Organisational influence on the co-production of fire science: overcoming challenges and realising opportunities

Keywords: actionable science, collaboration, co-production, research organisations, science-management interface, science-policy interface, translation, wildfire social science, wildland fire.

In the US, wildland fire is becoming increasingly severe and burned area is growing (Cattau et al. 2020). Concerns about loss of life, the health impacts of smoke, the cost of suppression, and the destruction of homes and infrastructure are becoming more acute (Calkin et al. 2015). In response, fire scientists are advancing new risk analyses, examining the role of climate in changing fire regimes, and building tools to enable more effective wildfire response. But fire science is often underutilised and the latest research is not always integrated into management decisions (see e.g. Adams et al. 2017). The Joint Fire Science Program (JFSP) was established in 1998 to bridge the ‘gap’ between fire science and management (LeQuire 2011), and has facilitated the use of fire science (Hunter 2016; Maletsky et al. 2018). Despite this success, a number of barriers continue to impede the application of fire science in management, including cultural differences between scientists and managers, a lack of trust that impedes communication, institutional and bureaucratic challenges, the inaccessibility of fire science, and the perception that fire science is often not relevant to management concerns (Kocher et al. 2012; Hunter 2016). These barriers are similar to the challenge of integrating science into land and environmental management more broadly. But, as Hunter et al. (2020) point out, ‘relative to other fields (e.g. climate change adaptation), the evaluation of the use of science for the management of wildland fire is immature’. Thus, there is a need for additional research on connecting science and management within fire management specifically (Colavito 2017). The present study examined a set of fire science projects that involved collaborations between researchers, managers and other partners to better understand the benefits and challenges of such work, and more specifically organisational influence on co-production of fire science.

A number of researchers have called for more interaction between fire researchers and practitioners (Adams et al. 2017) and even specifically for ‘co-production’ to ‘bridge the gap between fire science production and application’ (Maletsky et al. 2018). Hiers (2017) argues that co-production could be transformative for fire science. Co-production is a process that involves collaboration between scientists and managers in knowledge creation and problem solving, and is often assumed to increase the utility of science for management. Co-production of fire science could benefit both managers and scientists, improving the relevance of the research and the ability of managers to solve wildfire problems (Varner and Hiers 2020).

In the broader field of natural resource management, the need for partnerships between scientists and practitioners is increasingly recognised (see e.g. Cook et al. 2013; Lawson et al. 2017; Carter et al. 2020). In a recent review of the use of science in National Forest management, White et al. (2019) conclude that natural resource management in the United States will ‘likely require increased engagement between managers and scientists’ to create actionable knowledge (White et al. 2019, p. 13). Increased engagement is often promoted as a solution to the problem of the so-called ‘gap’ between research and practice, and the incongruence between science and decision-making more broadly (Roux et al. 2006; Wyborn et al. 2019).

A number of terms have been used to describe processes that bring scientists and managers together to improve the utility of research, including use-inspired research (Colavito 2017), translational ecology (Enquist et al. 2017), actionable science (Beier et al. 2017) and co-production (Norström et al. 2020). We use the term co-production because a recent review of research examining the use of fire science found a growing focus specifically on co-production (Hunter et al. 2020). Co-production involves collaboration between scientists, managers and other invested parties in knowledge creation and problem-solving (Wyborn et al. 2019; Norström et al. 2020). Roux et al. (2006) describe co-production as ‘a shift from a view of knowledge as a ‘thing’ that can be transferred to viewing knowledge as a ‘process of relating’ that involves negotiation of meaning among partners’ (Roux et al. 2006, p. 16). This ‘process of relating’ allows for diverse types of expertise to be integrated into a learning experience that reframes a problem and how to address it (Roux et al. 2006; Schuttenberg and Guth 2015; Beier et al. 2017). When scientists and managers work through this learning process, they are more likely to see the knowledge that they co-produce as legitimate, credible and salient, and the knowledge is more likely to be used in decision-making (Cash et al. 2003; Meadow et al. 2015; Clark et al. 2016). The relevance of research products can be improved by integrating the knowledge of managers and community members into the various stages of the research process, from development of research questions and interpretation of results to dissemination of findings, to ensure that research responds to the concerns of end-users (Cash et al. 2003; Dilling and Lemos 2011; Beier et al. 2017). Similarly, researchers can describe the limitations of the research, the types of questions the research can answer, and what these answers could be used for (Cash et al. 2003; Beier et al. 2017). Through this knowledge sharing, the research objectives, methods and products can be negotiated and informed by both scientists and non-scientists, which improves the likelihood that participants will perceive the results as legitimate (Cash et al. 2003). Deeper understanding of research processes and scientific uncertainty can also improve participants’ perceptions of knowledge as credible (Schuttenberg and Guth 2015; Beier et al. 2017).

Despite the numerous benefits of co-production, a number of barriers have been documented. Existing organisational structures and cultures may be unable to accommodate the power-sharing culture that often characterises co-production (Wyborn et al. 2019). Incentive systems for many scientists act as barriers to co-production to the extent that they do not reward the usability of science or direct engagement with decision-makers (Dilling and Lemos 2011; Norström et al. 2020). In a study on barriers to use of wildfire decision support systems (DSSs), Colavito (2021) identifies the absence of effective strategies to integrate new tools into decision-making processes and a lack of capacity for the development and integration of these tools as two important barriers to use of DSSs by managers. According to Varner and Hiers (2020), fire science suffers from a growing emphasis on disciplinary silos and managers who are sceptical of research that does not integrate management experience, highlighting both the need for and challenges of co-production of fire science.

While the literature abounds with guidelines, principles and recommendations for how to design co-production processes (Dilling and Lemos 2011; Clark et al. 2016; Beier et al. 2017; Morisette et al. 2017; Djenontin and Meadow 2018; Norström et al. 2020), there is no single recipe for successful co-production. Rather, each co-production endeavour responds to a specific context that includes different capacities and constraints, power dynamics and decision-making contexts (Littell et al. 2012; Van Kerkhoff and Lebel 2015; Norström et al. 2020; Turnhout et al. 2020). The organisations involved in co-production are an important part of this context because they can enable or constrain co-production processes and subsequent outcomes, but very little empirical research has focused on the organisational context of co-production. In response, Djenontin and Meadow (2018) call for additional research into how organisations can support and incentivise co-production and Wyborn et al. (2019) recommend further research into how co-production can be institutionalised at different scales, including at the level of organisations. Two types of organisations are particularly relevant in this context. Research organisations focus on scientific research and include universities, government labs and science agencies, and research-focused non-governmental organisations. Boundary organisations facilitate translation and collaboration across science and decision-making, and support production of use-inspired science (Hunter et al. 2020), and similarly include both government and non-governmental organisations. Several studies have examined the role of boundary organisations in facilitating collaborations between fire scientists and managers (e.g. Colavito et al. 2019; Hunter et al. 2020), but to date, no studies have explored how research organisations influence the co-production of fire science. To address this gap, we studied how the USDA Forest Service Rocky Mountain Research Station (RMRS) as a research organisation influences co-production of fire science, with the goal of identifying institutional changes that research organisations can employ to enable future co-production efforts. To understand organisational influence, we examined how participants in co-production processes view the need for, benefits of and challenges to co-production, including the role of the RMRS in enabling and constraining co-production.

The USDA Forest Service RMRS is one of six regional research units that constitute the Research and Development (R&D) division of the USDA Forest Service. Forest Service R&D was established by statute as a research organisation in 1928 and amended in 1978 (16 U.S. Code §1642; OLRC 1978). A key provision of these statutes was to promote a comprehensive program of forest and rangeland renewable resources research that operated independently from the rest of the Forest Service and served all forest and rangeland managers regardless of ownership or administration. In addition to this statutory direction, the US Congress appropriates funds separately for management and for research within the Forest Service to ensure that research scientists and managers maintain distinct roles. Despite these separate roles, over time, RMRS has taken on a large portfolio of fire-related research funded by non-R&D appropriations, specifically the hazardous fuels and fire suppression programs designed to improve the agency’s mitigation of and response to wildfire. Thus, RMRS was selected for this study because of its outsized role in conducting fire research in the United States. In sum, while RMRS was created primarily as a research organisation with a focus on producing new knowledge related to resource management, similarly to many research organisations it increasingly functions as a boundary organisation in facilitating communications and collaborations between research scientists, land managers and other natural resource stakeholders. Thus, emerging collaboration on wildfire-related research represents an expansion of its statutory role and history of separating management from research and may move RMRS more toward co-production and other research endeavours that involve boundary spanning across science and management.

To understand organisational influence on co-production, we studied seven RMRS fire science projects. These projects were selected by RMRS scientists and program managers for inclusion in the study because they all involved research–management partnerships and exhibited many of the features of co-production, specifically collaborations between scientists and managers in knowledge creation and problem solving. For more information on these projects, see Table 1. To protect the confidentiality of the interviewees, we do not identify the specific projects in this paper. The authors of this study were not part of any of these projects.

Table 1.  Rocky Mountain Research Station Projects.

We conducted in-depth, semi-structured interviews to gain a detailed understanding of how project participants perceived the need for and benefits of co-production as well as organisation support for and barriers to co-production. In-depth interviews are widely utilised in the social sciences, including wildfire social science (see e.g. Olsen and Sharp 2013; McCaffrey et al. 2015; Mylek and Schirmer 2016; Rapp et al. 2020). We interviewed 33 individuals, with four to seven participants from each project and a similar number of scientists, managers and community members from each project wherever possible. None of the participants were part of more than one project. Of the 33 participants, 14 were scientists (10 from RMRS, 4 from universities and other federal agencies), 9 were managers (all from federal agencies) and 10 were community members (including staff from city and county government and non-governmental organisations). We aimed to interview the individuals who were most involved with each project, including the RMRS scientists working on the projects. Like many research organisations, RMRS employs both scientists who conduct research (known within Forest Service R&D as Research Grade positions) and professional staff who support research activities and help ‘deliver’ research to managers and otherwise work on science application (known at RMRS as Professional Grade). RMRS scientists (Research Grade and Professional Grade), as well as scientists from universities and other federal agencies are referred to throughout as ‘scientists.’ Decision-makers in federal agencies and the agency staff who work to implement these decisions are referred to as ‘managers’ throughout. All participating city and county government staff, representatives of non-governmental organisations and community members are referred to as ‘community members.’

We conducted nearly all of the interviews over the phone, with the exception of a handful that were conducted in person. All of the individuals we contacted participated in an interview, with the exception of one person. We utilised an interview guide (see Appendix A) to provide continuity and comparability across interviews while allowing for follow up questions and emergent phenomena (Patterson and Williams 2002). Interview questions were informed by the co-production literature and preliminary review of documents on RMRS and the seven case studies. Interview questions focused on understanding why participants engaged in these projects, the benefits of these types of collaborations, and the role of the RMRS in enabling and constraining science–management collaboration. We did not use the academic jargon of ‘co-production’ during the interviews, but rather focused on key features of co-production as outlined in the literature.

Interviews were 45–90 min long, and were recorded, professionally transcribed, proofed and conducted in compliance with the University of Montana Institutional Review Board (IRB Protocol #183-18). Using NVIVO 12 software, each transcript was coded to identify relevant concepts related to the literature on co-production as well as emergent themes not previously addressed in the literature (Patterson and Williams 2002). Interviews were first analysed individually and then comparatively to identify broader patterns between and across individuals. To compare across interviews, the data within each code were analysed in depth to better understand patterns across the sample. Analysis involved an iterative process of reading and re-reading interview transcripts to code for different concepts and themes as larger patterns emerged, as well as comparing and contrasting results with existing theory on co-production. The first author developed the coding scheme, based on preliminary analysis of eight interviews and intensive group discussion (Saldaña 2013) with the full author team. Then, all interviews were coded by the first author. During the analysis, the author group met weekly to discuss the data, the codes and the coding process, and the post-coding cross-interview analysis. The data excerpts below were selected because they either represent a perspective held by a majority of people in the sample or to illustrate the range of views held by people in the sample. Throughout the results, we indicate the number of interviewees who share a particular perspective.

Throughout our interviews, scientists, managers and community members described the challenge of working across science and management (see Table 2 for additional data on these challenges and the need for collaboration). Approximately half of the participants (15) discussed a ‘gap’ or separation between scientists and managers, with one scientist explaining that, as a consequence of a ‘huge chasm between research and management’, scientific products may not be relevant to management concerns or easily understood. Other participants described this disconnect as a lack of understanding that goes both ways: scientists may not understand the difficulty and complexity of management, while scientists may not feel that managers are paying attention to their findings. One manager suggested the need for an ‘olive branch’, referring to the importance of building relationships and understanding between scientists and managers. The use of the term ‘olive branch’ indicates that this manager perceived the need for a gesture of goodwill. Participants (11) explained that one of the challenges to effectively connect science and management is that the two efforts often operate based on different epistemologies and ways of knowing, distinguishing between the analytical knowledge of researchers and the experiential knowledge of fire managers. One participant suggested that ‘researchers […] are clueless because they don’t know what the real world is like.’ Participants suggested that some managers may resist using science because they perceive scientists as presenting their knowledge with a kind of authority that managers do not perceive the scientists have. One scientist noted that managers may value their experiential knowledge over research products, again characterising this difference in terms of the epistemologies valued by managers and scientists, and the challenge it posed for collaboration. Despite the challenging relationship between science and management, participants (12) noted that they depend on their partners to meet their broader objectives. Accordingly, the analytical knowledge of science is required alongside the experiential knowledge of managers to address the complexity of wildfire management. Participants suggested that collaboration enables them to achieve much more than they could independently, working solely within the capabilities and expertise of their own organisations.

Table 2.  Additional data on the disconnect between science and management and the need for co-production.

To close the ‘gap’ between science and management, participants were co-producing knowledge through a variety of collaborative processes (see Table 3 for additional data on the benefits of co-production). Participants (17) described co-designing research projects, explaining that dialogue enabled them to refine research questions, methods and analyses to enhance the relevance of the research to management decisions. One RMRS scientist talked about a particular project where, through conversations with managers and community members, they had learned what form of data was easiest for them to use and subsequently adjusted the project to better meet management needs. Participants (18) talked about opportunities for learning and the benefits of building a shared understanding through these projects, which led to adjustments to research questions, methods and management approaches. More specifically, one scientist discussed learning how to communicate across different perspectives to cultivate shared understanding.

Table 3.  Additional data on the benefits of co-production.

Many participants (24) emphasised the importance of valuing different kinds of knowledge in collaborative processes. One scientist described a project to develop maps of wildfire risk that managers could use when responding to a wildfire. The different zones on the maps integrated managers’ knowledge about the locations and terrain where they could effectively engage and suppress a fire, manager and community member perceptions of values at risk in that landscape, and scientists’ models of fire behaviour. A manager suggested that ‘both of those kinds of bodies of knowledge are incredibly important’. A community member similarly emphasised the need to ‘be willing to work with people and to see value in those different perspectives’, suggesting that integrating different perspectives into a project means that ‘you’re going to have a much more robust answer to your questions’.

Many participants (22) suggested that collaborations between scientists and managers can also integrate management knowledge into existing research outputs. One community member described how their experiential knowledge of fire fighting was used to improve the utility of landscape-scale modelling of wildfire behaviour. In this case, managers and community members helped scientists ‘ground-truth’ and refine how the model characterised the risk of projected wildfire behaviour, based on local knowledge of on-the-ground conditions. One manager suggested that integrating knowledge from management and research ‘works both ways where we’re enhancing the research and the research is kind of enhancing our ability to make decisions on the ground’.

Participants (15) also described how these collaborations enhanced ‘transparency’ around research and the way that information is used by managers, noting that this transparency can build support for both the research outputs and the management decisions that are informed by them. One manager explained that the outputs of their project, which integrated professional and scientific knowledge, will be useful for informing decision-making for fire management, as well as explaining how management decisions were made. One scientist suggested that, by incorporating a broad range of participants into the research process and informing communities about the work, the projects and their outputs, gain more ‘credibility’ and legitimacy because more people understand how they came about, and project outputs ‘will be more relevant’ and ‘will be more trusted’. One manager conveyed that when different kinds of people engage in a process, a broader range of people will see the types of people they trust involved and thus find the outcomes more credible.

One of the goals of this research was to better understand how research organisations enable and constrain co-production (see Table 4 for additional data on addressing the challenges of co-production). To this end, all participants were asked to reflect on the ways in which their organisation and RMRS specifically supported these projects and their work on them. While participants painted a nuanced picture of the organisational context within which they worked, there was widespread agreement regarding how RMRS might address some of the barriers to collaborative research.

Table 4.  Additional data on addressing the challenges of co-production.

Some participants (10) described RMRS as supportive of collaborations between scientists and managers. One scientist shared that their project was ‘well-received’ and they were invited to present the project to a variety of audiences. According to another scientist, RMRS is ‘very happy about [collaborative work]’ and ‘they want us to be interacting with managers’. Another scientist conveyed that ‘[RMRS] has always been very supportive of this work… I’ve never had anyone not support this work’. Upon further reflection, some participants (15) conveyed that while RMRS values collaborative work, the organisation provided inadequate resources for collaboration or insufficient recognition of this work in job descriptions and performance evaluations.

Many participants (21) explained that these types of collaborations are not part of their normal job duties and entail additional work that is not as easily recognised or valued in their performance evaluations. More specifically, participants (10) argued that Research Grade scientists within RMRS as well as the university scientists who partner with them are evaluated primarily based on research publications. One scientist described the trade-offs between addressing management concerns and working on peer-reviewed publications, explaining that scientists sometimes choose to work collaboratively at the expense of working on publications, or vice versa. When asked about the importance of producing publications from their collaborative work, one scientist commented, ‘It’s critical. Otherwise, I couldn’t be a part of it.’ To address this challenge, some participants recommended adjustments to performance evaluations to incentivise and reward both scientists and managers to work collaboratively and to value outputs for applied audiences.

Funding structures were also viewed as a barrier to collaboration. Out of the seven projects studied, only one received additional funding from RMRS specifically for collaboration, including funding for some of their partners. Nearly all of the people who felt they had adequate financial support to engage in collaborative work were part of this project and the project is expected to produce several publications, ensuring that scientists receive ‘credit’ in performance evaluations. One of the scientists from this project described their program manager, who is their direct supervisor, as ‘unbelievably supportive’, saying that this supervisor supported them spending time on the project and helped them identify additional collaborators. They shared that they ‘have felt so supported by the research station’ and are ‘lucky to have experienced that’.

Of the other six projects, only one participant described receiving financial incentives or support for engaging in this kind of collaboration, describing the receipt of awards and commenting, ‘That’s giving a person money and kudos too, so that’s a good way to reward someone and incentivise their work.’ In contrast, another scientist referenced these types of awards and argued that they do not reflect an incentive appropriate to the amount of investment by the scientists. Most participants regarded funding as largely inadequate to support the large investment of time required for collaborative work. One scientist explained that, because RMRS is unable to provide additional financial resources for collaborative work, scientists depend on external funders to support these time-consuming efforts. Some participants (10) also argued that external funding can make collaboration challenging because it’s often short-term and these are ‘big efforts’ that ‘take a while’. As one scientist put it, it is ‘hard to develop a program and things that’ll last if your money is year to year’. To address funding challenges, some participants (10) suggested that grants should be adjusted to adequately compensate participants for the time investment and to better match the duration of these efforts, which would appropriately incentivise participation.

Beyond the constraints that individual scientists experience to engaging in collaborative processes, some participants (11) also describe that as a team or organisation, they lack the staff capacity to meet the demand for these types of collaborations. One scientist argued that with more personnel they could do more collaboration, saying, ‘If we had more staff and capacity, we could do it a lot’, and that they currently have to turn away collaborative opportunities. Another scientist discussed the challenge of securing additional funding for new positions to grow collaborations. One manager argued that important management questions go unanswered because there is limited funding for researchers and too few scientists. To address this challenge, some participants (11) recommended greater investment in scientist positions to build capacity to engage in collaborative research.

Participants (16) also described a pressing need for translation of research into products that are useful to managers. Some participants specifically emphasised the need for more Professional Grade scientists who are not incentivised to publish, but instead focus primarily on translation. One scientist discussed the importance of ‘tech-transfer’ activities like adjusting tools to be more easily usable, training individuals on how the tools work, maintaining them over time, and providing feedback to scientists on any problems that users are experiencing. Another scientist explained that tech-transfer work can help translate and maintain research products for management use, and enable researchers to continue creating new tools, instead of spending their energy maintaining their successful products. In addition to building capacity to translate research into products that are useful to managers, one scientist argued that Professional Grade scientists need career ladders that value and incentivise their work to address frequent turnover in these positions due to lack of upward mobility.

Participants in this study suggested that fire science may not always be relevant to managers, and that differences in organisational culture can cause difficulties in communication and manifest as different epistemologies. By integrating different forms of knowledge, participants argued that they had enhanced the relevance of fire science for specific contexts, incorporating decision contexts, financial resources, community values and diverse goals into research products. They also suggested that the on-going learning involved with co-production can build trust and improve transparency, thereby increasing the credibility, salience and legitimacy of fire science. It is important to note that this research examined seven wildland fire projects that were characterised as co-production, and thus these results may be biased toward participants who have been more successful at overcoming organisational barriers to co-production. Despite this success, however, participants in this study still experienced important organisational constraints. As Colavito (2021) points out, the ‘disconnect between science and management’ is not unique to fire science, but rather ‘pervades many complex natural resource issues that seek to integrate scientific information into decision-making’. Thus, even though this study focused on a specific research organisation, RMRS, the incentives and barriers described in this paper may apply to co-production more broadly, given the similar cultures and incentive systems that are present in many research organisations, from government labs to universities.

Like previous research, we found that support and incentives for co-production are vital (Dilling and Lemos 2011; Meadow et al. 2015; Clark et al. 2016; Beier et al. 2017), and many institutional structures continue to limit this type of engagement (Djenontin and Meadow 2018; Wyborn et al. 2019; Turnhout et al. 2020). As a research organisation whose mission includes producing science to assist Forest Service managers, co-production is aligned with the RMRS mission. However, RMRS still exists within a broader science culture and structure that values, conducts, rewards and funds science in particular ways. As such, RMRS provides a window into both the benefits of co-production as well as the ways that co-production challenges conventional models of science. Our results suggest a number of institutional changes that can improve the ability of research organisations to support co-production, within fire science and beyond.

Adjustments to performance evaluations could help incentivise co-production, and better connect fire science and management. More specifically, participants in this study argued that performance evaluations for scientists tend to value peer-reviewed scientific publications more than engagement in co-production. Previous research has found that incentive structures for scientists that focus on peer review publications can discourage participation in co-production (Dilling and Lemos 2011; Beier et al. 2017). Furman et al. (2018) recommend that performance evaluations in academia consider relationship-building on par with outputs and recognise the long-term nature of collaborative research processes (Furman et al. 2018). These kinds of structural adjustments to performance evaluations could prompt a broader cultural shift within the scientific community toward valuing direct engagement with decision-makers (Clark et al. 2016).

The Forest Service guide for evaluation of Research Grade scientists was recently revised to emphasise impact and collaboration. The guide states that ‘information and technology transfer (ITT) is an important aspect of disseminating research results’ and that ‘research accomplishments have no ultimate effect until someone uses the scientist’s findings, or puts new information into the hands of those that can apply the results’ (USDA Forest Service 2019, p. 45). The revision also specifically recognised the value of ‘conducting joint research with potential end users or research results’ and noted ‘much Forest Service research is also the result of collaborative relationships with research partners and land management’ (p. 45). These changes had not yet been formally adopted at the time the interviews for this study were conducted, and it may take time to shift organisational structure and culture toward valuing and rewarding impact and collaboration alongside other scientific work.

In addition to changes to performance evaluations, greater investment in positions specifically designed to facilitate dialogue across science and end-users (often called science application or science delivery positions) is needed. These positions often focus on translating research into products that meet the needs of end-users and helping scientists better understand the needs of managers. In recognition that this work requires spanning the needs and cultures of both science and management, these individuals are increasing called ‘boundary spanners’ (Cash et al. 2003; Bednarek et al. 2018). Colavito (2021) describes boundary spanners as individuals who ‘bridge the spheres of science and management’ and ‘provide communication and translation functions that help to facilitate the exchange of knowledge to co-develop’ science, decision support tools and other knowledge for fire management. Matso and Becker (2015) recommend that projects build in capacity for integration through dedicated resources for boundary spanning to improve the utility of science for decision-making. Note that boundary spanners do not need to be employed by boundary organisations and can reside in research organisations. But regardless of location, career ladders and compensation that adequately value the work of boundary spanners are needed (Bednarek et al. 2018).

While investments in boundary spanning positions could help integrate manager knowledge and needs into existing research products, if designed as one-way science delivery, these investments could further separate managers and scientists from directly engaging with one another. This highlights an important trade-off, because many of the benefits of co-production come from iterative engagement between scientists and managers throughout the process, such as shared learning and relationship building (Sarkki et al. 2015; Schuttenberg and Guth 2015; Van Kerkhoff and Lebel 2015; Norström et al. 2020). Thus, to improve the ability of research organisations to create science that is useful for management and addresses the management questions that go unasked, it may be necessary to build capacity for boundary spanning and better incentivise scientists to engage with managers.

It is important to recognise, as Cash et al. (2003) have noted, that perceptions of the salience, credibility and legitimacy of knowledge are subjective, and efforts to increase these characteristics for some can often decrease them for others. When research organisations make a shift to provide more salient outputs for managers, they may jeopardise the credibility of their research within the scientific community, which currently assigns more value to questions and knowledge that builds from and contributes to the scientific literature. Further, many policies require that federal agencies like the Forest Service integrate the ‘best available science’ into their decisions and ‘best available science’ is typically defined as peer-reviewed science. This presents a conundrum for researchers whose work is relevant to natural resource management, especially if increased investments in co-production come at the expense of publishing. Thus, research organisations need to enable scientists to engage with managers to ensure that their research is actionable and to publish their results so that management decisions are legally defensible.

Funders also play an important role in encouraging research processes that connect science with action (Arnott et al. 2020). According to Matso and Becker (2014, 2015), funders can support actionable science by modifying the review process to place more emphasis on science that is connected to decisions, requiring that scientists and end-users engage throughout the research, and dedicating funding to support this engagement. Including decision-makers on review panels and conducting impact analysis can also help advance research that meets the needs of end-users (Nyboer et al. 2021). Echoing these calls, Varner and Hiers (2020) recommend that fire science funders require that proposals demonstrate how projects will meet the needs of managers and outline how managers will be involved in all stages of the research. Longer-duration grants and dedicated funding can enable science–management partnerships that are focused on the problems that managers face (Carter et al. 2020; Nyboer et al. 2021). To the extent that fire science funders provide training and resources on co-production, as recommended by Varner and Hiers (2020), funders can help build the capacity for co-production.

Hunter et al. (2020) point out that co-production processes are time-consuming and potentially prohibitive. Boundary organisations can help address these barriers by facilitating ongoing interactions between fire scientists and managers, to advance fire science that is relevant to decision-makers (Kocher et al. 2012; Colavito et al. 2019; Hunter et al. 2020). For example, Maletsky et al. (2018) found that the JFSP Fire Science Exchange Network facilitated the use of fire science in management decisions. In the present study, several of the projects included external partnerships with local government, NGOs, or other federal agencies, but none of these partners play the role of a boundary organisation. Research organisations do not necessarily need to evolve into boundary organisations. However, research organisations can enhance their capacity to engage in boundary-spanning work, and leverage their resources and skills to support co-production to improve the utility of science, including fire science, for management.

Our results suggest that research organisations can make adjustments to better support the co-production of actionable science, and thus help agencies respond more effectively to the ever-growing risk of wildfires. The growing body of research on the benefits of co-production, including this study, indicates that collaborative engagements between scientists and decision-makers increase the relevance and utility of research products. But because co-production means doing science differently, these efforts require institutional support. Research organisations, from government agencies to universities, are uniquely positioned to incentivise collaborations that create more useful and useable science through investments in co-production via modified funding arrangements, performance evaluations, career laddersa and resources for translation.

Raw data are not available owing to confidentiality requirements.

The authors declare no conflicts of interest.

This research was supported by the USDA Forest Service, the University of Montana BRIDGES program and the National Science Foundation Division of Graduate Education (1633831). The findings and conclusions in this publication are those of the authors and should not be construed to represent any official USDA or US Government determination or policy. Any use of trade, product, or firm names is for descriptive purposes only and does not imply endorsement by the US Government.