) is vague, as it purposefully does not name these regions. Ambiguity, on the other hand, has a wider scope as it refers to incommensurability. The concept of energy security is ambiguous not because it lacks clarity, but because multiple definitions of energy security can co-exist. This distinction is relevant to the case studies presented below. 3 Applications of QST This paper draws on four case studies assessing innovations with QST in the MAGIC project: biofuels, shale gas, electric vehicles, and alternative water resources (Boxes 1-4). The innovations were chosen due to their relations with the nexus elements of water, food, energy and climate. In this section, we briefly describe the main characteristics of each case study (Table 1) and discuss commonalities and differences with regards to their purpose, methods, and results. 8 BOX 1 - Biofuels as an example of the climate-energy-food nexus in the EU. Innovation: Biofuels refer to liquid fuels for transport produced from biomass (EC 2018, 104). The technological development of this innovation can be defined as mature for first generation biofuels, i.e., food crops biofuels, and emerging for advanced biofuels, i.e., biofuels produced from various types of non-food biomass, as some of them are still at a non-commercial stage (Cheng and Timilsina 2011). Nexus relevance: The production of biofuels can be intuitively linked to multiple nexus dimensions, such as the competition of first generation biofuels with food for land and water; and the energetic and economic convenience of advanced biofuels. European policies: Policies for biofuels and advanced biofuels have been in place since 2003 with the biofuels directive (Directive 2003/30/EC). The revised Renewable Energy Directive (REDII) (2018) introduced a 14% REStransportation energy target and a 3.5% advanced biofuels sub-target by 2030 (EC 2018). BOX 2 - Shale gas as an example of the water-energy nexus in the EU . Innovation: Shale gas is an unconventional source of natural gas held in shale stone pores. Its exploitation increases the availability of natural gas. It is considered as a means to increase energy security and geopolitical power related to energy, and as a driver for rural development. Nexus relevance: Shale gas is extracted with hydraulic fracturing or fracking, a technology that drills into the earth and injects a high-pressure mixture of waters and chemicals to fracture layers of shale stone, releasing gas. Thus, the process of fracking is deeply tied to impacts on water and land. European policies: Following the U.S. experience and in preparation for a situation of severe natural gas supply shortage, the EU considered in the early 2010s if and how shale gas should be developed and a regulatory framework to guide a potential implementation of the innovation was built (EC 2014b). The EU Energy Roadmap 2050 sees shale gas as a potentially important new source of natural gas to support the energy transition (EC 2012). 9 Biofuels Shale gas Electric vehicles Alternative water resources REFERENCES CadilloBenalcazar et al. 2021; M. Ripa, CadilloBenalcazar, and Giampietro 2021 Mad rid Lopez 2020 Di Felice et al. 2020; Di Felice, Renner, and Giampietro 2021 Cabello et al. 2020, 2021 SCOPE Purpose of the QST cycle Check robustness of dominant justification narratives Critical reflection on uncertainties and alternatives Governanc e level European European European Local/Regional Geographi cal scale National/ European National/ River Basin/ European (Regionalize d) European Local/ Regional METHODS Narrative analysis Policy text analysis; scientific literature review Policy text analysis, interviews (5) Policy text analysis Semistructured interviews (27) to actors holding knowledge on water, food and energy management Quantitati ve analysis Analysis of production, imports, and exports of biofuels in the Netherlands and Europe GIS scenario of waterenergy nexus for demograph ic patterns of Analysis of nexus dimensions of electric vehicles: energy consumption, GHG Multi-level diagnosis of water-food supply nexus patterns of farming systems. Energy cost 16 extraction wells emissions, local air emissions, labor investments and material requirements. of water sources. Numbersnarratives relation Negative feedback, falsification Negative and positive feedback, explore uncertainties ENGAGEM ENT Purpose Validation (or not) of the falsification of dominant narratives Reflexivity on the narratives, co-creation of new narratives Number and type of activity 1 workshop Interviews Failed workshop 2 workshops Interviews 2 workshops in the Canary Islands Engaged actors Energy modelers and EU policymakers EU policymaker, industry, academics, and relevant NGOs. Regional cooperatives and civil society groups, urban planners and decisionmakers at EU Regions & Cities week, PhD students Agricultural and water management organizations, public authorities, academics, civil society, students 4 Discussion Innovations are not just material objects. They are grounded in imaginaries that rely on values, emotions, and expectations beyond what can be discussed solely through quantitative assessments. A relational nexus lens can open the assessment of innovations to include the narratives and imaginaries that these innovations are grounded in, and which they contribute to, relating them to material dimensions. Our QST processes highlighted the dominant narratives justifying the implementation of nexus innovations in European 17 policies. In the biofuel and shale gas cases, we observed expectations for an energy-secure Europe grounded in political needs to overcome the bottleneck created by the dependence on imported oil and gas. To avoid a political stall, governance shifts to the management of expectations (Borup et al. 2006). On the other hand, the promotion of electric vehicles and alternative water resources relies on expectations of these technologies as effective solutions to pressing social-ecological concerns such as climate change and water scarcity. Ultimately, the four innovations provide an avenue for reconciling the tensions between conflicting policy goals (Kovacic, Rommetveit, and Strand 2020): with economic growth being continuously coupled with environmental protection, the EU may pursue its leadership for sustainable growth while depoliticizing the material requirements and impacts of such growth (Levidow 2013). Dominant narratives placed the innovations within nexus imaginaries (Cairns and Krzywoszynska 2016). By framing each solution as the most desirable one in relation to multiple and sometimes changing concerns, uncertainties associated with the innovations are minimized or ignored, tradeoffs are purportedly resolved, and possible alternatives are cast aside. Our quantitative assessments served to explore whether the innovations’ dominant narratives were coherent when checked from a multi-scale nexus perspective. For biofuels and shale gas, justifications were questioned with biophysical evidence, in order to falsify them; for electric vehicles and alternative water resources, the dominance of the normative-justification relationship was questioned, opening the space for alternatives to emerge. Looking back at Section 2, we asked whether the innovations meet their material expectations. For biofuels, this is clearly not the case. For shale gas, material expectations unfold across different scales. Individual wells provide a surplus of energy and economic revenue that for some landowners may justify the impacts over water and land. However, when it comes to the EU’s expectations of using shale gas to transition to clean energy, this would not be possible at the large scale due to water and land requirements. For EVs, the answer is also “it depends”: our case study highlighted the various sources of uncertainty tied to those expectations (for example, studies claiming that GHG emissions will be reduced with EVs, and studies claiming the opposite). For alternative water resources, material expectations are mostly met: these resources sustain a variety of farming systems and water tables have stopped dropping. In addition, there is a strong 18 sense of security derived from the lack of freshwater resources (controlled in a speculative private market). Farmers now know where they can get water, at stable supply and prices. We also asked what uncertainties surround the implementation of these innovations. On one hand, ambiguity of goals allows for policy measures such as biofuels and shale gas to persist despite contrasting biophysical evidence. On the other, strict uncertainty and indeterminacy allow for electric vehicles and alternative water resources to be framed in a positive light, with ambiguity also playing a part in the justifications associated with these innovations. This view of uncertainties and their underlying politics casts doubt on whether multi-scale assessments can directly shift dominant policy narratives and imaginaries (Turnhout et al. 2020). From an implementation perspective, our quantitative assessments were not discussed with EU policymakers, partially because of a strong resistance of actors from European institutions to get involved in the project. This led to the engagement of other actors who did not feel fully reflected in the narratives presented, while also having a limited capacity to directly influence European policies. The importance of engaging relevant actors from the onset of coproduction processes has been extensively signaled (Norström et al. 2020). Yet, we have learnt that when those actors reject participating, QST may still offer relevant avenues for exploration. A discussion built on both narratives and numbers offered the opportunity to collectively reflect on the ecosystem of narratives that surrounds complex sustainability challenges with a variety of actors. For example, in the electric vehicle workshops, existing evidence on how GHG emissions are tied to the size of vehicles sparked debates about the tension between environmental benefits of the vehicles, and the EU’s priority of boosting its automotive industry. In the biofuel workshop, when presented with EU biofuel justification narratives and contrasting evidence, energy modelers acknowledged existing trade-offs in the biofuel supply chain, suggesting that biofuels tend to play a small part in the construction of their scenarios. In this sense, the modelers were already aware of the biophysical limitations played by biofuels. Grounding the discussion in both narratives and metrics created the opportunity to discuss how their own views of the role of biofuels differ from the EU’s official framing. This outcome questions 19 the preference for using solely quantitative tools in the nexus literature (Albrecht, Crootof, and Scott 2018) by suggesting that qualitative and plural framings of nexus relations are crucial to bring politics into reflexive assessments (van Gevelt 2020). In addition to discussing dominant narratives and their evidence base, for the cases of electric vehicles and alternative water resources the engagement processes served to open conversations on context and alternatives. The workshops on electric vehicles held space for imagining alternative solutions to the unsustainability of the transport system, by envisioning the future of cities and of rural-urban divides. The discussion focused on the potential roles of carsharing, public transport, walking and cycling in shaping future transport systems. This, in turn, led to broader conversations about where responsibility lies, and about the local dimensions of mobility, suggesting that electric vehicles may be useful in some contexts, and less so in others. The workshops on alternative water resources focused on discussing what new roles could be envisioned for these innovations within futures with more sustainable farming practices. Framing the conversation around desirable futures helped expand the debate from the technicalities of water management to the difficulties small farming systems face to stay afloat within an export-imports based European food policy. Our outcomes in terms of contextualizing and broadening discussions, as well as empirical and methodological insights to STS, add to calls for “opening up” the assessment of innovations and of the nexus (Stirling 2008; Urbinatti et al. 2020). 20 New narratives & imaginaries Normative narratives Dominant nexus imaginary Justification narratives QST: open the relationship up to plurality QST: falsify the relationship Diffuse impacts Diffuse impacts ? ? Justification narratives Normative narratives Justification narratives Normative narratives Figure 2. QST as a means to question dominant imaginaries Building on our shared experiences, QST can be re-framed from the closed cycle of Figure 1 to a means of opening the debate of science-policy relations to wider audiences, discussing dominant imaginaries as shown in Figure 2. When the promotion of an innovation responds to vested interests, like we observed in the cases on biofuels and shale gas, the QST route of falsification can question whether dominant narratives are coherent by amplifying critical evidence and alternative narratives. This is a strategy to include “unknown knowns” in the for our four case studies, dominant EU narratives tie together economic growth, security, environmental protection, and local development. Nexus innovations are seen as a way to simultaneously address these multiple concerns, confirming 21 the existence of what has been termed in the literature as “nexus imaginaries” (Cairns and Krzywoszynska 2016). In this context, simply presenting policymakers with biophysical evidence is not enough to generate change in their narratives. On one hand, innovations are inevitably tied to different types of uncertainty that may serve political purposes and cannot be reduced by producing additional quantitative assessments. On the other hand, there may be limited interest for these kinds of critical results, as our attempts to engage with policymakers suggest. This, however, does not mean that analyses such as QST, where relationships between numbers and narratives are outlined, do not have a place in the current ecosystem of science for policy. What we saw through our engagement events is that QST may be used to question existing narratives and co-create new imaginaries. In this way, by discussing innovations beyond their material dimensions, the nexus can become a lens that accommodates plural visions. Given the experimental nature of the case studies, there are many limitations to QST in its current form, and to the meta-analysis that can be extracted from the four innovations. While we provided two main “routes” which may be taken through QST (falsification vs. opening up to plurality), the operationalization of QST at the methodological level is not yet formalized – in this first iteration, different researchers made different decisions regarding their own case studies. We hope that this paper can serve as a point of reflection for researchers to move forward in more coordinated steps. This includes engagement activities, an important part of QST that was implemented in different ways. Further discussion is needed to reflect upon the role of engagement in QST and more broadly in the analysis of science-policy relations. On this, the difficulty in engaging relevant actors from the onset was particularly limiting, and future iterations of QST may have to re-focus objectives by keeping this limitation in mind. When it comes to impacts, we mentioned in our discussion how QST may lead to diffuse impacts, which are harder to quantify. While we do not believe that these impacts should be reduced to metrics, it would be fruitful to think of ways to track these processes over time, for example through the analysis of the evolution of narratives, or the engagement of the same actors across the years. This would require multiple iterations of QST, while what we presented here was an individual “cycle.” We are aware that this kind of 22 work would require extensive time and energy, and that funding bodies may be reluctant to invest in tracking long-term, diffuse impacts of experimental work of this sort. Still, we hope that our reflections may provide food for thought for those researching nexus innovations and their role in science-policy relations, and that the insights generated by our first iteration of QST can enrich existing methods and approaches. process (Giampietro and Funtowicz 2020). Alternatively, when opportunities for pluralistic processes arise, one might opt for a co-creation route attending to different uncertainties in symmetrical dialogical relations (Giatti 2019). Rather than direct policy or narrative change, the outcomes from these processes might be diverse and diffuse. They extend beyond the observed timeframe and are harder to trace or quantify. In the long run, routes of this kind may contribute to shared alternative imaginaries that may, in turn, lead to changes in the governance of nexus innovations. Importantly, Figure 2 emerged from a first run of experimentation of QST: as such, steps are not yet formalized. While we believe that a degree of openness in the methodology is necessary to accommodate a variety of case studies, actors, and objectives, further testing and implementation are needed to discuss how QST can be used in consistent ways across diverse case studies. 5 Concluding remarks Scientists are increasingly asked to produce policy-relevant research, intended as research that can have direct, quantifiable impacts on decisionmaking processes. In this paper, we have presented and reflected upon our experiences of implementing QST, an approach initially aimed at assessing dominant innovation narratives and empirically testing them through multi-scale nexus assessments, to replace them with narratives that are grounded in biophysical reality. Our experiences speak to the role played by socio-technical imaginaries in shaping policy solutions, particularly innovations. We found that for our four case studies, dominant EU narratives tie together economic growth, security, environmental protection, and local development. Nexus innovations are seen as a way to simultaneously address these multiple concerns, confirming the existence of what has been termed in the literature as “nexus imaginaries” (Cairns and Krzywoszynska 2016). 23 In this context, simply presenting policymakers with biophysical evidence is not enough to generate change in their narratives. On one hand, innovations are inevitably tied to different types of uncertainty that may serve political purposes and cannot be reduced by producing additional quantitative assessments. On the other hand, there may be limited interest for these kinds of critical results, as our attempts to engage with policymakers suggest. This, however, does not mean that analyses such as QST, where relationships between numbers and narratives are outlined, do not have a place in the current ecosystem of science for policy. What we saw through our engagement events is that QST may be used to question existing narratives and co-create new imaginaries. In this way, by discussing innovations beyond their material dimensions, the nexus can become a lens that accommodates plural visions. Given the experimental nature of the case studies, there are many limitations to QST in its current form, and to the meta-analysis that can be extracted from the four innovations. While we provided two main “routes” which may be taken through QST (falsification vs. opening up to plurality), the operationalization of QST at the methodological level is not yet formalized – in this first iteration, different researchers made different decisions regarding their own case studies. We hope that this paper can serve as a point of reflection for researchers to move forward in more coordinated steps. This includes engagement activities, an important part of QST that was implemented in different ways. Further discussion is needed to reflect upon the role of engagement in QST and more broadly in the analysis of science-policy relations. On this, the difficulty in engaging relevant actors from the onset was particularly limiting, and future iterations of QST may have to re-focus objectives by keeping this limitation in mind. When it comes to impacts, we mentioned in our discussion how QST may lead to diffuse impacts, which are harder to quantify. While we do not believe that these impacts should be reduced to metrics, it would be fruitful to think of ways to track these processes over time, for example through the analysis of the evolution of narratives, or the engagement of the same actors across the years. This would require multiple iterations of QST, while what we presented here was an individual “cycle.” We are aware that this kind of work would require extensive time and energy, and that funding bodies may be reluctant to invest in tracking long-term, diffuse impacts of experimental work of this sort. Still, we hope that our reflections may provide food for thought for 24 those researching nexus innovations and their role in science-policy relations, and that the insights generated by our first iteration of QST can enrich existing methods and approaches. Acknowledgements This research has been funded by: the European Union’s H2020 project MAGIC: Moving Towards Adaptive Governance in Complexity (MAGIC GA No. 689669); European Union’s FP7 project IANEX: Integrated Assessment of the Nexus: the case of hydraulic fracturing, Marie Curie International Outgoing Fellowship GA No. 623593; Spanish Ministry of Science’ Juan de la Cierva Fellowship IJC2019038847-I/ AEI / 10.13039/501100011033; Maria de Maetzu CEX2019-000940-M. The present work reflects only the authors' view and the funding Agencies cannot be held responsible for any use that may be made of the information it contains. The authors are grateful to the Sci-Hub free knowledge platform from which many of the referenced papers could be accessed. References Albrecht, Tamee R., Arica Crootof, and Christopher A. Scott. 2018. “The WaterEnergy-Food Nexus: A Systematic Review of Methods for Nexus Assessment.” Environmental Research Letters 13 (4). IOP Publishing: 043002. Allen, Timothy F. H., and Mario Giampietro. 2006. “Narratives and Transdisciplines for a Post-Industrial World.” Systems Research and Behavioral Science 23 (5): 595–615. doi:10.1002/sres.792. Berridge, G. R., and Alan James. 2003. “Constructive Ambiguity.” A Dictionary of Diplomacy. Palgrave MacMillan. Borup, Mads, Nik Brown, Kornelia Konrad, and Harro Van Lente. 2006. “The Sociology of Expectations in Science and Technology.” Technology Analysis & Strategic Management 18 (3–4). Taylor & Francis: 285–98. Brugnach, Marcela, and Helen Ingram. 2012. “Ambiguity: The Challenge of Knowing and Deciding Together.” Environmental Science & Policy 15 (1): 60–71. doi:10.1016/j.envsci.2011.10.005. Cabello, Violeta, Ana Musicki, David Romero, Ângela Guimarães Pereira, and Baltasar Peñate. 2020. “Solving Water Problems vs. Creating New Ones: The Use of Alternative Water Resources for Irrigation.” MAGIC (H2020-GA 689669), Project Deliverable 6.7, 28 February 2020. Cabello, Violeta, David Romero, Ana Musicki, Ângela Guimarães Pereira, and Baltasar Peñate. 2021. “Co-Creating Narratives for WEF Nexus Governance: A Quantitative Story-Telling Case Study in the Canary Islands.” Sustainability Science 16 (4): 1363–74. doi:10.1007/s11625-02100933-y. Cadillo-Benalcazar, Juan J., Sandra G. F. Bukkens, Maddalena Ripa, and Mario Giampietro. 2021. “Why Does the European Union Produce Biofuels? Examining Consistency and Plausibility in Prevailing Narratives with Quantitative Storytelling.” Energy Research & Social Science 71 (January): 101810. doi:10.1016/j.erss.2020.101810. 25