Sustainability Science (2023) 18:1235–1250 https://doi.org/10.1007/s11625-022-01265-1 ORIGINAL ARTICLE Implementing post‑normal science with or for EU policy actors: using quantitative story‑telling Kirsty L. Blackstock1 · K. A. Waylen1 · K. B. Matthews2 · A. Juarez‑Bourke1 · D. G. Miller2 · A. Hague1 · D. H. Wardell‑Johnson2 · M. Giampietro3,4 Received: 25 February 2022 / Accepted: 8 November 2022 / Published online: 20 January 2023 © The Author(s) 2023 Abstract There is increasing recognition of the wicked nature of the intertwined climate, biodiversity and economic crises, and the need for adaptive, multi-scale approaches to understanding the complexity of both the problems and potential responses. Most science underpinning policy responses to sustainability issues, however, remains overtly apolitical and focussed on technical innovation; at odds with a critical body of literatures insisting on the recognition of systemic problem framing when supporting policy processes. This paper documents the experience of implementing a mixed method approach called quantitative story-telling (QST) to policy analysis that explicitly recognises this normative dimension, as the methodology is part of a post-normal science (PNS) toolkit. The authors reflect on what was learnt when considering how QST fared as a tool for science–policy interaction, working with European Union (EU) level policy actors interested in sustainable agricul- ture and sustainable development goal 2. These goals—also known as UN Agenda 2030—are the latest institutionalisation of the pursuit of sustainable development and the EU has positioned itself as taking a lead in its implementation. Thus, the paper illustrates our experience of using PNS as an approach to science policy interfaces in a strategic policy context; and illustrates how the challenges identified in the science–policy literature are amplified when working across multiple policy domains and taking a complex systems approach. Our discussion on lessons learnt may be of interest to researchers seeking to work with policy-makers on complex sustainability issues. Keywords Quantitative story-telling · Post-normal science · Science–policy interactions · European Union · Agriculture Introduction The publication of the UN 2030 agenda and its 17 sustain- able development goals (SDGs) renewed rhetorical com- mitment to sustainability. The Agenda and SDGs promote shared responsibility between the global North and South; Handled by Lars Coenen, Western Norway University of Applied and the involvement of societal actors from the public, pri- Sciences, Norway. vate and civic domains (Simon and Schiemer 2015). How- ever, policy processes associated with the SDGs have been * Kirsty L. Blackstock Kirsty.blackstock@hutton.ac.uk critiqued for retaining a managerial approach focussed on technical innovations (Koff and Maganda 2016). Failing to 1 Social, Economic and Geographical Sciences, James take a systemic approach to delivery of the SDGs (Norström Hutton Institute, Craigiebuckler, Aberdeen AB15 8QH, et al. 2014) and giving insufficient attention to political Scotland, UK choices and power imbalances perpetuates inequitable and 2 Information and Computational Sciences, James Hutton unsustainable outcomes (Blühdorn 2011). Institute, Craigiebuckler, Aberdeen, Scotland, UK Understanding sustainable development as a ‘wicked 3 Institute of Environmental Science and Technology (ICTA), problem’ (Termeer et al. 2015) highlights the political and Universitat Autònoma de Barcelona, Bellaterra, Barcelona, Spain subjective nature of how sustainability problems and solu- 4 tions are framed. Although scientific knowledge plays a ICREA, Pg. Lluís Companys 23, Barcelona, Spain 13 Vol.:(0123456789) 1236 Sustainability Science (2023) 18:1235–1250 critical role in the understanding of interconnected sustain- PNS and SPIs. First, fundamental to effective SPIs are con- ability challenges and responses (Clark and Dickson 2003), cepts of Credibility, Relevance (salience) and Legitimacy science is not neutral and separate to these wicked problems. (CRELE) (Cash et al. 2002) to ensure that scientific findings Facilitating appropriate use of science (and other) knowl- are judged by policy-makers as relevant and timely, based edges for sustainability is at the heart of science–policy on robust methodologies using transparent and inclusive interfaces (SPIs) literature. processes. More recently, Sarkki et al. (2015) added ‘itera- Decades of scholarship on science–policy interfaces or tivity’ to credibility, relevance and legitimacy, to recognise SPIs (e.g. van den Hove 2007) and the role of science in the dynamics involved in these policy–science exchanges decision-making (e.g. Jasanoff and Wynne 1998) plus cur- (van Hulst and Yanow 2016). PNS scholars use the TRUST rent recommendations on transdisciplinary research (Berg- (transparency, robustness, uncertainty, sustainability and mann et al. 2021) provide a variety of insights and recom- transdisciplinary) mnemonic (Kønig et al. 2017). mendations for improving scientific knowledge use in policy Transparency and robustness are shared with CRELE processes. In particular, there is a need to better acknowl- whilst a focus on sustainability is one manifestation of rel- edge political factors inherent to science knowledge (Wes- evance. However, uncertainty, particularly ontological uncer- selink et al. 2013); better value and reflect different knowl- tainty, arises for non-equivalent perspectives on the same phe- edge systems (Cornell et al. 2013; e.g. Norström et al. 2020); nomena, and emphasises the importance of exploring multiple and the multiple factors shaping policy processes (Cairney problem frames rather than assuming problems are agreed or 2012); and recognise limits to our certainty and control of solutions straightforward. This is because it is problematic socio-ecological systems (Partelow 2015; Miller et al. 2013). to impose structured solutions on wicked problems that may In practice, the SPI literature draws attention to the need, not recognise the wider systemic issues (Hisschemöller and at minimum, to co-produce research agendas (Sutherland Hoppe 2001). With respect to SPIs, a PNS approach implies et al. 2012) and ideally, to ensure that scientific knowledge that policy actors should work with scientists to ‘find the prob- is produced, interpreted and used with the policy actors for lem’ (Hoppe 2018) before they can discuss solutions to avoid whom the science was designed to support. Instead of knowl- screening out difficult systemic considerations. edge transfer, a dialogue is needed to comprehend and shape Trans-disciplinarity highlights how knowledge is held by how knowledge is needed, understood, and used (Young many actors beyond formally qualified scientific experts, so et al. 2014). However, these ideas are not always reflected by problem structuring requires the input of ‘transdisciplinary practice; how to foster new co-produced approaches remains teams’ of varied societal actors (Alonso-Yanez et al. 2019). unclear (Cundill et al. 2019), and aspirations for transforma- Therefore a PNS approach to SPIs requires policy actors to tive sustainability science are often disappointed (Jaganna- be active knowledge producers as well as knowledge users. than et al. 2020). These disappointments may be due to the Furthermore, to tackle wicked problems, such as sustainable failure to take a politically informed approach to knowledge development, SPI actors need to be responsive and inter- co-production in SPIs (Maas et al. 2022). ested in difficult political challenges (Termeer et al. 2015). This suggests that a ‘post-normal science’ perspective on Many papers on SPIs highlight the relational and longitudi- sustainability might be useful. Post-normal science (PNS) nal demands of the process on individuals, e.g. Sarkki et al. is suggested for when the "facts [are] uncertain, values in (2019) and Bergmann et al. (2021). dispute, stakes high and decisions urgent" (Funtowicz and These individual capacities are enacted within a wider Ravetz 1993). It is therefore relevant to many contempo- context (Sarkki et al. 2019), in this case governance pro- rary sustainability challenges. PNS is a normative philoso- cesses, that balance steering and coordination by the EU phy not a precise methodological prescription (Turnpenny with the autonomy of member states (Schout and Jordan et al. 2011); providing room for heterodox interpretations. 2005; Rüffin 2020). These processes include the European Despite the potential application of PNS approaches to pol- Commission, its agencies and the political arenas of the Par- icy support, there are very few examples of PNS being used liament, Council of Europe and Council of Ministers. Pol- with EU policy actors in transdisciplinary teams. Instead, icy development, implementation and evaluation processes PNS research focus tends to be whether practices are post- are steered by collective discourses and behaviours (Can- normal (Karpińska 2018; Rose 2018; Ainscough et al. del et al. 2014; Voelker et al. 2019; Gossling et al. 2016). 2018); focussed on a specific controversy (Palliser and Dod- Therefore, the interplay between individual and institution son 2017); or working with specific local actors (Tsukahara conditions PNS SPIs (Guimarães Pereira and Saltelli 2017). 2017). This paper analyses the use of PNS with policy actors Policy-relevant research, working with specific actors, is within EU institutions and considers how PNS informs the central to an approach called quantitative story-telling (QST) science–policy interface. (Di Felice et al. 2021). The purpose of QST is to ensure that To understand how PNS works to support science–policy sufficient attention is given to how quantified metrics and interfaces, we consider attributes seen as relevant to both indicators are produced and interpreted by specific actors, at 13 Sustainability Science (2023) 18:1235–1250 1237 a particular time, in a particular context for a specific purpose The paper contributes to sustainability science through pro- (Kovacic 2018; Voelker et al. 2019). QST is a particular form viding an individual case study on a PNS-SPI involving multi- of participatory mixed methods to explore the performance ple policy institutions within the European Commission. The of a system. QST follows a cycle of ‘semantic’ activities paper now describes the case, the policy context, the stages (engagement with stakeholders, identification and articula- of QST and how the data were collected and analysed. The tion of dominant framings and narratives) that are used to discussion highlights how the case compares to other empiri- inform the ‘formal’ aspects of QST (quantified results from cal SPI papers. The paper therefore has relevance to those models or accounting frameworks). These results are deliber- practising SPIs related to sustainability, even if the reader is ated on with attention to how different actors understand and not committed to using the QST approach specifically. respond to these metrics. The process of QST is relevant to SPI literatures, as the formal-analytic outcomes are appraised in terms of credibility, relevance and legitimacy (Giampietro Materials and methods et al. 2006; Allen and Giampietro 2016). QST is designed to be consistent with a post-normal To examine the use of QST in a policy context, we imple- approach in several ways. First, through a focus on uncov- mented a QST cycle within the EU agricultural policy con- ering and discussing multiple problem frames at the start text (Matthews et al. 2017, 2019). This was the focus of one and end of the cycle to expose ontological uncertainties. work-package within a broader multi-partner project called Second, through a focus on ‘transdisciplinary teams’ inter- MAGIC,1 funded by the European Commission within the acting throughout the QST cycle. Third, most applications Horizon 2020 programme—this paper refers to the instance of QST have used a ‘societal metabolism accounting frame- of implementing QST in an EU wide policy setting and work’ (Giampietro and Mayumi 2000) combining multiple focusses on the procedural findings with the accounting simultaneous perspectives on flows and funds within both findings in a sister paper (Matthews et al. 2021). the biosphere and technosphere; across different decomposi- tions of space and economic or demographic sectors. This Focus multi-faceted and systemic approach to accounting includes externalities and was useful to enable an evaluation of sus- The QST approach was implemented in two cycles, each tainability. Thus, QST ensures the science part of a SPI is starting with a different (albeit connected) focal ‘story’. systemic and recognises multiple knowledge claims whilst The first QST cycle (2017–18) focussed on “CAP aims to the QST process highlights the importance of the interface ensure European agricultural competitiveness in the world to agree the problem frame before findings and solutions are market and aims to deliver public goods, such as biodiversity debated. Policymakers are enrolled as part of the transdisci- conservation, water quality, and climate change mitigation. plinary team choosing the stories to follow and making sense These aims are in opposition.” This focal story responded of the quantified results. This paper does not describe the to the proposed revision of Common Agricultural Policy quantitative accounting aspects of QST, which can be found (CAP) objectives and pressure to integrate agricultural pol- in existing literature (Ripoll Bosch et al. 2018; Krol et al. icy with delivery of climate and environmental objectives 2018; Giampietro et al. 2017). In short, the paper locates (Hodge et al. 2015). Calls for a systems approach to CAP itself within the need for SPIs to advance sustainability solu- (Kuhmonen 2018; Scown et al. 2019; Scown and Nicholas tions, explores the benefits of taking a PNS approach that 2020) made the CAP an interesting focus for a PNS QST recognises contestation and uncertainty, and the potential application. The CAP is a huge and complex policy, with a of the QST methodology to deliver a PNS-SPI. We consider plethora of policy instruments nested underneath its objec- how to address the complexity of multi-dimensional and tives. The first cycle of QST was focussed on the outcomes uncertain aspects of steering the delivery of SDGs across of farming activities (as expressed through metrics generated multiple policy units. This challenge requires research with from the Farm Accountancy Data Network2) and associa- policy actors in transdisciplinary teams, rather than research tions with material, financial and human inputs. It was not on EU sustainable development policy-making for policy focussed on a particular policy instrument, farming sector or actors. commodity, precisely to give a holistic overview. The paper has the following research objectives: 1. Highlight the benefits and challenges experienced work- 1 The H2020 MAGIC project (Moving Towards Adaptive Govern- ing with EU level policy actors steering sustainable agri- ance in Complexity: Informing Nexus Security) is described and reported at https://​cordis.​europa.​eu/​proje​ct/​id/​689669 cultural policy as part of the implementation of SDG2 2 https://​ec.​europa.​eu/​info/​food-​farmi​ng-​fishe​ries/​farmi​ng/​facts-​and-​ 2. Consider whether a PNS approach like QST can facili- figur​es/​farms-​farmi​ng-​and-​innov​ation/​struc​tures-​and-​econo​mics/​ tate a systemic approach to sustainability. econo​mics/​fadn_​en 13 1238 Sustainability Science (2023) 18:1235–1250 The second QST cycle (2019–20) positioned itself in evant to the themes identified in stage 1. The accounting terms of how the CAP was becoming enrolled in delivery framework is populated with quantitative data to become of the Sustainable Development Goals. The focus for the a quantified (and sometimes spatial) representation of second cycle was “Sustainable agriculture is an important the system. The first cycle used a societal metabolism societal goal: it helps to support environmental policies as approach focussed on associations between different well as being a key component of SDG2. The European Farm Accounting Data Network (FADN) indicators, Commission claim CAP enables sustainable agriculture, whereas the second cycle used the full Multi-Scale Inte- supports environment policies, such as WFD and N2K, and grated Analysis of Societal and Ecosystem Metabolism supports progress to the SDGs. However, agricultural exter- (MuSIASEM) accounting methodology (Giampietro and nalities harm European progress to achieve its sustainability Mayumi 2000). goals both within and beyond Europe”. At the time of the 4. Contextualise and present metrics: Metrics are gener- second QST cycle, the EU’s SDG implementation had been ated that summarise the feasibility (within biophysical criticised for being too focussed on encouraging develop- limits) and viability (within socio-economic limits) and ment in non-EU States and neglecting its internal unsustain- desirability (distributional, burden sharing and accept- able trends (Szopik-Depczynska et al. 2018; Montéville and able outcomes) of the system as it is represented by the Kettunen 2019). Although the literature suggests focussing quantified data. on the interactions of all 17 SDGs (Pradhan et al. 2017), we 5. Discuss interpretations and implications: This stage sees selected SDG2 (sustainable agriculture, nutrition, food secu- the deliberative interpretation of the outputs of the QST rity and ending hunger) as a more tractable focus, broaden- analysis with policy actors. It is both a process of ‘clos- ing the CAP-environment discussion of the first cycle into ing the loop’ and the shaping of any further cycles— the wider food system and its impacts. either with new stories or with alternative cases. Although both cycles take a story and model simulation approach also used in participatory scenario analysis (Kok The aim is to complete the QST cycle in transdisciplinary et al. 2015; Swart et al. 2004); the focus was on widening teams to generate meaningful outputs that stimulate delib- the frame by which past and current sustainability trends are erations about the implications of these metrics for tackling understood, not on potential future actions. sustainability challenges (Matthews et al. 2019). QST Data collection The overall stages for each QST cycle described above are At the start of the first QST cycle, eight policy actors from introduced in our project report (Matthews et al. 2020). the European Commission and its agencies were interviewed The cycle explains the sequence of methodological steps regarding CAP evolution and its relationship to sustainable undertaken to derive the results presented below. development (see Voelker et al. 2019). A content analysis was conducted of CAP in 2016–17 and these data were sup- 1. Identify key themes relevant to policy: This stage draws plemented by field notes from a small cross-policy focus on desktop analysis and primary data to identify the group with members of staff from different Directorates issues and ideas which are of interest to policy actors. General (DG) —DG Agri, DG Energy, DG Environment and This initiates the formation of ‘Transdisciplinary teams’ the Parliament’s Environment Committee held in May 2017. (comprising policy stakeholders as well as MAGIC sci- Deliberation over a choice of stories for QST was held via entists). email and an online meeting with a further five policy actors 2. Decide what to represent in societal metabolism analy- from DG Agri in August 2017. Once societal metabolism sis: This stage develops a more specific shared under- analyses were completed and choices about which metrics standing of what will be analysed and how to make to present and how to present them were made, a workshop the issues identified in stage 1 amenable to societal was held in October 2018. The one and a half hour workshop metabolism (flows of energy and material) accounting. was held in DG Agri meeting rooms in Brussels, facilitated This stage moves progressively from deciding higher- by a contact in DG Agri and six DG Agri participants, some level priorities (the type and numbers of issues to be of whom had not previously been involved in MAGIC. The analysed), towards decisions on the specific aspects of workshop consisted of a presentation introducing the project, systems that have to be represented that will shape the the methodological approach, and the findings, ending with quantitative analysis (e.g. setting system boundaries, a series of questions for discussion. scales of analysis, useful indicators). To initiate the second QST cycle, we interviewed five 3. Compile data and carry out societal metabolism account- new actors (two from DG Agri, two from an NGO and one ing: This stage sees the formalisation of the systems rel- from the Joint Research Centre) to fill in the gaps from our 13 Sustainability Science (2023) 18:1235–1250 1239 policy review of the EU implementation of SDGs during Table 1  Analytical approach May 2019. Field notes from attending the Commission’s Requirement How these requirements were operationalised in QST Green Week 2019 were also used for context. Three dif- ferent workshops were held in November 2019, January Systemic Multi-dimensional metrics Multi-scale and multiple system perspectives 2020 and June 2020 (originally planned for March 2020 but Multiple policy domains delayed due to COVID-19). The first workshop was held in Post-normal Uncertainty Brussels, at Scotland House, a venue independent from the Unstructured problem finding official EU institutions, attended by fifteen participants from Inclusive transdisciplinary teams DG Devco, DG Clima, DG Agri, DG Research, EASME Normative (political) approach and a think tank. As with the first cycle, the one and a half Science–pol- Credible icy interac- Relevant hour workshop summarised the purpose of the research, the tions Legitimate methodology and the findings. The second workshop was a Iterative ‘parliamentary breakfast’ hosted by a member of European Parliament, attended by twelve participants from the Com- mission and Parliament. The third workshop was a webi- nar with eight participants from the European Environment Agency and associated topic centre researchers. The sec- ond and third workshops used the same structure as the first workshop. In both cycles, the last QST stage was designed to ask the policy actors to share their knowledge and expertise both during workshops and afterwards privately through feedback forms. A workshop report was provided to the participants with an invitation to follow up any points of interest with the MAGIC scientific team. To improve our understanding of how QST was experienced by the non-MAGIC participants, a colleague not involved in the workshops undertook four feedback interviews with participants from the Commission and agencies during April–June 2020. The nature of the QST process means data were collected from a self-selecting sample of those motivated to engage with PNS and QST. Researchers also attempted to engage with people from other relevant Commission agencies, but were either unable to, as in the case of DG Sante, or had Fig. 1  An overview of the quantitative story-telling (QST) cycle limited success, in the case of DG Env. Therefore, the results must be interpreted within these caveats. The interviews were recorded and transcribed. In all workshops, full field as illustrated in Table 1 below. The approach is a simplifica- notes were taken by a dedicated researcher. The experience tion of the multiple literatures introduced in the Introduction of the MAGIC science team was captured through internal to provide three clear aspects with which to consider imple- meeting notes and research memos. The research received menting a PNS approach to sustainability with EU policy authorisation from the James Hutton Institute Research Eth- actors. ics Committee. Data analysis Results The transcripts, field notes, meeting notes and feedback form The results cover the main learning points from the two data were managed using the qualitative research database cycles of QST (see Fig. 1) before addressing how QST sup- NVIVO 12 and labelled with case attributes to allow analy- ported policy for sustainability using the analytical frame- sis by type of actor, policy domain and different settings work (Table 1). Each section is a snapshot of the main (workshop or interview). Initial thematic coding illustrated themes summarising each aspect before the implications of what was learnt in the QST process for technical reporting the findings are discussed in Sect. 4. (Matthews et al. 2020). Further focussed analysis, embracing theoretical multiplicity (Karpouzoglou et al. 2016), was used 13 1240 Sustainability Science (2023) 18:1235–1250 What did we learn about implementing QST? MuSIASEM for most of the authors (Matthews et al. 2021). In both cycles, the application was in ‘diagnostic’ Here we (the authors) set out the aspects that generated mode (analysis of current state) not ‘simulation’ mode. the most reflection about the practice of undertaking QST, Running QST in simulation mode to discuss potential discussed according to each stage of QST (Matthews et al. futures could be less confronting than discussing cur- 2018, 2020). rent policy choices. As with stage 1, it was important to balance the potential to deepen analyses with reserving 1. Identify key themes relevant to policy: In both QST resources for stages 4 and 5, as well as for further cycles. cycles, the scope was set by the project specification, 4. Contextualise and present metrics: In both QST cycles, but there were many potential themes. This stage proved the societal metabolism applications produced metrics difficult to achieve sufficient depth and breadth of analy- for environmental pressures (feasibility) and socio-eco- sis for relevant themes within the available time. More nomic pressures (viability), but desirability remained a resources were spent in the first cycle on formal docu- product of deliberation reserved for stage 5. The envi- ment analysis and interviews with relevant policy actors ronmental pressures were converted to spatial impacts than on the second cycle. Policy actors were not asked in the second QST cycle. The main challenge for QST to select the theme in the second cycle as this process was considering which metrics to represent visually for extended the time taken on stage one without improving deliberation in stage 5 (Cortes Arevalo et al. 2020). The the quality of the discussion in stage five of the cycle. It final technical report contained complex Circos dia- was important to anticipate and iterate between stage 1 grams (Krzywinski et al. 2009); relationship maps and and stage 5 (the feedback stage) in terms of identifying extent-intensity matrices (Matthews et al. 2020), but in themes relevant to identified actors. It was more impor- both cycles, the presentations for policy actors focussed tant to invest in the last stage of the cycle than perfect mainly on using bar charts, tables and two dimensional the first stage. graphs, which were more familiar to the anticipated 2. Decide what to represent in societal metabolism anal- audience and required less explanation. Contextualising ysis: This stage proved very important for the shared and explaining the ‘take-home’ message and the depth learning within the MAGIC researchers. We used the of analysis leading to this message took considerable Eurostat Farm Accountancy Data Network (FADN) as effort. As with stage 2, stage 4 was a very important the best dataset on which to base a pan-EU farm sys- step in the QST cycle, as it required many choices about tems sustainability assessment but generated constraints what to represent and how to represent the data. These regarding what data were available, at what resolution processes are underreported in academic literature, yet and decomposition by scale or sector (Matthews et al. are fundamental in shaping deliberations, particularly 2019, 2021). Iteration occurred between stage 1 and when there is limited time requiring a reduced selec- stage 2, reappraising which themes were most promis- tion of metrics. It was particularly difficult in the second ing given the data available. Stages 1 and 2 shape stage cycle given the huge extent of data combinations pos- 3, regarding how thorough the accounting should be, and sible. Therefore, this stage needs more time than one stages 4 and 5 regarding presenting the data. Using dif- might expect. ferent societal metabolism methodologies in the first and 5. Discuss interpretations and implications: In both cycles, second stages, to advance scientific knowledge, created we chose to use an existing seminar slot, recommended extra work—using familiar methodologies in further by policy actors in stage 1 of the first cycle as the timing iterations would make this stage more tractable. Feed- and location would be convenient for the participants. back on the final choice of representations was shared However, this introduced time constraints (maximum with the policy actors during the first cycle. Discussion 90 min) and perhaps generated expectations of a tradi- with policy actors at the end of the first cycle about tional ‘show and tell’ scientific presentation. The pres- what the FADN dataset could support helped guide the entations asked about the political and bureaucratic con- choices made in the second cycle. figurations that block or enable transformation. In the 3. Compile data and carry out societal metabolism second phase, we provided our own view of the implica- accounting: Although the first QST cycle did not use a tions before asking workshop participants whether they full MuSIASEM application, both QST cycles shared agreed with us. However, very little of the subsequent the challenge of taming the exponential growth in poten- discussion addressed these questions. Provision of feed- tial metrics being produced. There were many techni- back forms, to enable people to express views privately cal challenges, such as discontinuity in the FADN data also did not generate much insight. Even when discuss- available, limited access to the data, missing data values, ing the results with environmental policy actors, most of particularly given it was the first experience of using the discussion was regarding the technicalities of soci- 13 Sustainability Science (2023) 18:1235–1250 1241 income (before and after agricultural subsidy); and in the second cycle, more information, such as reliance on family labour and seasonal labour, was generated. The second cycle introduced data on nutritional demand and dietary choices. Central to societal metabolism analyses is the production of both extents (total hectarage, number of livestock, hours worked) and intensity (such as stocking rate per hectare, energy inputs per hour). In both cycles, visualisations were used to illustrate data in both extents and intensities, often on the same figure (e.g. soil erosion losses per ha, and total soil losses for each member state). Multi-scale analysis was presented in both QST cycles. In the first cycle, environmental and social data were pre- sented by member state and NUTS2 regions and by differ- ent farm type; and intensity metrics presented by hectare, hour or euro. In the first cycle, change over time was indi- cated, but the second phase tended to produce static data for a single year. In both cycles, EU commodity imports and exports were explored; with the second phase introducing the dependence of EU agricultural commodities on imported materials and labour, illustrating the difficulties of reducing inputs from beyond the EU. Multiple policy domains were involved in both cycles, Fig. 2  Different levels of the agri-food system with participants from different DGs (see Sect. 2.3) with interests in climate, development, agriculture and the envi- ronment. The primary policy focus was the CAP, but the etal metabolism and comparisons with lifecycle analysis. results for water, soil and pollinators were relevant to the The innovation of ‘feedback’ interviews was extremely environmental fiche; and GHG results relevant to climate useful to get insights into how the policy actors experi- action. In the second cycle, CAP was a nexus for multiple enced QST and we would recommend this as an impor- aspects of the SDGs, but the data were not presented against tant part of the QST cycle. This stage was both a way to policy objectives nor individual policy instruments. Instead, close the QST cycle, but also the springboard for further the data were presented against the four main components of QST cycles. SDG2, to locate CAP and its sister policies within a wider sustainability framing. How did the policy actors respond to the systemic Overall, there was a mixed response to the systemic nature of the QST? nature of the QST results. Whilst most policy actors were interested in the different aspects of the system, the soci- The first iteration of QST was less ambitious in coverage, etal metabolism accounting was unfamiliar and difficult to focussed primarily on arguments around economic efficiency understand (see Sect. 3.3). It was also time-consuming to and environmental management by farm types and across explain the extent and intensity metrics. Participants were member state regions. The second iteration of QST was given printed slides at the workshops, which allowed them more ambitious, extending beyond the farm gate into the to flick between guidance on how to read the diagrams and agri-food supply system (see Fig. 2). the results. The discussion may have been impeded by the Multi-dimensional metrics (money and material) were lack of understanding, although there were moments when used in both QST cycles. The metrics were presented by participants suddenly recognised the implications—such as standard farm types (based on the European Farm Account- the focus on intensity of GHG emissions in places like the ancy Data Network typology). The metrics addressed envi- Netherlands neglected the cumulative effect of low GHG ronmental pressures (and where possible, impacts) on soil, emissions over large extents in places like France. There water quality, pollinators and water quantity and GHG emis- was some resistance to the implication that EU agriculture is sions in the first cycle; and this was extended in the second not sustainable from the DG participants, although broadly cycle—for example, differentiating green and blue water. accepted by the NGOs, agency and MEP participants. Many In the first cycle, metrics across EU28 were also produced DG Agri participants wanted to focus on how the CAP was for viability aspects, such as farm productivity, net farm evolving to be more focussed on environmental and social 13 1242 Sustainability Science (2023) 18:1235–1250 protection. Whilst the systemic presentation was often seen research team regarding how to undertake each step of the as intriguing, for most actors, it was not seen as particularly QST cycle meant we informed rather than actively engaged relevant (see Sect. 3.3). the policy actors regarding choices in stages 2, 3 and 4 of both cycles. In the first cycle, a member of DG Agri offered How did the policy actors respond to champion the project with their colleagues, but they left, to the post‑normal nature of the QST? and we did not find a replacement. Engaging beyond DG Agri in the second cycle took time to research multiple DGs The interviews and workshops did not use the language of and other units within the Commission. However, providing post-normal science explicitly, but the main tenets of PNS the opportunity for interaction between DGs (beyond the were addressed in the design and content of the QST cycles. formal Inter-Service Steering Groups convened for specific Uncertainty analysis is extremely important to PNS policy processes) was most appreciated according to feed- approaches, as it reinforces the philosophy of scientific back sheets and interviews. Thus, whilst the transdiscipli- bounded rationality and knowledge as always in the mak- nary teams were not as active as originally hoped, we did ing. However, in both cycles, there was limited time to dis- engage a range of EU level policy actors in both cycles, cuss uncertainty and sensitivity analyses. The metrics were which we believe was a partial success. intended to spark discussion and were presented as partial A normative approach is the final tenet of PNS, recog- and illustrative. Participants seemed to understand that we nising how values and politics are integral to how science were discussing a partial view on complex and adaptive is understood. QST is not about value-neutral presentation socio-ecological systems. of facts but acknowledges the values associated with gen- Unstructured problem finding, ensuring that the nature of erating and interpreting the societal metabolism metrics. the problems is understood and agreed on, is central to PNS. This was probably the most dissonant aspect in the three Whilst there is a growing academic critique of CAP and workshops. The field notes suggest that the general atmos- the EU’s approach to agri-food sustainability, we wanted to phere was polite and reserved, with the discussion focussed explore how policy actors understood these problems before on technical matters of professional expertise rather than discussing how to transform the system using policy levers. the political implications or personal passions. There were Problem framing is important when policy levers can worsen exceptions when individuals grew animated in the work- the systemic problems through not accounting for wider con- shops. The interviews, feedback forms, and follow-up emails sequences. It proved very difficult to get policy actors to revealed some normative views, but these varied from per- discuss systemic problems (see Sect. 3.1). Instead, the delib- son to person. Some individuals still focussed on what they eration focussed on how policy actors were making progress did and how they did it, with little commentary on why it in resolving the problems. In other words, where critiques mattered or how it felt. There was much less dissonance were acknowledged, they were recast as being resolved by with the parliamentarians, who were willing to acknowledge technical solutions. The feedback interviews revealed that how the implications might affect them. We debated many the participants had expected the scientists to present solu- ways of trying to elicit values to complement the stage five tions and confirmed that asking policy actors to work with discussions, such as using a multi-criteria approach. How- scientists to develop a shared understanding of the problem ever, given some participants struggled to understand the was highly unusual. Although we emphasised knowledge method, and debated the metrics, such structured exercises co-creation in the recruitment material, we realised that indi- were inappropriate. Ideally, there would be multiple interac- viduals attended the workshop in passive knowledge recipi- tions to split up the presentation of the results from debating ent mode. In the future iterations, building capacity around the implications for their roles and responsibilities. However, knowledge co-creation is needed. time constraints meant both were addressed in the single Inclusive transdisciplinary teams underlie QST. We workshop. hoped to engage policy actors as co-researchers through- Overall, there was a mixed response to the PNS aspects out multiple cycles. Whilst we did engage some actors of QST. Taking a PNS approach meant our workshops were throughout both cycles (see Sect. 2.3), there was much less very different to previous SPIs. There was some resistance to knowledge co-creation than planned. This was partly due structuring the problem, with the actors either accepting that to their limited availability and different priorities—policy there were major problems and expecting us to have solu- actors were navigating changing policy priorities, from the tions; or arguing that the Commission was already working past EU sustainability strategy to the delivery of the SDGs, on suitable solutions, so they did not want to discuss if the working on the ex-ante impact evaluation of the proposed systemic nature of problems was fully understood. Policy new CAP and delivering the Juncker Commission. Being actors are heterogeneous. There were some engaged actors relevant (see Sect. 3.4) meant targeting extremely busy and who were positive about the potential for QST to assist their focussed actors. The ongoing scientific learning within the work. Feedback interviews revealed that there were some 13 Sustainability Science (2023) 18:1235–1250 1243 actors who wanted more engagement with the researchers. responsibilities. Indeed, some wanted us to explore technical We had followed up requests for information, but as we were aspects of production steps or sequential pathways (Fig. 2), unfamiliar with the individuals, we didn’t understand they which related to specific policy instruments (e.g. support for were signalling a desire for further active science–policy organic farming). Towards the end of the project, contrast- interactions. ing the quantified results from societal metabolism analysis with the results presented under the Common Monitoring How did policy actors respond to the science–policy and Evaluation Framework for CAP was suggested as one interface aspect of the QST? solution. Legitimacy QST aims to make choices explicit at each Here we briefly reflect on how QST relates to the four tenets stage of the cycle. It was difficult to fully explain the quan- of positive SPIs. tification process and impossible to present all the quanti- Credibility The quantification in the QST cycle used soci- fied data in the time-limited workshops, making the QST etal metabolism. This systemic accounting process is not process not fully transparent. The workshops were designed a typical approach and its ontology is opposed to many of to ensure inclusive participation and to encourage a range the conventional policy assessment methodologies such as of values to be expressed. The range of participants from cost–benefit analyses (Giampietro et al. 2013; Saltelli and across different DGs and with different roles was appreci- Giampietro 2017b). It was difficult to explain the methods ated by participants, but the limited discussion did not really and how to interpret the findings in a short time and partici- improve mutual understanding between DGs or between the pants may not have fully understood the purpose of societal researchers and the rest of the ‘transdisciplinary teams’. The metabolism. Lack of familiarity could have reduced partici- presentation of the results as partial positioned all actors as pants’ ability to judge the quality of the evidence. Given the knowledgeable, rather than researchers imparting knowledge constraints on overall workshop length, the presentations to the policy actors, but this did not work particularly well. were highly selective and provided a very partial perspec- The interviews helped us to understand policy processes but tive. In both cycles, participants were unused to having did not involve discussion of the QST results. Whilst more ‘draft’ results presented to them for discussion and improve- one-to-one interactions at stage five would improve mutual ment. Together, these aspects may have undermined the understanding, it would fail to build collective mutual under- credibility of the results, particularly if judged as evidence standing between the DGs. The authors felt a shared prob- for policymaking rather than starting points for co-creating lem framing within the Commission is necessary to support pathways for improved policy. It is unlikely that participants SDG2. would directly challenge the research team’s credibility, but Iterativity QST is evolutionary, building on reflection at the questions about methodology could be read as an oblique each step to deepen relationships and foster improved under- test of our expertise (Waylen et al. 2023). standing. It is not simple repetition but extending contextu- Relevance The authors explicitly selected EU policies as alised knowledge production. Between the first and second the focus for the QST to test how QST might be used to QST cycle, we faced a choice about whether to continue with support policy cycles (Matthews et al. 2020). Participants the same narrative (CAP’s tension between competitiveness found it useful to focus on competitiveness as this was the and public goods) or to broaden the focus. The former would topic of the European Semester (framework for coordi- have allowed further iterations. It may have been more rel- nating economic analyses). Given feedback from the first evant to one or two actors, but it diverted the focus away phase interviews, we switched from the original language of from a systemic approach to sustainability into a specific Water–Energy–Food Nexus to language of the CAP objec- aspect on production steps. Instead, we chose to extend the tives; and participants appreciated the workshop in 2018 QST to look more explicitly at how CAP was enrolled in having been informed by prior interactions with DG Agri. delivering of the UN SDGs, through coherence with other By the second cycle, the Commission were being challenged policies (see Sect. 2.1). This allowed us to build on the first for more detail on their approach to the SDGs; whilst the cycle but to bring in new institutional settings, practices and Green Deal (European Commission 2019) and the Farm- actors beyond DG Agri. This amplified the learning for the to-Fork Strategy (European Commission 2020) were under researchers and provided a greater network of policy actors development. The QST cycles had to support a fast-moving with whom to learn. However, it meant there was less time policy landscape. This situation was challenging given that to consolidate learning about specific policies and actors the specific focus, language and framing of these policies and perhaps less ability to improve our credibility or the were often not visible and had to be accessed through inter- quality of these relationships. Originally, we planned to run views and informed guesswork at the start of each cycle. both QSTs in simulation mode as well as diagnostic mode, The holistic approach in both QST cycles meant that the to consider possible future trajectories. This is another set analyses lacked relevance to individual policy actors with of choices in how and when to iterate. We chose to focus on 13 1244 Sustainability Science (2023) 18:1235–1250 analysis of the current systemic metrics, rather than hypo- targets for the creation of our ‘transdisciplinary teams’. thetical alternatives, as we felt this was more relevant and The approach was also more relevant for agencies, such more credible. However, comparing our experiences with as the European Environment Agency or Eurostat. How- others using QST in the project (Krol et al. 2018; Cabello ever, to focus on these other institutions does not address et al. 2021; Di Felice et al. 2021) albeit with different actors the need to reframe EU polices to address the transforma- and topics, simulations may prove better for building rela- tion required to meet the UN and EU 2030 agenda. These tionships as they do not entail any evaluation of past or cur- observations reinforce the challenge of using QST to address rent policy performance. SDGs with Commission actors, rather than using QST to As with the responses to the systemic and PNS aspects of address a particular issue within a DG and policy domain. QST, our experiences with QST as a SPI were also mixed. Given the importance of iteration, further cycles of QST Our experiences suggest that legitimacy was somewhat in could have been used to build up through the system levels tension with credibility, as trying to co-construct knowl- (see Fig. 2) and beyond the original starting DG, but more edge through a new methodology in a collective setting was slowly than we attempted. The process of using QST for very challenging. These observations reinforce the need for SDG2 was beginning to mature when the project ended. We iterativity to build familiarity with the methodology and to compare our learning points with relevant published experi- build relationships of trust where values and positions can be ences, to discuss the potential of QST to support policy for expressed. In the future, more frequent meetings are needed. sustainability. We were too ambitious in the shift between first and second cycles, which impacted on our ability to be relevant in the right way, with the right actors, at the right time. However, Discussion: trade‑offs and governance we believe QST could be a very useful approach for future challenges SPIs, building on the lessons we have learnt. This paper adds to the growing literature on how SPIs Supporting policy for sustainability involve choices and trade-offs. In particular, there is a wealth of literature reflecting on transdisciplinary research on sus- Policy coherence is central to the delivery of the UN and tainability (Schneider and Buser 2018) and the SDGs (Sch- EU 2030 agendas, requiring both a broad and deep under- neider et al. 2019); as well as advice on how to practise PNS standing of how objectives, instruments and implementation appropriately (Kønig et al. 2017). Our findings acknowledge practices can be supported across policy domains. However, where—for lack of resources, knowledge or foresight—the it was very difficult to assess how policy coherence for the implementation of our QST cycles did not meet these exact- SDGs was practised within the Commission. During the ing requirements. For example, we did not have sufficient first stage of the second cycle, considerable time and effort ongoing relationship building resources (Bergmann et al. were expended on trying to trace how the SDGs were being 2021) and struggled to have the intense involvement required implemented, given there was limited public information on (Brandt et al. 2013). Therefore, our case reflects on imple- how the lists of policies and strategies would be combined. menting QST as a PNS SPI tackling pan-EU policy and we We also struggled to identify how different parts of the highlight areas of struggle as well as success and our reflec- Commission were involved. Whilst it was clear that the Sec- tion may be useful to other researchers embarking on similar retariat General’s department was responsible for the overall processes. implementation of the SDGs, we were unable to get more Using a less complex methodology that did not address detail on these processes. Approaches to the Secretariat multiple simultaneous dimensions might have made it eas- General’s office were redirected to the DGs, yet individual ier for non-scientists to engage with the metrics and their DG staff were responsible for aspects of a single policy. We implications. However, the choice to use societal metabo- were aware of, but unable to access, Inter-Service Steering lism accounting within our PNS SPI responds to critiques Groups such as the group consulted on the proposed new that many analyses of SDGs have limited systemic analy- CAP objectives. It would have been ideal to use QST with ses (Miola et al. 2019). Spending less time on describing such a collective. Instead, we involved some actors from the and implementing the accounting may have allowed more CAP inter-service group, but most actors were very focussed frequent QST iterations and improved the SPI practice on the implications for their own role within a specific policy (Sarkki et al. 2015). Therefore, our case suggested a trade- domain. The fact that cross-DG attendance at seminars was off between highlighting all dimensions of sustainability welcome but unusual suggests that these inter-DG relation- wicked problems, and the tractability of using complex sys- ships are still nascent and not a routine part of the job. tems methodology in transdisciplinary teams where there In fact, the MEPs and NGO actors were the most inter- was limited social capital and low familiarity with societal ested in a systemic overview, yet these were not the original metabolism accounting. We maintain that it is important to 13 Sustainability Science (2023) 18:1235–1250 1245 aspire to a truly systemic approach within SPIs but with philosophy, could have built collective capacity in transdis- hindsight, attempting to tackle SDG2 in the 2nd cycle took ciplinary teams, which could have been harnessed to later resources away from relationship and capacity building pro- acknowledge and address wicked problems (Termeer et al. cesses within DG Agri. 2015). Trying to go inside the ‘policy-making black box’ (Gos- Positive process design cannot always defuse unpalatable sling et al. 2016) was ambitious to attempt QST in a new content. The QST results (Matthews et al. 2017, 2019) that institutional setting where we had no existing networks or found how the current approach by the EU is unlikely to prior experience to help us negotiate knowledge co-produc- achieve SDG2, is echoed in official reports (Eurostat 2019; tion. For example, we read policy-makers’ request to focus European Environment Agency 2020). Likewise, the find- on organics as deflection to an incremental focus on techni- ing that the current framing of the problem is too narrow cal solutions, avoiding a discussion on the systemic prob- to understand and respond to the sustainability challenges, lems with CAP (Kuhmonen 2018). With hindsight, it might is also found elsewhere (Scown and Nicholas 2020). It is have been an attempt to bolster the position for agro-ecol- tempting to blame our challenges faced on the post-political ogy in the new Commission and increase the environmental nature of the European Commission (Wilson and Swynge- ambition of the Farm-to-Fork Strategy. Likewise, we discov- douw 2014), whereby solutions are framed within a neo- ered it is important to keep contacting policy actors despite liberal incremental approach to sustainability. This would their limited interest or opportunity to engage. Approaches imply that the SDGs can be achieved by existing policies like QST are emergent and co-produced by the specific time implemented more efficiently and more effectively, but and context—they are therefore more difficult to predict, without challenging the status quo. It is difficult to balance plan and control (Chilvers and Kearnes 2019) and require a critique of the status quo with a more positive solutions continuous adaptation (Oliver et al. 2021). focus; given true transformation often requires wholescale However, attention to the SPI design and good practice systems change, not incremental policy improvements (Chan is necessary but not sufficient. We identified the policy win- et al. 2020; Dorninger et al. 2020). Our experiences with dows (Rose et al. 2017); but improved design and greater QST may have been difficult as we were asking for a deep determination could not address the problematic governance transformation of the socio-ecological system, which chal- across silos (Ruddy and Hilty 2008; Candel and Biesbroek lenges the status quo (Ojha et al. 2022). Non-state actors 2016). Our experience—that SDGs were of interest to every- may be more willing to discuss the politics of the problems one, but the responsibility of few—was also highlighted by and embrace the challenge (Linnenluecke et al. 2017). others (European Parliament 2019; Montéville and Kettunen QST asked policy actors to go beyond commenting on the 2019). This is a common governance challenge, whereby credibility and relevance of system knowledge (Schneider the more holistic the issues, the more diffuse, contested and and Buser 2018), to reveal the unstated play of the policy- invisible the processes of accountability become (Bevir making game, rather than the formally recorded rules of 2011; Kraft and Wolf 2018). It is difficult to be relevant the game. Our process emphasised the political dimensions when few have a remit for tackling wicked problems [also and choices involved in the framing of societal challenges. highlighted when working with Swedish policy-makers Highlight the politics risks delegitimising the presentation (Höjer et al. 2011)]. Having a policy entrepreneur to advo- of policy-making as rational-, linear-, and evidence-based cate for change would have helped identify systemic cham- (Cairney 2017, 2022). If our results are ‘uncomfortable pions to whom the societal metabolism metrics were most knowledge’ (Rayner 2012) then QST becomes an uncom- relevant (Timmermans et al. 2014). fortable process. Being willing to participate in an uncom- Many other SPIs use future methodologies to explore fortable process requires commitment by these policy actors plausible but hypothetical pathways to sustainability to sup- when there is little institutional support for doing so (Maas port policy development and implementation (Höjer et al. et al. 2022). It is possible therefore to reframe our limited 2011; Carlsson et al. 2015). Indeed sustainability science engagements, particularly with those actors who stuck with often uses science to tackle urgent societal challenges; with the process over 3 years, as quite a success. SPIs co-producing solutions (Bergmann et al. 2021; Miller Despite many warnings that attempting to do QST on et al. 2013). However, our PNS approach asked policy-mak- SDGs within the Commission DGs was impossible, we have ers to (re)consider how current policy choices were gen- achieved some small steps forward. The quantified findings erating unsustainable trends, across multiple domains and provide a resource, with many others, to help shift how the spatial scales throughout the EU. Far from offering solu- Commission addresses SDG2, for example to reframe the tions, QST tries to reframe the debate away from incremental Farm Accountancy Data Network to the Farm Sustainability progress towards acceptance that more radical change was Data Network (European Commission 2021) and to address needed (Saltelli and Giampietro 2017a). However, starting impacts on non-EU biodiversity when considering the sus- with potential solutions, whilst somewhat contrary to PNS tainability of intensive livestock rearing (Cadillo-Benalcazar 13 1246 Sustainability Science (2023) 18:1235–1250 et al. 2020). Echoing other findings (Guimarães Pereira and As existing SPI literature highlights, QST should be Saltelli 2017), it is clear that some institutional changes with guided by a policy entrepreneur and seek inputs from a the Commission are required that make space for plurali- diverse ‘transdisciplinary team’. Users should expect to ties of science practices and encourage greater reflexivity by expend significant time on decisions around selecting and both scientists and policy actors. We need engaged policy visualising metrics, combined with additional time to reflect actors in the creation and not only consumption of sustain- on the effects of these decisions. Using QST requires not ability knowledge (van Hulst and Yanow 2016); and actors only a strong academic team with multiple scientific special- willing to accept the breadth and scale of wicked problems. isms but also wider transdisciplinary skills, such as relation- ship maintenance, political sensitivity, determination, and confidence. Conclusion: using quantitative story‑telling However, using QST also has implications for non-aca- with EU policy actors demic members of QST transdisciplinary teams (Stokols 2006). QST is a means to an end—a way to highlight the This paper summarises the learning points derived from need to act more urgently on sustainability problems—and using two cycles of a post-normal science (PNS) methodol- researchers cannot enact sustainability transformations ogy, quantitative story-telling, with EU policy actors. Each alone. Therefore, QST requires actors able and willing to QST cycle covered issue framing, running societal metabo- enact change, beyond incremental technical improvements to lism analyses, selecting and presenting metrics and delib- policy. Individual policy actors therefore need to step out of erating over what the metrics mean for policy actors (see their bureaucratic remits and explore challenging and uncer- Sect. 2.2). The first cycle focussed on a potential tension tain topics. Although the current climate, biodiversity, and between EU agricultural competitiveness and provision of COVID-19 crises require transformation to resolve them in public goods; and the second cycle on delivery of SDG2 the longer term, they are also immediate priorities for policy (zero hunger). Operationalising QST with busy EU policy actors, further reducing their capacity and appetite for more actors and using science–policy interfaces (SPIs) with PNS apparently esoteric analyses. systemic approaches was challenging (see Sects. 3.2, 3.3, We offer this paper as a contribution to the growing body 3.4). We achieved a partial success in stimulating new think- of literature that reflects honestly on implementing SPIs in ing about policies for SDG2 by Europe (see Sect. 3.5). practice. We document what has been learnt when using Some shortcomings could be overcome, by having more a PNS approach at a pan-EU scale. Our observations add resources, experience and familiarity with the context and to those from other QST research (Cabello et al. 2021; Di actors—to better prepare participants about what to expect Felice et al. 2021) and we would encourage further research from PNS approaches; build capacity in understanding soci- using QST as a PNS SPI approach. It would be interesting etal metabolism and visualisations; and hold more frequent to see if others fare better when using QST for SDGs; or if iterations, building on prior QST cycles. However, other QST has more traction with other types of actors (activists, shortcomings remain more challenging—such as trying to citizens) or when focussed on a specific policy. Our expe- move from incremental business as usual improvements rience also adds to wider scholarship on taking a critical towards a more fundamental transformation and identifying approach to SPIs. Such research will strengthen a post-nor- the actors to whom such a proposition is relevant. Like all mal response to sustainability transformation that is sorely SPIs, QST requires a strongly relational approach, which needed. asks for policy actors’ time, energy and trust. Unlike many Acknowledgements The Horizon 2020 MAGIC project (2016-2020) SPIs, this is required whilst challenging many of the formal received funding from the European Union’s Horizon 2020 Research institutions on which the Commission’s authority is based. and Innovation Programme under Grant agreement No. 689669. Whether or not future academics adopt the societal Authors KB, KAW, KBM, AJB, DGM, AH and DWJ acknowledge metabolism accounting approach (MuSIASEM), we pro- the underpinning funding from the Rural & Environment Science & Analytical Services Division of the Scottish Government. Author MG pose the QST cycle as a useful framework for PNS SPIs. acknowledges financial support by the Spanish Ministry of Science Although challenging to implement, QST is rare in balanc- and Innovation (MICINN) through the “María de Maeztu” program ing multiple perspectives on problem framing with delib- for Units of Excellence (CEX2019-000940-M). This work reflects the eration over integrated, holistic and cross-scale quantitative authors’ view only; the funding agency(ies) is(are) not responsible for any use that may be made of the information it contains. assessment of current situations. Explicit attention to prob- lem framing before seeking solutions is fundamental to a Data availability Many of the datasets generated by the H2020 MAGIC PNS and deep sustainability. Potentially, the challenging the project used for this article are available in the Zenodo repository: problem framing of current policies could be combined with https://​zenodo.​org/​commu​nities/​magic/​search?​page=​1&​size=​20&​ type=​datas​et. However, some of the datasets analysed for this article other solution- or future-orientated SPI approaches that are (e.g. the interview and seminar transcripts) are not publicly available more attractive to policy actors. 13 Sustainability Science (2023) 18:1235–1250 1247 due to participants being promised anonymity and confidentiality as Cairney P (2022) The myth of ‘evidence-based policymaking’ in a part of their informed consent. When all information that could identify decentred state. Public Policy Adm 37(1):46–66. https://​doi.​org/​ the individual or their organisation is removed from these transcripts, 10.​1177/​09520​76720​905016 they no longer have much value as research data, as the information Candel JJL, Biesbroek R (2016) Toward a processual understanding is context specific. of policy integration. Policy Sci 49(3):211–231. https://​doi.​org/​ 10.​1007/​s11077-​016-​9248-y Open Access This article is licensed under a Creative Commons Attri- Candel JJL, Breeman GE, Stiller SJ, Termeer CJAM (2014) Disen- bution 4.0 International License, which permits use, sharing, adapta- tangling the consensus frame of food security: the case of the tion, distribution and reproduction in any medium or format, as long EU Common Agricultural Policy reform debate. Food Policy as you give appropriate credit to the original author(s) and the source, 44:47–58. https://​doi.​org/​10.​1016/j.​foodp​ol.​2013.​10.​005 provide a link to the Creative Commons licence, and indicate if changes Carlsson J, Eriksson LO, Öhman K, Nordström E-M (2015) Com- were made. The images or other third party material in this article are bining scientific and stakeholder knowledge in future scenario included in the article's Creative Commons licence, unless indicated development—a forest landscape case study in northern Sweden. otherwise in a credit line to the material. If material is not included in For Policy Econ 61:122–134. https://​doi.​org/​10.​1016/j.​forpol.​ the article's Creative Commons licence and your intended use is not 2015.​08.​008 permitted by statutory regulation or exceeds the permitted use, you will Cash D, Clark W, Alcock F, Dickson N, Eckley N, Jager J (2002) Sali- need to obtain permission directly from the copyright holder. To view a ence, credibility, legitimacy and boundaries: linking research, copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. assessment and decision making (RWP02-046) Chan KMA, Boyd DR, Gould RK, Jetzkowitz J, Liu J, Muraca B, Naidoo R, Olmsted P, Satterfield T, Selomane O, Singh GG, Sumaila R, Ngo HT, Boedhihartono AK, Agard J, de Aguiar References APD, Armenteras D, Balint L, Barrington-Leigh C, Cheung WWL, Díaz S, Driscoll J, Esler K, Eyster H, Gregr EJ, Hashi- Ainscough J, Wilson M, Kenter JO (2018) Ecosystem services as a moto S, Hernández Pedraza GC, Hickler T, Kok M, Lazarova T, post-normal field of science. Ecosyst Serv 31:93–101. https://​ Mohamed AAA, Murray-Hudson M, O’Farrell P, Palomo I, Say- doi.​org/​10.​1016/j.​ecoser.​2018.​03.​021 sel AK, Seppelt R, Settele J, Strassburg B, Xue D, Brondízio ES Allen TFH, Giampietro M (2016) Narratives and transdisciplines for (2020) Levers and leverage points for pathways to sustainability. a post-industrial world. Syst Res Behav Sci 23:595–615. https://​ People Nat 2(3):693–717. https://​doi.​org/​10.​1002/​pan3.​10124 doi.​org/​10.​1002/​sres.​792 Chilvers J, Kearnes M (2019) Remaking participation in science and Alonso-Yanez G, House-Peters L, Garcia-Cartagena M, Bonelli S, democracy. Sci Technol Hum Values 45(3):347–380. https://​doi.​ Lorenzo-Arana I, Ohira M (2019) Mobilizing transdisciplinary org/​10.​1177/​01622​43919​850885 collaborations: collective reflections on decentering academia Clark WC, Dickson NM (2003) Sustainability science: the emerg- in knowledge production. Glob Sustain 2:e5. https://​doi.​org/​10.​ ing research program. Proc Natl Acad Sci 100(14):8059–8061. 1017/​sus.​2019.2 https://​doi.​org/​10.​1073/​pnas.​12313​33100 Bergmann M, Schäpke N, Marg O, Stelzer F, Lang DJ, Bossert M, Gan- Cornell S, Berkhout F, Tuinstra W, Tàbara JD, Jäger J, Chabay I, de tert M, Häußler E, Marquardt E, Piontek FM, Potthast T, Rho- Wit B, Langlais R, Mills D, Moll P, Otto IM, Petersen A, Pohl C, dius R, Rudolph M, Ruddat M, Seebacher A, Sußmann N (2021) van Kerkhoff L (2013) Opening up knowledge systems for better Transdisciplinary sustainability research in real-world labs: suc- responses to global environmental change. Environ Sci Policy cess factors and methods for change. Sustain Sci 16(2):541–564. 28:60–70. https://​doi.​org/​10.​1016/j.​envsci.​2012.​11.​008 https://​doi.​org/​10.​1007/​s11625-​020-​00886-8 Cortes Arevalo VJ, Verbrugge LNH, Sools A, Brugnach M, Wolterink Bevir M (2011) Governance and governmentality after neoliberalism. R, van Denderen RP, Candel JHJ, Hulscher SJMH (2020) Sto- Policy Polit 39(4):457–471. https://d​ oi.o​ rg/1​ 0.1​ 332/0​ 30557​ 310X​ rylines for practice: a visual storytelling approach to strengthen 550141 the science–practice interface. Sustain Sci. https://​doi.​org/​10.​ Blühdorn I (2011) The politics of unsustainability: COP15, Post-Ecol- 1007/​s11625-​020-​00793-y ogism, and the Ecological Paradox. Organ Environ 24(1):34–53. Cundill G, Harvey B, Tebboth M, Cochrane L, Currie-Alder B, Vincent https://​doi.​org/​10.​1177/​10860​26611​402008 K, Lawn J, Nicholls RJ, Scodanibbio L, Prakash A, New M, Brandt P, Ernst A, Gralla F, Luederitz C, Lang DJ, Newig J, Reinert F, Wester P, Leone M, Morchain D, Ludi E, DeMaria-Kinney J, Abson DJ, von Wehrden H (2013) A review of transdisciplinary Khan A, Landry ME (2019) Large-scale transdisciplinary col- research in sustainability science. Ecol Econ 92:1–15. https://d​ oi.​ laboration for adaptation research: challenges and insights. Glob org/​10.​1016/j.​ecole​con.​2013.​04.​008 Chall. https://​doi.​org/​10.​1002/​gch2.​20170​0132 Cabello V, Romero D, Musicki A, Guimarães Pereira Â, Peñate B Di Felice LJ, Cabello V, Ripa M, Madrid-Lopez C (2021) Quantita- (2021) Co-creating narratives for WEF nexus governance: tive storytelling: science, narratives, and uncertainty in Nexus a quantitative story-telling case study in the Canary Islands. innovations. Sci Technol Hum Values. https://​doi.​org/​10.​1177/​ Sustain Sci 16(4):1363–1374. https:// ​ d oi. ​ o rg/ ​ 1 0. ​ 1 007/​ 01622​43921​10538​19 s11625-​021-​00933-y Dorninger C, Abson DJ, Apetrei CI, Derwort P, Ives CD, Klaniecki Cadillo-Benalcazar JJ, Renner A, Giampietro M (2020) A multiscale K, Lam DPM, Langsenlehner M, Riechers M, Spittler N, von integrated analysis of the factors characterizing the sustainability Wehrden H (2020) Leverage points for sustainability transfor- of food systems in Europe. J Environ Manag 271:12. https://​doi.​ mation: a review on interventions in food and energy systems. org/​10.​1016/j.​jenvm​an.​2020.​110944 Ecol Econ 171:106570. https://​doi.​org/​10.1​ 016/j.​ecole​con.​2019.​ Cairney P (2012) Complexity theory in political science and public 106570 policy. Political Stud Rev 10(3):346–358. https://​doi.​org/​10.​ European Commission (2019) Communication From the Commission 1111/j.​1478-​9302.​2012.​00270.x to the European Parliament, the European Council, the Council, Cairney P (2017) Evidence-based best practice is more political than The European Economic and Social Committee and the Com- it looks: a case study of the “Scottish Approach.” Evid Policy mittee of the Regions: The European Green Deal. COM (2019) 13(3):499–515. https://d​ oi.o​ rg/1​ 0.1​ 332/1​ 74426​ 416X1​ 46092​ 6156​ 640 final, 24 5901 13 1248 Sustainability Science (2023) 18:1235–1250 European Commission (2020) Communication from the Commission Hoppe R (2018) Rules-of-thumb for problem-structuring policy design. to the European Parliament, the Council, the European Economic Policy Des Pract 1(1):12–29. https://​doi.​org/​10.​1080/​25741​292.​ and Social Committee and the Committee of the Regions. A 2018.​14274​19 Farm to Fork Strategy for fair, healthy and environmentally- Jagannathan K, Arnott JC, Wyborn C, Klenk N, Mach KJ, Moss RH, friendly food system Sjostrom KD (2020) Great expectations? Reconciling the aspira- European Commission (2021) Conversion of FADN to a Farm Sustain- tion, outcome, and possibility of co-production. Curr Opin Envi- ability Data Network (FSDN): Roadmap Consultation. https://​ ron Sustain 42:22–29. https://​doi.​org/​10.​1016/j.​cosust.​2019.​11.​ ec.e​ uropa.e​ u/i​ nfo/l​ aw/b​ etter-r​ egula​ tion/h​ ave-y​ our-s​ ay/i​ nitia​ tives/​ 010 12951-​Conve​rsion-​to-a-​Farm-​Susta​inabi​lity-​Data-​Netwo​rk-​ Jasanoff S, Wynne B (1998) Science and decision making. In: Rayner FSDN-_​en. Accessed 20 Oct 2021 HS, Malone EL (eds) Human choice and climate change, the European Environment Agency (2020) The European environment— societal framework, vol 1. Battelle Press, Columbus, pp 1–87 state and outlook 2020: knowledge for transition to a sustainable (Reprinted from: ordered 29/11) Europe Karpińska A (2018) Post-normal science. The escape of science: from European Parliament (2019) Europe's approach to implementing the truth to quality? Soc Epistemol 32(5):338–350. https://​doi.​org/​ Sustainable Development Goals: good practices and the way for- 10.​1080/​02691​728.​2018.​15311​57 ward. https://​www.​europ​arl.​europa.​eu/​cmsda​ta/​160360/​DEVE%​ Karpouzoglou T, Dewulf A, Clark J (2016) Advancing adaptive gov- 20stu​dy%​20on%​20EU%​20SDG%​20imp​lemen​tation%​20for​mat- ernance of social-ecological systems through theoretical mul- ted.​pdf. Accessed 3 Apr 2019 tiplicity. Environ Sci Policy 57:1–9. https://​doi.​org/​10.​1016/j.​ Eurostat (2019) Sustainable development in the European Union: moni- envsci.​2015.​11.​011 toring report on progress towards the SDGS in an EU context Koff H, Maganda C (2016) The EU and The human right to water Funtowicz SO, Ravetz JR (1993) Science for the post-normal age. and sanitation: normative coherence as the key to transformative Futures 25(7):739–755. https://​doi.​org/​10.​1016/​0016-​3287(93)​ development. Eur J Dev Res 28(1):91–110. https://​doi.​org/​10.​ 90022-L 1057/​ejdr.​2015.​77 Giampietro M, Mayumi K (2000) Multiple-scale integrated assessment Kok K, Bärlund I, Flörke M, Holman I, Gramberger M, Sendzimir of societal metabolism: introducing the approach. Popul Environ J, Stuch B, Zellmer K (2015) European participatory scenario 22(2):109–153. https://​doi.​org/​10.​1023/A:​10266​91623​300 development: strengthening the link between stories and mod- Giampietro M, Allen TFH, Mayumi K (2006) The epistemological els. Clim Change 128(3):187–200. https://​doi.​org/​10.​1007/​ predicament associated with purposive quantitative analysis. s10584-​014-​1143-y Ecol Complex 3(4):307–327. https://​doi.​org/​10.​1016/j.​ecocom.​ Kønig N, Børsen T, Emmeche C (2017) The ethos of post-normal sci- 2007.​02.​005 ence. Futures 91:12–24. https://​doi.​org/​10.​1016/j.​futur​es.​2016.​ Giampietro M, Aspinall RJ, Bukkens SGF, Cadillo Benalcazar J, 12.​004 Diaz-Maurin F, Flammini A, Gomiero T, Kovacic Z, Madrid C, Kovacic Z (2018) Conceptualizing numbers at the science–policy inter- Ramos-Martín J, Serrano-Tovar T (2013) An innovative account- face. Sci Technol Hum Values. https://​doi.​org/​10.​1177/​01622​ ing framework for the food-energy-water Nexus. Application of 43918​770734 the MuSIASEM approach to three case studies. http://​www.​fao.​ Kraft B, Wolf S (2018) Through the lens of accountability: analyz- org/​docrep/​019/​i3468e/​i3468e.​pdf. Accessed 21 June 2016 ing legitimacy in environmental governance. Organ Environ Giampietro M, Aria M, Cabello V, Cadillo Benalcazar J, D'Ambrosio 31(1):70–92. https://​doi.​org/​10.​1177/​10860​26616​680682 A, de la Fuente JA, Di Felice L, Iorio C, Zovacic Z, Krol M, Krol MS, Cabello Villarejo V, Cadillo-Benalcazar J, de Olde E, di Madrid C, Matthews K, Miller D, Musicki Savic A, Pandolfo G, Felice L, Giampietro M, Muscat A, Renner A, Ripa M, Rip- Penate Suarez B, Renner A, Ripa M, Ripoll Bosch R, Serrano- oll Bosch R, Serrano-Tovar T, Verburg CCA (2018) Report on Tovar T, Siciliano R, Staiano M, Velasco Fernandez R (2017) exploratory applications of the MuSIASEM Toolbox in Quantita- Report on Nexus Security using Quantitative Story-Telling. tive Story Telling for anticipation. MAGIC (H2020-GA 689669), MAGIC (H2020-GA 689669), Project Deliverable 4.1. https://​ Project Deliverable 4.3. https://​magic-​nexus.​eu/​docum​ents/​d43-​ magic-​nexus.​eu/​docum​ents/​d41-​report-​nexus-​secur ​ity-​using-​ report-​explo​ratory-​appli​catio​ns-​musia​sem-​toolb​ox-​quant​itati​ve-​ quant​itati​ve-​story-​telli​ng. Accessed 12 Jan 2018 story-​telli​ng. Accessed 21 May 2019 Gossling S, Cohen SA, Hares A (2016) Inside the black box: EU policy Krzywinski MI, Schein JE, Birol I, Connors J, Gascoyne R, Horsman officers’ perspectives on transport and climate change mitiga- D, Jones SJ, Marra MA (2009) Circos: an information aesthetic tion. J Transp Geogr 57:83–93. https://​doi.​org/​10.​1016/j.​jtran​ for comparative genomics. Genome Res. https://d​ oi.o​ rg/1​ 0.1​ 101/​ geo.​2016.​10.​002 gr.​092759.​109 Guimarães Pereira Â, Saltelli A (2017) Post-normal institutional iden- Kuhmonen T (2018) Systems view of future of wicked problems to be tities: quality assurance, reflexivity and ethos of care. Futures addressed by the common agricultural policy. Land Use Policy 91:53–61. https://​doi.​org/​10.​1016/j.​futur​es.​2016.​11.​009 77:683–695. https://​doi.​org/​10.​1016/j.​landu​sepol.​2018.​06.​004 Hisschemöller M, Hoppe R (2001) Coping with intractable controver- Linnenluecke MK, Verreynne M-L, de Villiers Scheepers MJ, Ven- sies: the case for problem structuring in policy design and analy- ter C (2017) A review of collaborative planning approaches for sis. In: Hisschemöller M, Dunn WN, Ravetz JR (eds) Knowledge, transformative change towards a sustainable future. J Clean Prod power, and participation in environmental policy analysis. Trans- 142:3212–3224. https://​doi.​org/​10.​1016/j.​jclep​ro.​2016.​10.​148 action Publishers, London, pp 47–72 Maas TY, Pauwelussen A, Turnhout E (2022) Co-producing the Hodge I, Hauck J, Bonn A (2015) The alignment of agricultural and science–policy interface: towards common but differentiated nature conservation policies in the European Union. Conserv responsibilities. Humanit Soc Sci Commun 9(1):93. https://​doi.​ Biol 29(4):996–1005. https://​doi.​org/​10.​1111/​cobi.​12531 org/​10.​1057/​s41599-​022-​01108-5 Höjer M, Dreborg KH, Engström R, Gunnarsson-Östling U, Svenfelt Matthews KB, Blackstock KL, Rivington M, Waylen KA, Miller DG, Å (2011) Experiences of the development and use of scenarios Wardell-Johnson D, Kovacic Z, Renner A, Ripa M, Giampi- for evaluating Swedish environmental quality objectives. Futures etro M (2017) Delivering more than the “Sum of the Parts”: 43(4):498–512. https://​doi.​org/​10.​1016/j.​futur​es.​2011.​02.​003 using Quantitative Storytelling to address the challenges of conducting science for policy in the EU land, water and energy nexus. In: Syme G, Hatton MacDonald FDB, Piantadosi J (eds) 13 Sustainability Science (2023) 18:1235–1250 1249 MODSIM2017, 22nd International Congress on Modelling and in sustainability research. Nat Sustain. https://​doi.​org/​10.​1038/​ Simulation. Modelling and Simulation Society of Australia and s41893-​019-​0448-2 New Zealand, pp 7. Online: https://​www.​mssanz.​org.​au/​modsi​ Ojha H, Nightingale AJ, Gonda N, Muok BO, Eriksen S, Khatri D, m2017/​Keyno​te/​matth​ews.​pdf. Retrieved from https://​www.​ Paudel D (2022) Transforming environmental governance: criti- mssanz.​org.​au/​modsi​m2017/​Keyno​te/​matth​ews.​pdf cal action intellectuals and their praxis in the field. Sustain Sci. Matthews KB, Blackstock KL, Waylen KA, Juarez-Bourke A, Miller https://​doi.​org/​10.​1007/​s11625-​022-​01108-z DG, Wardell-Johnson DH, Rivington M (2018) Report on the Oliver TH, Benini L, Borja A, Dupont C, Doherty B, Grodzińska- Quality Check of the Robustness of the Narrative behind the Jurczak M, Iglesias A, Jordan A, Kass G, Lung T, Maguire C, Common Agricultural Policy (CAP). https://​magic-​nexus.​eu/​ McGonigle D, Mickwitz P, Spangenberg JH, Tarrason L (2021) docum​ents/​d55-​report-​narra​tives-​behind-​cap. Accessed 21 Nov Knowledge architecture for the wise governance of sustainability 2018 transitions. Environ Sci Policy 126:152–163. https://​doi.​org/​10.​ Matthews KB, Waylen KA, Blackstock KL, Juarez-Bourke A, Miller 1016/j.​envsci.​2021.​09.​025 DG, Wardell-Johnson D, Rivington M, Giampietro M (2019) Palliser A, Dodson G (2017) Uncertainty, complexity and controversy Science for sustainability: using societal metabolism analysis to in dolphin threat management: a role for post-normal science? check the robustness of European Union policy narratives in the Environ Sci Policy 78:74–80. https://​doi.​org/​10.​1016/j.​envsci.​ water, energy and food nexus. In: Elsawah S (ed) MODSIM2019, 2017.​09.​004 23rd International Congress on Modelling and Simulation. Mod- Partelow S (2015) Coevolving Ostrom’s social–ecological systems elling and Simulation Society of Australia and New Zealand. (SES) framework and sustainability science: four key co-benefits. Canberra, Australia, p 7. https://​doi.​org/​10.​36334/​modsim.​2019.​ Sustain Sci. https://​doi.​org/​10.​1007/​s11625-​015-​0351-3 J5.​matth​ews. Accessed 2 Feb 2022 Pradhan P, Costa L, Rybski D, Lucht W, Kropp JP (2017) A systematic Matthews KB, Blackstock KL, Waylen KA, Juarez-Bourke A, Miller study of sustainable development goal (SDG) interactions. Earths DG, Wardell-Johnson DH, Rivington M, Hague A, Fisher D, Future 5(11):1169–1179. https://d​ oi.o​ rg/1​ 0.1​ 002/2​ 017EF ​ 00063​ 2 Renner A, Cadillo-Benalcazar J, Schyns J, Giampietro M (2020) Rayner S (2012) Uncomfortable knowledge: the social construction Report on EU sustainability goals: insights from Quantitative of ignorance in science and environmental policy discourses. Story Telling and the WEFE nexus. MAGIC (H2020–GA Econ Soc 41(1):107–125. https://​doi.​org/​10.​1080/​03085​147.​ 689669) Project Deliverable 5.1. https://​magic-​nexus.​eu/​docum​ 2011.​637335 ents/​deliv​erable-​51-​report-​eu-​susta​inabi​lity-​goals. Accessed 31 Ripoll Bosch R, Giampietro M, Cabello V, Cadillo Benalcazar J, de July 2020 Boer IJM, di Felice L, Krol M, Matthews K, Miller D, Mus- Matthews KB, Renner A, Blackstock KL, Waylen KA, Miller DG, cat A, de Olde E, Renner A, Ripa M, Serrano-Tovar T, Verburg Wardell-Johnson DH, Juarez-Bourke A, Cadillo-Benalcazar CCA, Wardell-Johnson D (2018) Report on EU socio-ecological J, Schyns JF, Giampietro M (2021) Old wine in new bottles: systems. MAGIC (H2020-GA 689669), Project Deliverable 4.2. exploiting data from the EU’s farm accountancy data network https://​magic-​nexus.​eu/​docum​ents/​d42-​report-​eu-​socio-​ecolo​ for Pan-EU sustainability assessments of agricultural production gical-​syste​ms. Accessed 3 Feb 2019 systems. Sustainability 13(18): 10080. https://​www.​mdpi.​com/​ Rose DC (2018) Avoiding a post-truth world: embracing post-normal 2071-​1050/​13/​18/​10080 conservation. Conserv Soc 16(4):518–524. https://​doi.​org/​10.​ Miller TR, Wiek A, Sarewitz D, Robinson J, Olsson L, Kriebel D, 4103/​cs.​cs_​17_​131 Loorbach D (2013) The future of sustainability science: a solu- Rose DC, Mukherjee N, Simmons BI, Tew ER, Robertson RJ, Vadrot tions-oriented research agenda. Sustain Sci 9(2):239–246. https://​ ABM, Doubleday R, Sutherland WJ (2017) Policy windows doi.​org/​10.​1007/​s11625-​013-​0224-6 for the environment: tips for improving the uptake of scientific Miola A, Borchardt S, Neher F, Buscaglia D (2019) Interlinkages and knowledge. Environ Sci Policy. https://​doi.​org/​10.​1016/j.​envsci.​ policy coherence for the sustainable development goals imple- 2017.​07.​013 mentation: an operational method to identify trade-offs and co- Ruddy TF, Hilty LM (2008) Impact assessment and policy learn- benefits in a systemic way, EUR 29646 EN, Publications Office ing in the European Commission. Environ Impact Assess Rev of the European Union, Luxembourg, 2019, ISBN 978-92-79- 28(2):90–105. https://​doi.​org/​10.​1016/j.​eiar.​2007.​05.​001 99556-9, https://​doi.​org/​10.​2760/​472928.​JRC11​5163 Rüffin N (2020) EU science diplomacy in a contested space of multi- Montéville M, Kettunen M (2019) Assessing and accelerating the EU level governance: ambitions, constraints and options for action. progress on Sustainable Development Goals (SDGs) in 2019, a Res Policy 49(1):103842. https://​doi.​org/​10.​1016/j.​respol.​2019.​ briefing to inform the UN High Level Political Forum (HLPF) 103842 and the SDG Summit in New York (9–18 July and 24–25 Sep- Saltelli A, Giampietro M (2017a) What is wrong with evidence based tember 2019). https://​ieep.​eu/​uploa​ds/​artic​les/​attac​hments/​52fc7​ policy, and how can it be improved? Futures 91:62–71. https://​ 03c-​0404-​4c58-​9819-​9e31c​92d3e​f1/​EU%​20SDG%​20bri​efing%​ doi.​org/​10.​1016/j.​futur​es.​2016.​11.​012 202019_​IEEP.​pdf?v=​63729​716887. Accessed 3 Mar 2020 Saltelli A, Giampietro M (2017b) What is wrong with evidence based Norström AV, Dannenberg A, McCarney G, Milkoreit M, Diekert F, policy, and how can it be improved? Futures 91:9. https://d​ oi.o​ rg/​ Engström G, Fishman R, Gars J, Kyriakopoolou E, Manoussi V, 10.​1016/j.​futur​es.​2016.​11.​012 Meng K, Metian M, Sanctuary M, Schlüter M, Schoon M, Schultz Sarkki S, Tinch R, Niemelä J, Heink U, Waylen K, Timaeus J, Young J, L, Sjöstedt M (2014) Three necessary conditions for establish- Watt A, Neßhöver C, van den Hove S (2015) Adding ‘iterativity’ ing effective sustainable development goals in the anthropocene. to the credibility, relevance, legitimacy: a novel scheme to high- Ecol Soc. https://​doi.​org/​10.​5751/​ES-​06602-​190308 light dynamic aspects of science–policy interfaces. Environ Sci Norström AV, Cvitanovic C, Löf MF, West S, Wyborn C, Balvanera P, Policy 54:505–512. https://d​ oi.o​ rg/1​ 0.1​ 016/j.e​ nvsci.2​ 015.0​ 2.0​ 16 Bednarek AT, Bennett EM, Biggs R, de Bremond A, Campbell Sarkki S, Balian E, Heink U, Keune H, Nesshöver C, Niemelä J, Tinch BM, Canadell JG, Carpenter SR, Folke C, Fulton EA, Gaffney O, R, Van Den Hove S, Watt A, Waylen KA, Young JC (2019) Man- Gelcich S, Jouffray J-B, Leach M, Le Tissier M, Martín-López aging science–policy interfaces for impact: interactions within B, Louder E, Loutre M-F, Meadow AM, Nagendra H, Payne D, the environmental governance meshwork. Environ Sci Policy. Peterson GD, Reyers B, Scholes R, Speranza CI, Spierenburg https://​doi.​org/​10.​1016/j.​envsci.​2019.​05.​011 M, Stafford-Smith M, Tengö M, van der Hel S, van Putten I, Österblom H (2020) Principles for knowledge co-production 13 1250 Sustainability Science (2023) 18:1235–1250 Schneider F, Buser T (2018) Promising degrees of stakeholder inter- Timmermans J, van der Heiden S, Born MP (2014) Policy entrepre- action in research for sustainable development. Sustain Sci neurs in sustainability transitions: their personality and lead- 13(1):129–142. https://​doi.​org/​10.​1007/​s11625-​017-​0507-4 ership profiles assessed. Environ Innov Soc Trans 13:96–108. Schneider F, Kläy A, Zimmermann AB, Buser T, Ingalls M, Messerli P https://​doi.​org/​10.​1016/j.​eist.​2014.​06.​002 (2019) How can science support the 2030 Agenda for Sustainable Tsukahara T (2017) Commentary: new currents in science: the chal- Development? Four tasks to tackle the normative dimension of lenge of quality, examining the discrepancies and incongruities sustainability. Sustain Sci 14(6):1593–1604. https://​doi.​org/​10.​ between Japanese techno-scientific policy and the citizens’ sci- 1007/​s11625-​019-​00675-y ence movement in post-3/11 Japan. Futures 91:84–89. https://​ Schout A, Jordan A (2005) Coordinated European Governance: self- doi.​org/​10.​1016/j.​futur​es.​2017.​04.​008 organizing or centrally steered? Public Adm 83(1):201–220. Turnpenny J, Jones M, Lorenzoni I (2011) Where now for post-normal https://​doi.​org/​10.​1111/j.​0033-​3298.​2005.​00444.x science? A critical review of its development, definitions, and Scown MW, Nicholas KA (2020) European agricultural policy requires uses. Sci Technol Hum Values 36(3): 287–306. http://w ​ ww.j​ stor.​ a stronger performance framework to achieve the sustainable org/​stable/​41149​056 development goals. Global Sustain 3:e11. https://​doi.​org/​10.​ van den Hove S (2007) A rationale for science–policy interfaces. 1017/​sus.​2020.5 Futures 39(7):807–826. https://​doi.​org/​10.​1016/j.​futur​es.​2006.​ Scown MW, Winkler KJ, Nicholas KA (2019) Aligning research with 12.​004 policy and practice for sustainable agricultural land systems in van Hulst M, Yanow D (2016) From policy “Frames” to “Framing”: Europe. Proc Natl Acad Sci USA 116(11):4911–4916. https://​ theorizing a more dynamic, political approach. Am Rev Public doi.​org/​10.​1073/​pnas.​18121​00116 Adm 46(1):92–112. https://​doi.​org/​10.​1177/​02750​74014​53314 Simon D, Schiemer F (2015) Crossing boundaries: complex systems, Voelker T, Blackstock K, Kovacic Z, Sindt J, Strand R, Waylen K transdisciplinarity and applied impact agendas. Curr Opin Envi- (2019) The role of metrics in the governance of the water-energy- ron Sustain 12:6–11. https://​doi.​org/​10.​1016/j.​cosust.​2014.​08.​ food nexus within the European Commission. J Rural Stud. 007 https://​doi.​org/​10.​1016/j.​jrurs​tud.​2019.​08.​001 Stokols D (2006) Toward a science of transdisciplinary action research. Waylen KA, Blackstock KL, Matthews KB, Juarez Bourke A, Hague Am J Community Psychol 38(1–2):79–93. https://​doi.​org/​10.​ A, Wardell-Johnson DH, Miller DG, Zovacic Z, Voelker T, Gui- 1073/​pnas.​18121​00116 marães Pereira A, Giampietro M (2023) Post-normal science in Sutherland WJ, Bellingan L, Bellingham JR, Blackstock JJ, Bloomfield practice: Reflections from scientific experts working on the Euro- RM, Bravo M, Cadman VM, Cleevely DD, Clements A, Cohen pean agri-food policy nexus Environmental Science & Policy AS, Cope DR, Daemmrich AA, Devecchi C, Anadon LD, Dene- (Accepted) gri S, Doubleday R, Dusic NR, Evans RJ, Feng WY, Godfray Wesselink A, Buchanan KS, Georgiadou Y, Turnhout E (2013) Tech- HCJ, Harris P, Hartley SE, Hester AJ, Holmes J, Hughes A, nical knowledge, discursive spaces and politics at the science– Hulme M, Irwin C, Jennings RC, Kass GS, Littlejohns P, Mar- policy interface. Environ Sci Policy 30:1–9. https://​doi.​org/​10.​ teau TM, McKee G, Millstone EP, Nuttall WJ, Owens S, Parker 1016/j.​envsci.​2012.​12.​008 MM, Pearson S, Petts J, Ploszek R, Pullin AS, Reid G, Richards Wilson J, Swyngedouw E (2014) The post-political and its discontents KS, Robinson JG, Shaxson L, Sierra L, Smith BG, Spiegelhal- spaces of depoliticisation, spectres of radical politics. Edinburgh ter DJ, Stilgoe J, Stirling A, Tyler CP, Winickoff DE, Zimmern University Press, Edinburgh RL (2012) A collaboratively-derived science-policy research Young JC, Waylen KA, Sarkki S, Albon S, Bainbridge I, Balian E, agenda. PLoS One 7(3):e31824. https://​doi.​org/​10.​1371/​journ​ Davidson J, Edwards D, Fairley R, Margerison C, McCracken al.​pone.​00318​24 D, Owen R, Quine CP, Stewart-Roper C, Thompson D, Tinch Swart RJ, Raskin P, Robinson J (2004) The problem of the future: sus- R, Van den Hove S, Watt A (2014) Improving the science-pol- tainability science and scenario analysis. Glob Environ Change icy dialogue to meet the challenges of biodiversity conserva- 14(2):137–146. https://​doi.​org/​10.​1016/j.​gloen​vcha.​2003.​10.​002 tion: having conversations rather than talking at one-another. Szopik-Depczynska K, Kedzierska-Szczepaniak A, Szczepaniak K, Biodivers Conserv 23(2):387–404. https://​doi.​org/​10.​1007/​ Cheba K, Gajda W, Ioppolo G (2018) Innovation in sustainable s10531-​013-​0607-0 development: an investigation of the EU context using 2030 agenda indicators. Land Use Policy 79:251–262. https://​doi.​org/​ Publisher's Note Springer Nature remains neutral with regard to 10.​1016/j.​landu​sepol.​2018.​08.​004 jurisdictional claims in published maps and institutional affiliations. Termeer CJAM, Dewulf A, Breeman G, Stiller SJ (2015) Governance capabilities for dealing wisely with wicked problems. Adm Soc 47(6):680–710. https://​doi.​org/​10.​1177/​00953​99712​469195 13