Facilitating cross-sectoral assessments of local climate change vulnerability

Abstract

The production of contextualised assessments of local climate change vulnerability serves to improve their usefulness in urban planning. For this purpose, a cross-sectoral participatory approach combining local and academic knowledge is vital. This study aims to contribute to the understanding of how such assessments can be effectively facilitated. Through the elaboration of a framework for joint knowledge production, the paper develops and applies ex-post evaluation criteria to analyse how the set-up and design of participatory assessments affected the identification of local climate vulnerability in two Swedish urban areas. These cases included a series of researcher-led stakeholder dialogues involving participants from various municipal departments, national agencies and research institutions. The results demonstrate that the project set-up affected the joint knowledge production by unifying relevant competences. However, occasionally, it also created conflict. The design of the dialogues influenced the understanding of local vulnerability by broadening the perspective on risks and opportunities and by creating common visual representations of abstract issues. The paper concludes that when facilitating participatory cross-sectoral vulnerability assessments, the consideration of two aspects is important. First, intermediaries, in the form of maps, interactive techniques and metaphors, can bridge organisational divides if designed with clear and negotiated aims. Second, the project set-up can spur motivation if the initial group composition is substantiated and the process is flexible enough to allow for relevant detours.

Keywords:

• boundary object
• climate change
• cross-sectoral planning
• joint knowledge production
• stakeholder participation
• vulnerability assessment

1. Introduction

Climate change vulnerability assessments are often intended to guide policy-making on issues related to local vulnerability by identifying manageable characteristics of vulnerability (O'Brien et al. 2004; Carter et al. 2007; Schneider et al. 2007). Vulnerability assessments are increasingly becoming interdisciplinary, considering more contextualised variables and management alternatives, and giving greater importance to a system’s capacity to adapt (Füssel & Klein 2006). The involvement of stakeholders representing a broad spectrum of interests is crucial in such assessments to contribute with different expertise and value-based knowledge, thereby providing insight into local planning and policy-making contexts (Larsen & Gunnarsson-Östling 2009; Malone & Engle 2011; da Silva et al. 2012). By involving a representation of several sectors into common deliberations, so-called ‘silo effects’, which involve the neglect of risks and opportunities, may be avoided (Keivani 2010).

However, to make new collaborations among heterogeneous participants effective, and to make outcomes relevant for all involved actors, knowledge needs to be produced at the intersection of different traditions (van Buuren & Edelenbos 2004). A critical challenge in such joint knowledge production efforts is to facilitate common understandings, aims and methods which unite divergent viewpoints and logics (Gieryn 1983; Jasanoff & Wynne 1998; Evans 2008). This demands a ‘reconciliation’ of divergent social worlds and understandings of an issue (Burger et al. 2007). Since integrated climate change vulnerability assessments, in general, and participatory assessments, in particular, are still at an early stage of development, systematic analysis of how to work across such organisational divides (e.g. between various sector departments, research organisations, companies and interest groups) is limited (Soares et al. 2012). Nevertheless, some previous studies have provided insight into important aspects for ‘boundary crossing’ within their individual project outcomes (e.g. Kloprogge & van der Sluijs 2006; Parkins & MacKendrick 2007; de la Vega-Leinert et al. 2008: Fazey et al. 2010; Moser & Ekstrom 2011; Larsen et al. 2012), and more general conditions for success for joint knowledge production (e.g. Lemos & Morehouse 2005; Hegger et al. 2012), which can be utilised to build an understanding of how such fruitful participant interaction can be assisted.

Based on previous literature and results from two analysed assessments, this study aims to deepen the understanding of how participatory cross-sectoral vulnerability assessments can be facilitated. The framework of Hegger et al. (2012) was complemented and elaborated to analyse how the set-up and design of the two recently conducted vulnerability assessments affected the joint production of knowledge among participants. The studied research project, in which these assessments were conducted, aimed to increase the understanding of climate change vulnerability in two Swedish urban areas by involving key stakeholders in cross-sectoral assessments (cf. Jonsson et al. 2012). The analysis of the project was guided by the following two research questions:

1. How did the set-up of the assessments affect interaction among participants, and which factors are crucial to consider?

2. How did the design of deliberative exercises affect the understanding of local vulnerability, and how can the joint production of a deeper understanding be facilitated?

The article is organised as follows: the next section presents the analysed project, the empirical material and the methods used. Section 3 presents the approach to evaluating the joint knowledge production. Section 4 discusses how the production of joint knowledge regarding local climate change vulnerability was facilitated and obstructed by the project set-up and design of deliberative exercises. Section 5 concludes and discusses how cross-sectoral vulnerability assessments more generally can be facilitated.

2. Research approach and method

The object of study in this analysis is a research project¹ that developed and tested methods for participatory integrated vulnerability assessments by involving researchers and local stakeholders in two municipalities along the river Göta älv in southeast Sweden. The area has been identified as being highly exposed to climate change, which affects the management of multiple sectors in the analysed urban environments (SOU 2007). The two municipalities are similar in terms of exposure to climate variability and change, but differ greatly in size (Glaas et al. 2010). The larger municipality, Gothenburg, is the second largest city in Sweden with a population of approximately 500,000 inhabitants and is located at the mouth of the Göta älv river on the Skagerrak coast. Flooding has long been a problem in the municipality due to its low elevation, many proximate rivers and its location on the coast. The smaller municipality, Lilla Edet, has a population of approximately 13,000 inhabitants and is situated 55 km upstream from Gothenburg on the river. The municipality has historically had problems with ground stability along the river bank due to the high amount of quick clay in the ground (Glaas 2013).

The integrated assessment approach applied in the two municipalities considered various drivers of change (i.e. climatic, economic and institutional) and involved representatives from key municipal departments and national agencies (i.e. in the building, water, environment, traffic, planning, energy, parks and nature, and railroad sectors) to broaden the assessment scope. To ensure high stakeholder participation, the project was built around a series of four stakeholder dialogues (Welp et al. 2006) in each municipality (A–D; see Table 1). The dialogues included research-driven deliberative exercises focusing on the exposure, sensitivity and adaptive capacity components of vulnerability in one local development area in each of the two municipalities.

Table 1. Stakeholder dialogues, themes and empirical material (cf. Glaas 2013).

Phase/stakeholder dialogue	Themes of exercises	Empirical material
Start-up		Individual interviews with officials (6) Research application
A	Scanning perceived climate-related risks	Dialogue transcripts
	Selecting relevant development areas
		Individual interviews with officials (6) and researchers (4)
		Group interviews with researchers (2)
B	Assessing exposure to climate change and identify knowledge needed for effective management	Dialogue transcripts
	Assessing exposure to socioeconomic change and identify knowledge needed for effective management
	Identifying factors influencing local adaptive capacity
		Individual interviews with officials (6) Group interviews with researchers (2)
C	Overlay of sensitive locations	Dialogue transcripts
	Identifying measures to increase adaptive capacity	Evaluations
	Identifying key stakeholders for local vulnerability management
		Group interviews with researchers (2)
D	Identifying the sectors, activities and groups most vulnerable to climate change effects	Dialogue transcripts
		Evaluations
	Conclusions regarding actions needed to reduce local climate change vulnerability
		Group interviews with researchers (2)

The stakeholder dialogues aimed to encourage joint knowledge production (cf. Lemos & Morehouse 2005) whereby knowledge is co-produced in collaboration between all invited participants, both municipal officials and researchers. Five to eight officials from the municipalities, plus two to four researchers, were included in each dialogue. The dialogues were facilitated by the researchers, and in most cases mainly officials participated in the actual discussions on the presented topics. However, researchers participated by asking follow-up questions, etc., and were thereby highly involved in the production of knowledge. The dialogues were held over a period of two years plus a build-up phase, during which the dialogues were planned by the researchers and preparatory information was sent to the officials, and a follow-up phase, during which the results were analysed.

Each of the four stakeholder dialogues included two to three exercises covering different aspects of local climate change vulnerability, current local and national approaches to vulnerability management, and potential adaptation measures. In total, the four dialogues included 10 research-driven exercises focusing on the local development projects. To facilitate detailed analysis of the exercise content, each exercise had clear aims which had been agreed upon by the researchers and communicated to the officials (Table 1). Diverse techniques were used to stimulate interaction among participants, ranging from brain-storming regarding climate risks and socioeconomic futures, to participatory mapping to identify sensitive areas. The exercises were designed to enable the exchange of knowledge between participants.

To enable an analysis of the challenges and opportunities of the joint knowledge production process, the stakeholder dialogues were followed up by individual and group interviews and group evaluations with participating officials and researchers. The individual interviews were semi-structured and aimed to map participants' expectations and observations made during the four stakeholder dialogues. The group interviews followed a rather detailed interview guide to focus the discussions on interaction elements and the aims of specific exercises. The group evaluations were held in direct conjunction with dialogue C and D and replaced the follow-up interviews with officials held after dialogue A and B (Table 1). Here, groups of officials were asked to reflect upon the discussions held during the dialogue.

The analysed empirical material is approximately 50 hours in length and is presented in Table 1.

All recorded material was transcribed and analysed following a stepwise coding process (cf. Burnard 1991). First, notes were taken during the interview to document potentially interesting themes raised by participants, which provided a preliminary content scanning. Second, the notes were used to guide the coding of the transcripts into themes. Third, the themes were analysed according to the analytical frame (Section 3) and then grouped into three main topics: (i) expectations of the process and its outcomes, (ii) aspects facilitating or obstructing joint knowledge production and (iii) specific discussions important for producing relevant results.

3. Theoretical departure

Stakeholder involvement has been crucial to assessing additional climate change drivers and to broadening our understanding of specific contexts in vulnerability assessments (Kloprogge & van der Sluijs 2006; Romero-Lankao et al. 2013). Facilitating joint knowledge production in such processes entails creating meaningful interaction among participants who come from different organisations and who are used to different forms of collaboration. This leads to two distinct challenges: first, creating a project set-up which is considered useful and legitimate by all participants (Lemos & Morehouse 2005; Blackstock et al. 2007) and, second, designing interaction in specific parts of the project where knowledge can be jointly produced by overcoming the constraints relating to the participants’ various, and sometimes divergent, viewpoints, interests and backgrounds (Owens et al. 2006; Lövbrand 2007; Friman 2010).

Hegger et al. (2012) have developed a framework for discussing the merits and limitations of such joint knowledge production in participatory climate adaptation research projects and for identifying good practices and enabling factors for social learning. The framework uses a policy arrangement approach as the point of departure and suggests seven conditions for successful joint knowledge production (Table 2).

Table 2. Success conditions for joint knowledge production (Hegger et al. 2012, p. 61).

Success conditions
1. Broadest possible actor coalition within existing limits
2. Shared understanding of goals and problem definitions
3. Recognition of stakeholder perspectives
4. Organised reflection on division of tasks among participating actors
5. Roles of researchers and their knowledge are clear
6. Presence of innovations in reward structures
7. Presence of specific resources, such as boundary objects, facilities, organisational forms and competences

In this article, the framework by Hegger et al. (2012) is used as departure for analysing the joint knowledge production process in the studied project. As presented below, the seven conditions have been grouped into two categories, the first focuses on the project set-up (Section 3.1), while the second focuses on interaction design issues (Section 3.2) following the two research questions and the research challenges presented above. To enable an evaluation of the fulfilment of the success conditions, these categories were complemented by more concrete success criteria derived from previous literature on participatory research and from the STS field. The evaluation criteria are presented in Table 3 (Section 3.3).

Table 3. Evaluation criteria derived from the success conditions of Hegger et al. (2012).

Success condition	Evaluation criteria
1. Participants	Do/did participants have the relevant competences needed to fulfil project aims?
	Can/could all participants voice their opinions?
2. Project goals	Can/could participants influence project goals?
	Are/were the project goals achievable within the project?
4. and 5. Roles	Is/was information and communication sufficiently transparent?
	Are/were the research structure flexible?
3. and 7. Interaction	Do/did materials create common visual representations of issues or places?
	Do/did portrayals create a common space for discussions?
6. Reward structures	Not relevant to this application

3.1. Evaluating the project set-up

The success conditions 1, 2, 4 and 5 by Hegger et al. (2012) relate to how a research project should be set-up to be considered legitimate, credible and salient by all involved stakeholders (cf. Cash et al. 2003). The first success condition concerns participation: who is involved and not, how are participants chosen and how does this affect the joint knowledge production process? A vital aspect of promoting effective interaction is involving participants with competences relevant to the research aims to enhance the understanding of specific phenomena (cf. Stirling 2007). This requires that participants feel that they bring relevant competences to the table, to support project aims and ensure that participants gain from each other’s competences (Schulz et al. 2003). Moreover, equally important for functional interaction is that group dynamics should be free from obvious inhibitory power structures (Blackstock et al. 2007). Power structures in group processes are not explicitly mentioned by Hegger et al. (2012), but are considered to be important for joint knowledge production processes to allow the surfacing of all ideas and to create a sense of shared trust in previous literature (Johnson & Johnson 1997). Thus, the possibility for all participants to voice their opinions is a relevant criterion for analysing the appropriateness of group composition.

The second success condition concerns project goals: what problem(s) do the goals address and how well do the goals relate to participants’ interests? Involving participants early in the project set-up process, for example, when defining the main problems, priorities and goals, is seen as decisive in facilitating interaction even in later project stages (Pain & Francis 2003). As argued by Hisschemöller et al. (2001), ‘people can only meaningfully participate if they can relate to the issue to be explored’ (p. 68). This can be fostered if participants have an opportunity to influence the goal formulation and thereby take ownership of the project, making this an important criterion for evaluating appropriateness of project set-up (Cornwall & Jewkes 1995). According to Larsen et al. (2012), this is important in order to make goals relevant throughout assessments, which is seen as a key to stimulating interaction among various actors involved. This is also in line with the findings of Fazey et al. (2010), who concluded that it is necessary for all participants to join in the analysis of outcomes responding to these goals. If project goals are openly discussed and negotiated, it is easier to create realistic goals that support the fulfilment of all participants’ interests in the project. Thus, to ensure such fulfilment, it is also important that the goals are achievable during the project period in order to be able to follow-up (Blackstock et al. 2007).

The fourth and fifth success conditions concern roles and responsibilities: how open is the process to participant influence, and how clear are the roles among participants? For joint knowledge production to be efficient, the roles of all participants must be adjusted to project aims and participant interests (Blackstock et al. 2007). In exploratory studies, participants can be more passive, while a more active role is required when issues concern, for example, policy formulation (Kloprogge & van der Sluijs 2006). In both cases, however, all participant roles, and the aims of various project parts, must be clearly and honestly discussed in a transparent and continuous two-way dialogue, making transparency of information and communication an important criterion (Glicken 2000; Larsen et al. 2012). Last, there is an argued need to develop a flexible research structure to strike a balance between various participant interests and to enable changes in the project set-up to utilise the various competences involved (Lemos & Morehouse 2005; de la Vega-Leinert et al. 2008).

3.2. Evaluating the interaction design

Success conditions 3 and 7 by Hegger et al. (2012) concern the facilitation of participant interaction in joint knowledge production during specific events. The authors distinguish between two types of artefacts that can be seen as either physical or mental when comparing their meaning with previous STS literature. Both types are described as providing a common space for discussions to assemble mutual trust and motivation by reconciling different epistemologies. In this study, the existence and effects of these intermediaries are used as criteria for measuring the fulfilment of success conditions 3 and 7 (as seen in Table 3).

Physical intermediaries (cf. success condition 7) can be described as material used in the project that has spurred or obstructed joint knowledge production among participants. These relate clearly to the previous literature on boundary objects, sociotechnical objects and intermediary objects. Boundary objects are explained as artefacts or arrangements ‘that allow different social groups to work together without consensus’ (Star 2010, p. 602) by creating a common understanding of a complex issue. Examples include physical prototypes, IT artefacts, design drawings, scenarios and standardised reporting forms (Star & Griesemer 1989; Carlile 2002; Gal et al. 2004; Levina & Vaast 2005; Sundberg 2007). Sociotechnical objects (Toderi et al. 2007) can help make it possible to reconcile divergent meanings and interests in a common social space, enabling the redefinition of interests and the building of shared meanings. Prime examples of such objects are dialogical tools and methods such as participatory GIS, visual aids, scenarios and theatre events (Boujut & Blanco 2003). Intermediary objects have been used to analyse technological and data tools and to mediate communication within and across organisational boundaries; examples include maps, figures, models, management plans and drawings (Thurk & Fine 2003).

Mental intermediaries (cf. success condition 3) can be described as portrayals that have spurred or obstructed joint knowledge production among participants and relate to previous literature on intermediary concepts and hybrid objects. Intermediary concepts are definitions or sets of words that can facilitate the building of collective goals around which participants can unite (Teulier & Hubert 2004). One example is ‘good ecological status’, used as an overarching goal formulation in the European Union’s Water Framework Directive (Steyaert et al. 2007). Hybrid objects, finally, are defined as unitary, real and uncontroversial objects that in practice reflect different framings, but link different worldviews. These are similar to intermediary concepts in that they are not physical objects or techniques, but instead storylines, narratives or metaphors of how, for example, global change is conceptualised as specific value-laden words such as ‘deforestation’ (Forsyth 2003). One example is the storm Gudrun which hit Sweden in 2005, a real event that has transformed into a mental intermediary or a common construct of (climate) vulnerability among a wide range of actors, paving the way for a coordinated adaptation approach (André & Jonsson 2013).

3.3. Applying the evaluation criteria

The elaborated evaluation criteria (Table 3), deriving from the seven success conditions for joint knowledge production, provide a framework that likely can be used in both designing and evaluating local participatory research projects in general. In the following, however, we will use the framework to analyse the joint knowledge production process in the analysed project.

As seen in Table 3, one success condition (number 6, concerning the reward structures of joint knowledge production) is not touched on here since it concerns design issues for knowledge production at the national and, to some extent, international levels, and is normally not within the jurisdiction of individual research projects or programmes.

4. Results and discussion

In this section, results from the two research questions are presented and discussed. Section 4.1 relates to the first research question by discussing the fulfilment of success condition 1, 2, 4 and 5. Section 4.2 relates to research question two by discussing the fulfilment of success condition 3 and 7.

4.1. Project set-up

The group dynamics of the stakeholder dialogues was frequently discussed in the interviews and evaluations, and seems to have largely affected the cooperation among participants both positively and negatively.

4.1.1. Participants

Regarding the first criterion, whether the relevant competences were included, interviewed officials from both groups initially expressed that all key sectors were represented. This seems to have followed a practical view of adaptation seen as principally a technical challenge related to building robust infrastructure. During the course of the project, a broader definition of vulnerability and adaptation was developed. In the smaller municipality, especially the influence of social problems on local climate change vulnerability was highlighted, whereas the influence of regional development for local adaptation was highlighted in the larger municipality. Though a deeper knowledge about these issues was lacking in the two groups, additional competences were not included. A more dynamic group composition, which would aim to invite new group members with additional competences when needed, could supposedly have contributed to a deeper understanding of local vulnerability. Nevertheless, such inclusion can at the same time influence group dynamics among initial members as exemplified occasionally in the project. Thus, to facilitate such decisions, an initial discussion among participants on how to handle a dynamic management of competences should be initiated already in the project planning.

In accordance with the second criterion, whether all participants could voice their opinions, perceptions differed between the two groups. In each municipality, one official in a managerial position in the administration, who had insight into the general issue, was involved in identifying and inviting participants. From a research perspective, this approach was initially perceived as positive since it facilitated the identification of officials involved in related work with different backgrounds. In the larger municipality, several of these invited participants had also previously worked together and were therefore familiar with and confident about the range of competences possessed, which enabled joint knowledge production to start quickly. Although some participants from the small municipality had also cooperated before, they had not cooperated on similar issues, which made them unsure of each other’s competences. These participants, therefore, initially spent time positioning themselves in relation to how they perceived the way in which climate change was being approached in the municipality. These problems decreased over time and were less apparent among officials who had been thoroughly interviewed before the dialogues, which suggests that initial individual consultations with participants before the actual interaction can help facilitate joint knowledge production. However, a bigger problem in the group from the smaller municipality was conflicts originating from the involvement of the official who invited the participants. Interviews revealed that some officials in this group felt ordered to participate, which negatively affected their motivation to participate and the issues they considered possible to raise. Clearly, one should investigate how best to involve such a cooperation partner and evaluate, on an ongoing basis, what issues are best discussed in stakeholder groups versus individually.

4.1.2. Project goals

Interview results indicate that goals for the project differed considerably among participants, particularly between researchers and officials. Officials were interested primarily in boosting municipal adaptation work through influencing local politicians and producing concrete knowledge foundations to support adaptive planning. For researchers, there was an apparent mismatch between building the best possible overview of local vulnerability and creating scientifically valuable empirical research material. For example, high researcher involvement stimulated group discussion, but affected the perceptions of officials and thus reduced insight into local logics, which sometimes led to a perceived disharmony between collecting useful empirical material and stimulating knowledge exchange between officials and researchers.

The differences in expectations created some confusion about the purpose of some components of the project design, occasionally making it more difficult to finalise discussions. This seems to have resulted from unspoken positions and expectations that were not thoroughly addressed in the initial and ongoing discussions of realistic and desired project aims and outcomes. These problems could likely have been reduced if researchers had initially allotted more time to co-formulating overall project goals. Unfortunately, external project funding conditions indirectly hamper such efforts since project goals must follow research applications.

In the project, goals for the actual stakeholder interaction were gradually developed, which facilitated discussion of the role of specific exercises and the evaluation of goal fulfilment. These goals – focusing on the process in itself and not only on outcomes – as well as to unite around realistic project outcomes were by the researchers seen key aspects in the facilitation of joint knowledge. To incorporate these perspectives, it may be a useful design choice to meet all participants individually before initiating common group processes. However, as demonstrated, this will colour participants’ understandings and expectations which, in turn, will reduce the opportunities for the researchers to capture more spontaneous (and thereby possibly more ‘real’) reactions to the questions posed.

4.1.3. Roles

Transparency of information and communication seems to have been lacking in parts of the assessments. For example, several officials constantly asked for more information about the project. This demand was met by sending out minutes from the meetings and presentations of coming exercises to participants by e-mail. Despite these efforts, several officials still claimed they lacked sufficient insights. In retrospect, this does not seem to be a problem of too little information about results. The interviews reveal that participants rather lacked information on how the results were analysed and how the analysis was going to be used. These problems could probably have been reduced if the researchers had more thoroughly related the results to common project goals.

In terms of the discussions held during the dialogues, officials perceived that the open and wide-ranging discussions of vulnerability drivers permitted ‘out-of-the-box’ thinking which broadened established perspectives on risks. Nevertheless, these discussions sometimes stole time from more tangible discussions on risks. To be more tangible, discussions should arguably build on concrete examples from the municipalities. However, an important lesson learned is to avoid focusing on too detailed consideration, as this may only engage a few participants, thereby inhibiting the joint group discussions.

Regarding the flexibility criterion, the project was considered flexible in terms of content (as presented above), but less flexible when it came to changing the allotted time for the meetings. The time between the dialogues was 4–12 months, which several participants thought was too long, even though individual interviews were held and e-mail updates were sent between meetings. The long interval between physical meetings also led to some group member turnover, necessitating long recaps of the research project at every dialogue, which reduced the length of the common discussions. On the positive side, however, the fact that the dialogues were held over a longer period of approximately two years made it possible to discuss related policy developments and practical work in the municipalities, which would not have been possible otherwise. In this perspective, reducing the length of the interactive phase would not have been a good alternative. In order to maintain an active joint knowledge production phase, some participants argued that extra meetings on specific topics with only a few participants should have been added between the stakeholder dialogues.

From the above reasoning, two facilitating factors for building a flexible research outline were identified: (1) to make time in the project to address related upcoming contextualised policy examples, and (2) to involve a few participants in more sector specific supplementary meetings between the dialogues. Such supplementary meetings could serve the double purpose of facilitating overarching discussion during the larger group dialogues (i.e. participants would know that identified details would be addressed later), and maintaining an active interaction.

4.2. Interaction design

Results concerning the design of the individual exercises during the stakeholder dialogues indicated that the joint knowledge production was both facilitated and constrained by identified intermediaries.

4.2.1. Physical intermediaries influencing interaction

Physical intermediaries which stood out in the empirical material included local case areas, maps and matrices which were familiar methods to officials from several of the included sectors. Thus, their use reduced start-up times, stimulated in-depth learning and averted misunderstandings in the deliberative exercises. The role of questionnaires was also frequently mentioned in the evaluations, albeit not in the same positive way as they were partly seen to obstruct cooperation.

4.2.1.1. Case areas

The local case areas, consisting of one development project in each municipality, stimulated interaction among participants by interlinking different worldviews. The project in the larger municipality concerns the urban renewal of a central area which functions as a communications hub for the region. The project in the smaller municipality concerns the establishment of a new area which is planned to be built close to a newly constructed railway station on the periphery of the region. In both municipal groups, and among researchers, these areas were considered concrete and well-known examples that connected the research project with the practical work within the municipalities and led to tangible discussions on climatic and socioeconomic risks. For example, by demonstrating the possible effects of appending climate change risks on specific infrastructure and buildings, and by quantifying these in economic terms, the case areas facilitated joint knowledge production which could be directly transferred to practical work within individual sector departments. Nevertheless, the case areas also highlighted issues at the intersection of current departmental responsibilities which will be important to clarify within the municipality in the near future. One such example, highlighted during stakeholder dialogue A in the larger municipality, was the discussion on what sector(s) or actor(s) should finance the extra costs emerging from new developments on contaminated or unstable land in the developing area. As one of the officials indicated, the use of the case area example opened up for a climate vulnerability framing of issues for which responsibility is currently unclear and which need to be resolved to improve local capacity to adapt to climate change:

…there is a lot to consider here since this will be a development area faster than we had expected. It is low lying, it contains both road tunnels and railroad traffic, it has contaminated zones, it is a filling area, and it has stability problems. (Stakeholder dialogue A, Gothenburg. Author’s translation)

The choice to focus the group discussions on these small, well-known and relevant areas was made by the researchers during the initial project design phase. However, all participants were involved in the final selection. This approach worked very well and would probably be a good introductory exercise for initiating the participatory process in other types of assessments.

4.2.1.2. Maps

Maps are another identified important physical intermediary familiar to the participants. Maps of the two case areas, highlighting current and planned future infrastructure, made overviews of current practical work and, in particular, vulnerable areas more tangible and easy to discuss across the sectorial divisions. It also facilitated productive discussions of locations which will probably be sensitive to the combined effects of urbanisation and climate change in the future. This was clearly exemplified in the first exercise during stakeholder dialogue C (Table 1). Here, participants were first asked to identify the most sensitive points in the case areas from their sectoral perspective today, and up to 2060, by marking them on a transparent map. Secondly, the maps were layered on top of each other to visualise intersections in sensitivity.

The overlay maps served the double benefit of localising vulnerability ‘hot-spots’ which spurred discussions of cross-sectoral adaptation measures and provided visionary discussions of how the case areas could be developed in the future and what demands this would put on common departmental activities. With reference to the first benefit, a few of the vulnerability hot-spots identified in the overlay exercise had not previously been prioritised by any of the included sectors individually in the larger municipality due to their considered low vulnerability. However, since these were highlighted by several of the sector representatives, the results suggest that these are important locations to analyse in further detailed municipal assessments. With reference to the second benefit, the maps served as the basis for discussions on the future sensitivity of buildings and infrastructure, which is exemplified by the following interaction between two officials from the smaller municipality during the evaluation after stakeholder dialogue C:

…when you are visioning 50 years ahead then you start thinking, well what have we here, what will we have here, and what has SMHI [The Swedish Metrological and Hydrological Institution] said about water levels? It [the discussions] becomes more forward-looking.

A fact is also that the houses which are built now should preferably stand there for 150–200 years, so we could easily have discussed how it will look also after 100 years. Look back, what houses here are not 100 years old? (Evaluation after stakeholder dialogue C)

As indicated by the quotes, the use of maps in the sensitivity exercise also facilitated discussions of local historical developments, which enriched the discussions. Due to their visual power when used in this forward-looking way, maps can also stimulate the incorporation of both historical and future impacts of climatic, technical and demographical changes in other types of participatory assessments.

The usefulness of maps in the sensitivity assessments was highlighted by almost all of the interviewed officials. More generally, however, opinions about how to use maps as project outcomes differed slightly between researchers, on the one hand, and officials, on the other hand, partly due to their different reasons for getting involved in the project. Officials argued that the resolution of the maps should have been higher to make them more useful in future municipal planning. Researchers, however, argued that the maps should not be too detailed in order to leave room for more general discussions of the determinants of vulnerability. This divide is one of the recurrent splits in the project and relates to the participants’ different goals and expectations. As mentioned above, the researchers’ main priority was to create an environment that was conducive for joint knowledge production, intentionally designing materials in a fairly ‘raw’ form, while officials prioritised useful information that could feed directly into the municipal work process. Due to the importance of these requirements for the group interaction, these matters should preferably be raised and discussed more explicitly in the start-up phase of the project.

4.2.1.3. Matrices

Matrices made overviews of local vulnerability determinants transparent which, in turn, facilitated structured discussions on risks and measures. Some of the officials interviewed explicitly expressed that risks are often presented in this way within their sectors and that the matrix form for presenting results in the project made it easier to discuss combinations of, and to rank, determinants affecting climate change vulnerability. One participant from the smaller municipality highlighted this role of matrices during the evaluation after stakeholder dialogue D:

It is smart to use a format for depicting results that we are familiar with. Matrices is something we are used to work with and that everyone can understand. (Evaluation after stakeholder dialogue D)

Matrices are, however, more commonly used in some sectors than others (especially for technical applications) and in slightly different ways. Thus, although most participants had used matrices before, it is important to discuss how they should be applied in the actual vulnerability assessment to avoid confusion.

4.2.1.4. Questionnaires

The last physical intermediary that seems to have had an influence on the joint knowledge production, and which was also discussed by several of the participants in the interviews and evaluations, is questionnaires. Questionnaires were sent out to participants prior to stakeholder dialogue B to prepare them for the first two exercises and to focus the discussions on issues where perceptions differed. Participants were asked to rank climatic and socioeconomic determinates according to their importance in influencing the municipality’s exposure to climate risks and to add determinants if necessary.

As hoped, the questionnaires both highlighted how officials had different perceptions of which determinants were important in influencing exposure and facilitated discussions of how these determinants interact with municipal practices. Technical development was considered to both increase and decrease local vulnerability to climate risks. It was perceived to decrease vulnerability through the development of flood protection devices, but increase vulnerability by making inhabitants more dependent on technical infrastructure (such as internet access) and thereby more vulnerable to disturbances.

However, although the questionnaires initiated a few fruitful discussions, many officials did not have time to complete them, while it also took a long time to present the long list of determinants at the stakeholder dialogues. This created some frustration among officials which seemed to negatively affect the joint knowledge production process. To increase the effectiveness of questionnaires in future participatory assessments, they should be short and quick to complete, while the number of responses should perhaps not be presented, and the results should be presented in a more generic form.

4.2.2. Mental intermediaries influencing interaction

Mental intermediaries which influenced the joint knowledge production included brainstorming techniques and metaphors/analogies. In different ways, these intermediaries seem to have created nodal points in the discussions and broadened the perspective on risks and possibilities as presented below.

4.2.2.1. Brainstorming techniques

The brainstorming technique was familiar to the participants and thus easy to initiate. Participants found that the technique broadened their perspectives on the risks and opportunities that might emerge from climate change in relation to existing work in the municipalities, revealing new differences and similarities between departments. It also helped to position climate vulnerability and adaptation in relation to other important municipal issues. During stakeholder dialogue A, participants were asked to come up with short- and long-term risks facing municipalities and then to discuss how these risks related to climate change. Several participants, both researchers and officials, saw this as a useful way to synthesise local and academic knowledge at the start of the joint knowledge production process as described by the following researcher:

The brainstorming about risks gave a lot since there was space for local experts to freely list, which gave a wider scope and magnitude of risks than we had expected. (Group interview with researchers after stakeholder dialogue A)

4.2.2.2. Metaphors/analogies

The use of specific metaphors and analogies favoured the general delineation of mental intermediaries, which created common images of abstract issues. The analogy of a solar system was used by the researchers during the third exercise of stakeholder dialogue C (Table 1) to identify actors deemed important in municipal adaptation management by likening them to planets in a solar system. As intended, the analogy spurred a broader discussion in the municipality which attempted to identify the important actors at the national and subnational levels for building capacity to adapt to climate change and implementing measures. As one official explained after dialogue C:

The approach of creating a planetary system, we have done something similar earlier, but not as open as now. The openness resulted in a wider ‘map’ where we identified not only the Swedish rail road administration, but also train passengers as relevant stakeholders. (Evaluation after stakeholder dialogue C)

Beyond its intended effects, however, the solar system analogy also spurred new ways of ranking the importance of, and possibility to cooperate with, the identified actors.

Another analogy identified in the empirical material was that of a poker game. This emerged spontaneously in an exercise in which participants were asked to identify and rank adaptive capacity factors. The exercise turned into a card game in which playing cards represented the weights of the various factors. The dramaturgical metaphor quickly generated results by creating an easily understood method for weighting and comparing climate risks in the form of playing cards, as supported by the following:

The exercise was fun and gave good results in a fast way. The card game analogy made the participants more alert. (Group interview with researchers after stakeholder dialogue B)

The metaphor also highlighted important relationships between the individual departments and the municipal board, as well as between the various factors affecting municipal adaptive capacity. Participants compared how they weighted climate risks from the perspective of their sector department and its role in municipal policy-making.

5. Conclusions

This study aimed to deepen the understanding of how participatory cross-sectoral vulnerability assessments can be facilitated by analysing the set-up and design of two municipal assessments in a recently completed research project. The applied evaluative framework of Hegger et al. (2012) highlighted aspects of project design and participant interaction which were important in facilitating joint knowledge production.

The success conditions of Hegger et al. (2012) were considered too generally formulated to allow in-depth evaluations or to stimulate process planning and reflection. To realise the potential of the framework as a tool for guiding scientific ex-post analyses, process planning and associated reflexive discussions, we developed a set of operationalised evaluation criteria that finalised the framework (Table 3). These criteria, which were derived from the literature on participatory research and science and technology studies, guided the analysis of the research project and the two municipal assessments. The criteria are suitable for planning and evaluating participatory assessments in other contexts.

The application of the evaluative criteria to the extensive empirical material enabled an understanding of how the set-up and design of the two analysed assessments influenced the participants’ joint knowledge production. The results revealed that choices made regarding the set-up of the project had far-reaching positive and negative consequences for cooperation in the two assessment groups. For example, ‘outsourcing’ the selection of participants to local officials in a managerial position resulted in the inclusion of broad and relevant competences. However, this also meant that some participants felt like they were ordered to participate, which influenced the issues participants felt they could raise. Building trust during the start-up phase emerged as important for developing joint knowledge production. The issue of trust is not explicitly included in the framework of Hegger et al. (2012). However, according to the results of this and previous studies (e.g. Blackstock et al. 2007), trust should be systematically analysed to build a deeper understanding of any stakeholder deliberations. Moreover, goals seem to have differed largely among participants. Though all goals were important, they were not negotiated thoroughly in the beginning and it was therefore difficult to follow-up on them during the project. Last, the assessment approach, following a sequence of stakeholder dialogues over two years, was thought to include too few meetings to maintain active joint knowledge production. However, the long time span from the first to the last dialogue allowed the incorporation of practical policy examples from the municipalities which made the discussions more tangible.

Physical and mental intermediaries both facilitated and obstructed the creation of joint knowledge on local climate change vulnerability during the deliberative exercises. Important physical intermediaries included: local cases, which led to concrete discussions of climatic and socioeconomic risks and connected the research to practical work in the municipalities; maps, which made overviews of current practical work and particularly vulnerable areas more easy to discuss across organisational divisions; matrices, which were a familiar format for presenting risks gave a clear overview of results; and questionnaires, which stimulated discussions on the interaction of various vulnerability components, but created some irritation in the groups since officials did not complete them in time. Mental intermediaries which significantly influenced the joint knowledge production included brainstorming techniques, which broadened perspectives on risks and possibilities; and metaphors/analogies, which created common images of abstract issues.

To conclude, this study has identified aspects which are worth considering when planning future participatory cross-sectoral vulnerability assessments, of which three stood out as extra important. The first is flexibility, which earlier has been raised by Lemos and Morehouse (2005) and de la Vega-Leinert et al. (2008). The results revealed that motivation can be raised and the production of practical and implementable knowledge can be stimulated by allowing participants to analyse important local vulnerability components in smaller groups between ordinary meetings demanding an open and flexible research approach.

Second, exercises which have clearly formulated aim(s) prior to the stakeholder interaction are easier to start-up, communicate and evaluate. Intermediaries can, if used to facilitate such specific interactive aims, spur valuable discussions across organisational divisions by creating common reference points. By thoroughly discussing the interactive aim of each planned intermediary, their success rate will likely increase.

Finally, the creation of conditions that foster commitment and create a sense of shared trust (cf. Johnson & Johnson 1997) was here found vital. The results suggest that interviews or consultations held individually with participants before the actual participatory phase may serve multiple facilitating purposes such as making the participatory parts easier to plan and ensuring that participants can voice their perspectives in the project. Such consultations can also lay the groundwork for the joint negotiation of project goals and desired outcomes, which are important to discuss and agree upon to ensure group trust and motivation.

Funding

The research was funded by the Swedish Research Council for Environment, Agricultural Sciences and Spatial Planning (FORMAS) and by NordForsk/TFI through the excellence centre for Nordic Strategic Adaptation Research (Nord-Star).

Acknowledgements

The authors wish to thank the stakeholders who participated in this study as well as the researchers Karin André, Mattias Hjerpe, Louise Simonsson and Yvonne Andersson-Sköld within the analysed project for their assistance and for sharing their ideas. The authors also wish to thank the two anonymous reviewers for their constructive comments on an earlier version of this article.

Notes

1. ‘Enhancing cities’ capacity to manage vulnerability to climate change’, financed by FORMAS and conducted between 2007 and 2012.