Integrating collaborative research in marine science: Recommendations from an evaluation of evolving science-industry partnerships in Dutch demersal fisheries

Integrating collaborative research in marine science: Recommendations from an evaluation of

evolving science-industry partnerships in Dutch demersal fisheries

Steins, Nathalie A.; Kraan, Marloes L.; Reijden, van der, Karin; Quirijns, Floor J.; van

Broekhoven, Wouter; Poos, Jan Jaap

Published in:

Fish and Fisheries

DOI:

10.1111/faf.12423

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from

it. Please check the document version below.

Document Version

Publisher's PDF, also known as Version of record

Publication date:

2019

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):

Steins, N. A., Kraan, M. L., Reijden, van der, K., Quirijns, F. J., van Broekhoven, W., & Poos, J. J. (2019).

Integrating collaborative research in marine science: Recommendations from an evaluation of evolving

science-industry partnerships in Dutch demersal fisheries. Fish and Fisheries, 21(1), 146-161.

https://doi.org/10.1111/faf.12423

Copyright

Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policy

If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

146  

|

Received: 18 May 2019 

|

|

O R I G I N A L A R T I C L E

Integrating collaborative research in marine science:

Recommendations from an evaluation of evolving

science-industry partnerships in Dutch demersal fisheries

Nathalie A. Steins

 | Marloes L. Kraan

 | Karin J. van der Reijden

 |

Floor J. Quirijns

 | Wouter van Broekhoven

 | Jan Jaap Poos

This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes.

© 2019 The Authors. Fish and Fisheries published by John Wiley & Sons Ltd 1_{Wageningen Marine Research, IJmuiden,}

The Netherlands

2_{Environmental Policy Group, Wageningen} University, Wageningen, The Netherlands 3_{VisNed, Urk, The Netherlands}

4_{Aquaculture and Fisheries Group,} Wageningen University, Wageningen, The Netherlands

Correspondence

Nathalie A. Steins, Wageningen Marine Research, PO Box 68, 1970 AB IJmuiden, The Netherlands.

Email: nathalie.steins@wur.nl Present address

Karin J. van der Reijden, University of Groningen, Groningen, The Netherlands Jan Jaap Poos, Aquaculture and Fisheries Group, Wageningen University, Wageningen, The Netherlands

Funding information

The research leading to this manuscript did not receive funding. However, the Dutch Ministry of Agriculture, Nature and Food Quality Coöperatieve Visserijorganisatie, Nederlandse Vissersbond, VisNed, and Ekofish Group(co)funded some of the projects that were reviewed, while others received grants from the former European Fisheries Fund or the current European Maritime and Fisheries Fund., Grant/Award Number: n/a

Abstract

The increasingly complex nature of marine resource management calls for stronger stakeholder participation in advancing knowledge and developing management ap-proaches. Studies on stakeholder involvement in marine resource management have primarily focussed on participation in resource use negotiation and not on partici-pation in science. Using fishers' knowledge research frameworks, we evaluate over 15 years of science-industry research collaboration (SIRC) in Dutch demersal fisher-ies. Four key lessons emerge: (a) Capacity building in SIRC works multiple ways and triggers shifts in the fishers' knowledge research spectrum; (b) Successful SIRC de-pends on acceptance of industry collected data for scientific advice, which necessi-tates close involvement of end-users from the outset to provide feedback and obtain support; (c) (Fisher) participation raises often-overlooked equity questions and may result in selection bias; and (d) The governance context strongly influences struc-ture of SIRC and integration of SIRC knowledge. To ensure a sustainable, continuous process of stakeholder participation and use of their knowledge in marine resources research, collaborative research should be embedded in the institutional frameworks for science and management. It demands continuous maintenance of the relationship between scientists and stakeholders in the context of management developments, calls for reflection about selection and equity considerations, and requires continu-ous attention for communication with all parties involved at different levels. The les-sons learnt in science-industry research collaboration in fisheries are also relevant for the wider field of marine science, where stakeholder participation is necessary but not yet common.

K E Y W O R D S

collaborative research, fishers' knowledge research, knowledge co-creation, marine resource management, marine science, stakeholder participation

1 | INTRODUCTION

In science and management of marine renewable resources, there is an ongoing call for stronger participation of stakeholders (Kaplan & McCay, 2004; Mackinson & Middleton, 2018; Mackinson, Wilson, Galiay, & Deas, 2011; Mauser et al., 2013; Röckmann, Kraan, Goldsborough, & Hoof, 2017; Stange, 2014). Two key factors con-tribute to this call. First, the problems in managing marine resources involve many different and high stakes. They are characterized by different underlying values, have no definitive or objective solu-tions or answers, and call for trade-offs (Funtowicz & Ravetz, 1993; Pielke, 2007). Addressing these so-called wicked problems (Jentoft & Chuenpagdee, 2009) asks for participation of those affected.

Second, there is a growing understanding of the linkages and in-terdependencies between the social and ecological system (Berkes & Folke, 2000). Successful management of fisheries requires better knowledge of the social dimension of resource use to complement the ecological knowledge that marine sciences traditionally focus on (Fulton, Smith, Smith, & Putten, 2011). Fishers have extensive knowledge about the drivers for the choices they make in exploiting resources, thus providing valuable insights in this social dimension.

Given that stakeholder communities possess extensive knowledge about both the ecological and social functioning of marine socio-eco-logical systems, it is helpful to include them in the science process that informs policy. Their involvement should go beyond being part of a peer review community (Funtowicz & Ravetz, 1993). It should include active participation in the full scientific process, comprising hypothe-sis formulation, data collection, data interpretation and review.

Much has already been written about participation of stakehold-ers in the management process of marine renewable resources (Kraan, Hendriksen, Hoof, Leeuwen, & Jouanneau, 2014; Reed, 2008; Steins & Edwards, 1999). However, stakeholder participation in the scientific process has been underexposed. A major source of experience with stakeholder participation in science can be found in fisheries (Hartley & Robertson, 2006; Holm et al., in press; Johnson & Van Densen, 2007; Johnson & Van Densen, 2007; Mackinson et al., 2011). Since the early 2000s, the annual number of scientific publications about col-laborative research in fisheries has grown substantially (Figure 1). The focus of this research is shifting from passive participation towards active collaboration (Dörner et al., 2015; Holm et al., in press; Mangi et al., 2018). In the former, researchers use fishing vessels as a plat-form to collect data (Kaplan & MacCay, 2004; Mangi et al., 2018). In the latter, fishers are actively engaged in the development of research questions, project design, collection and interpretation of data, and communication of results (Hartley & Robertson, 2006; Holm et al., in press; Johnson & Van Densen, 2007; Johnson & Van Densen, 2007; Mackinson et al., 2011; Mangi et al., 2018; Thompson, Stephenson, Rose, & Paul, 2019; Wendt & Starr, 2009).

Scientists, fishers and their representatives, managers, and NGOs advocate science-industry research collaboration (SIRC) for multiple reasons. First, SIRC allows for cost-efficient data col-lection with increased temporal and spatial coverage, improving the knowledge base for fisheries management (Armstrong, Payne,

& Cotter, 2008; Johnson & Van Densen, 2007; Kraan, Uhlmann, Steenbergen, Helmond, & Hoof, 2013; Lordan, Cuaig, Graham, & Rihan, 2011; Mangi et al., 2018; Stephenson et al., 2016; Wendt & Starr, 2009). Second, SIRC increases transparency and communi-cation between scientists and fishers. This results in mutual un-derstanding of data collection and interpretation, increased trust amongst fishers in science-based management and increased legit-imacy of the management framework (Hartley & Robertson, 2006; Johnson & Van Densen, 2007; Wendt and Starr (2009); Kraan et al., 2013; Mauser et al., 2013; Dörner et al., 2015; Stephenson et al., 2016; Mangi et al., 2018; Thompson et al., 2019). Third, SIRC contributes to capacity building of the industry. Fishers and their representatives learn how to interpret data they collected and how these affect stock assessments and management advice

1. INTRODUCTION 147

2. CONCEPTS AND METHODS 148

2.1 Fishers' knowledge frameworks 148

2.2 Methods 149

3. THE EVOLUTION OF SIRC IN THE NETHERLANDS

149

3.1 Building trust 149

3.2 Broadening scope 152

3.3 Increasing degree of integration of fishers'

(experiential) knowledge 152

3.4 Funding and communication issues 152

3.5 Arrival of industry scientists 154

3.6 Participation of NGOs 155

4. OUR LESSONS FROM SIRC 155

4.1 Lesson 1: Capacity building is multidirectional, leading to shifts in the FKR spectrum

155 4.2 Lesson 2: Embedding end-users is crucial for

successful SIRC 156

4.3 Lesson 3: Equity questions and selection bias need to be considered

157 4.4. Lesson 4: The governance system impacts

structure and integration of SIRC

157 5. RECOMMENDATIONS FOR COLLABORATIVE

MARINE SCIENCE AND MANAGEMENT

158 5.1 Recommendation 1: Embed collaborative

research in institutional framework

158 5.2 Recommendation 2: Keep reflecting on

selec-tion bias and equity quesselec-tions 159

5.3 Recommendation 3: Make communication an integral part of SIRC

159

6. CONCLUSION 159

Acknowledgements 159

Data availability statement 160

(Johnson & Van Densen, 2007). Fourth, SIRC improves the soci-etal relevance of research by ensuring it addresses urgent man-agement problems of end-users (e.g. gear technology, real-time closures) (Johnson & Van Densen, 2007; Stephenson et al., 2016). Finally, SIRC may provide alternative sources of income for partic-ipant fishers (Johnson & Van Densen, 2007).

Nevertheless, SIRC has several potential pitfalls that, if not prop-erly addressed, will undermine research cooperation. Scientists and fishers may have different incentives to participate, different project objectives and different interpretation of results (Kraan et al., 2013; Lordan et al., 2011; Mangi et al., 2018; Steins & Edwards, 1999). Also, differences related to education and experiences can cause miscom-munication between fishers and scientists (Verweij, Densen, & Mol, 2010). Finally, scientists and policymakers may not accept data from SIRC in assessment models and decision-making because they may be delivered in incompatible formats, provide an incomplete view in time and/or space, may be inconsistent with other data or are not trusted because they were collected by fishers who have a vested interest (Hamilton, Giningele, Aswani, & Ecochard, 2012; Kraan et al., 2013; Mangi et al., 2018). As a result, these data may not be used at all.

In this paper, we share lessons learned from more than 15 years of evolving SIRC experience in Dutch demersal fisheries, using fishers' knowledge research (FKR) frameworks as developed by Johnson and Van Densen (2007) and Stephenson et al. (2016). These frameworks aid in evaluating the type of knowledge used in fisheries research, ranging from fishery observations to fishers' experiential knowledge. Also, the FKR frameworks help to assess the extent to which this knowledge is integrated in fisheries science and management. We argue that the les-sons learnt in SIRC go beyond use in biological and technical fisheries research. They are also relevant for the wider field of marine science, where stakeholder participation is necessary but not yet common.

2 | CONCEPTS AND METHODS

2.1 | Fishers' knowledge frameworks

In 2007, Johnson & Van Densen published an FKR framework pro-viding guidance for SIRC, based on an analysis of three forms of

collaborative research in the Northeastern USA and Northwestern Europe: (a) fisheries-dependent data collection, (b) industry-based surveys and (c) gear selectivity studies. Their framework is organ-ized according to five stages of a research project cycle: (a) problem identification, (b) research approach and design, (c) data collection, (d) data processing and analysis and (e) communication of results. In all five stages, communication to the industry at large is an essen-tial element and all data users including fishers, stock assessment scientists and managers must be involved during all stages.

Johnson & Van Densen's FKR framework focuses on fishers as a platform to collect data. This leads to a situation where mainly fish-eries science benefits directly from information provided by fishers. The authors acknowledge that there are also indirect benefits of SIRC, particularly capacity building: SIRC empowers fishers by im-proving understanding and appreciation for information produced through scientific research and how this translates to management advice (Johnson & Van Densen, 2007).

Nearly a decade after Johnson and Van Densen (2007) published guidelines for organizing SIRC, Stephenson et al. (2016) published a paper on fishers' knowledge research. They argue that "fishers’ knowledge includes, but is much greater than, basic biological fishery information. It includes ecological, economic, social and institutional knowledge, as well as experience and critical analysis of experiential knowledge." They distinguish between fisheries observation knowl-edge (FOK) and fishers' experiential knowlknowl-edge (FEXK). Scientists gather FOK by using fishers as platforms to collect data, whilst FEXK is unique knowledge fishers possess, derived from fishing as a social practice. Stephenson et al. (2016) point out that there are different degrees to which information from FOK and FEXK are integrated in fisheries assessment and management: from fishers only providing data to a fully participatory governance regime in which fishers' in-formation and knowledge is used in shaping management (Figure 2). They conclude that “fishers’ knowledge is best implemented in a par-ticipatory process designed to receive and use it.” In doing so, they reemphasize the case made by Johannes, Freeman, and Hamilton (2000) who convincingly show the value of using fishers' ecological knowledge to improve stock and ecosystem management.

In this paper, we define SIRC as collaborative projects between scientists and fishers aimed at improving the knowledge base for

F I G U R E 1   Annual number of scientific

publications (1970–2018) on Web of Science about the collaboration of scientists and fishermen in fisheries with "cooperative research," "participatory research," or “collaborative research” and “fisheries” in title, abstract or keywords. Search performed on August 20th 2019

fisheries management, in which the fishing industry is involved in at least one of the five project stages identified by Johnson and Van Densen (2007) and which uses FOK and/or FEXK (Stephenson et al., 2016).

2.2 | Methods

We selected 15 Dutch SIRC projects using the following criteria: (a) they focused on demersal fisheries in the North Sea; (b) they focused on improving the knowledge base for fisheries management (e.g. stock assessments, reduction of unwanted bycatch); (c) they were multi-annual and completed; and (d) they could be evaluated by a sin-gle lead scientist over the entire project duration. We excluded SIRC projects in the Dutch freshwater, shellfish, crustacean and pelagic fisheries. By analysing SIRC projects related to one fisheries subsec-tor, we could assess if interactions amongst projects occurred.

We evaluated the selected projects (Table 1) using the FKR frameworks by Johnson and Van Densen (2007) and Stephenson et al. (2016). Project lead scientists received an open-ended ques-tionnaire. We asked them to evaluate the collaboration between fishers and scientists for each of the five project stages identified by Johnson and Van Densen (2007). Where evaluations resulted in questions, we spoke with project leaders. We coded responses to the questionnaires using ATLAS.ti, allowing for qualitative anal-ysis. We then analysed where fishers' knowledge was positioned in the FKR spectrum developed by Stephenson et al. (2016), and whether and how it moved within this spectrum during the project

duration. We also analysed interactions amongst projects. Finally, we assessed lessons learnt in relation to stakeholder involvement.

The analysis did not include formal interviews with participating fishers or industry representatives due to resource limitations. In one case, however, both the lead scientist and a scientist employed by the fishing industry each independently filled out a question-naire about the same project. A comparison of responses showed no conflicting views. Furthermore, all authors have had many in-formal conversations with fishers and representatives through their work in SIRCs and information from these conversations contributed to our findings. We also presented our findings about the evolution of Dutch demersal SIRCs to representatives from the fishing industry, government and NGOs in a "SIRC Platform” meet-ing. This meeting confirmed our main observations and lessons.

3 | THE EVOLUTION OF SIRC IN THE

NETHERL ANDS

3.1 | Building trust

The first SIRC in the Netherlands started in 2002 from a deep re-lational crisis between the national fisheries institute and the fish-ing industry. An ongofish-ing discussion about the accuracy of the stock assessments for European plaice (Pleuronectus platessa, Pleuronectidae) and common sole (Solea solea, Soleidae) in the North Sea by the International Council for the Exploration of the Sea (ICES) escalated to a "them against us" situation on both sides.

F I G U R E 2   Fishers' knowledge research spectrum, developed by Stephenson et al. (2016). All but the italicized cell are considered

T A B LE 1  O ve rv ie w o f s el ec te d S IR C p ro je ct s i n t he D ut ch N or th S ea f la tf is h f is he rie s i n t hi s s tu dy SI RC p ro jec t Ty pe C om mo n o bj ec tiv es a St ar t En d In iti at or Fu ndin g F-pr oj ec t - W or k P ac ka ge ( W P) 2 B et te r u se of f is he rie s d at a ( F-pr oj ec t) C at ch m on ito rin g f or st ock a ss ess m en t F-pr oj ec t a s a w ho le : I m pr ov in g c ol la bo ra tio n a nd co m m un ic at io n b et w ee n g ov er nm en t, s ci en tis ts a nd fis he rm en o n I C ES s to ck a ss es sm en t o f p la ic e a nd s ol e in r es po ns e t o d is ag re em en ts o ve r t he q ua lit y o f t he se as se ss m en ts . W P2 : D ev el op a L an di ng s p er U ni t o f Ef fo rt ( LP U E) f or u se i n s to ck a ss es sm en ts . 20 02 20 07 M ini st ry M in is tr y a nd i nd us tr y Se lf-sa m pl in g o f d is ca rd s i n d em er sa l f is he r-ie s ( SS D is ) D is ca rd s m on ito rin g f or st ock a ss ess m en t O bt ai n r ea lis tic e st im at es o f t he q ua nt ity a nd c om po si -tio n o f d is ca rd s i n t he b ea m t ra w l f is he rie s. 20 04 20 09 In dus tr y In dus tr y O bs er ve rs o n-bo ar d s ci en tif ic s ur ve ys (O B SS ) Su rv ey f or s to ck ass ess m en t In cr ea se t ra ns pa re nc y r eg ar di ng t he d at a c ol le ct io n pr oc es s o n-bo ar d s ci en tif ic s ur ve ys a nd i m pr ov e fis he rs ' u nd er st an di ng a bo ut f is h s to ck s ur ve y me tho do lo gy . 20 07 O ng oi ng , no e nd date F-pr oj ec t M in is tr y ( D C F) Se lf-sa m pl in g o f d is ca rd s i n d em er sa l f is he r-ie s ( SS D is -D C F) D is ca rd s m on ito rin g f or st ock a ss ess m en t O bt ai n r ea lis tic e st im at es o f t he q ua nt ity a nd c om po si -tio n o f d is ca rd s i n d em er sa l f is he rie s a nd c om pl y w ith D at a C ol le ct io n F ra m ew or k R eg ul at io n ( D C F) . 20 09 O ng oi ng , no e nd date In dus tr y M in is tr y ( D C F) In du st ry s ur ve y p la ic e a nd s ol e ( IS PS ) Su rv ey f or s to ck ass ess m en t Ev al ua te a de qu ac y o f t he s ci en tif ic b ea m t ra w l S ur ve y (B TS) . 20 09 20 13 In dus tr y In du st ry w ith g ra nt f ro m E FF Pu lse m on ito rin g (Pu ls M on ) C at ch m on ito rin g O bt ai n e st im at es o f d is ca rd s c om po si tio n a nd q ua n-tit y i n f la tf is h p ul se g ea rs t o c om pl y w ith l ic en se re qu iremen ts . 20 11 20 13 In dus tr y In dus tr y C at ch m on ito rin g c od u si ng C C TV ( C C TV ) C at ch m on ito rin g Te st p ot en tia l o f c lo se d-ci rc ui t t el ev is io n ( C C TV ) a s m et ho d f or i m pr ov ed m on ito rin g c od c at ch es i n demer sa l f is her ie s. 201 2 201 5 M ini st ry In du st ry w ith g ra nt f ro m E FF N et i nn ov at io n d em er sa l I ( N IK O -I ) Fi she rie s te ch no lo gy In cr ea se s el ec tiv ity o f d em er sa l f is he rie s t hr ou gh n et in no va tio ns . 20 13 201 5 In dus tr y In du st ry w ith g ra nt f ro m E FF C od m on ito rin g ( C od M on ) C at ch m on ito rin g M on ito rin g o f c od b yc at ch i n o tt er tr aw l ( TR ) g ea rs t o as se ss t he a ct ua l c on ve rs io n f ac to r f or f is hi ng e ff or t (d ay s-at -s ea c or re ct ed f or c od b yc at ch r at es ) b et w ee n TR a nd b ea m t ra w l ( B T) g ea rs i n t he l ig ht o f t he Eu ro pe an C od A vo id an ce P la n. 20 13 20 14 M ini st ry an d In dus tr y M ini st ry B es t p ra ct ic es I ( B P-I) C at ch m on ito rin g Ev al ua te s oc io -e co no m ic c on se qu en ce s o f t he l an di ng ob lig at io n. 20 13 201 5 In dus tr y In du st ry w ith g ra nt f ro m E FF D is ca rd s ur vi va l I ( SU RV -I ) Fi sh s ur vi va l s tu di es D et er m in e c ha nc es o f s ur vi va l o f p la ic e, s ol e, a nd d ab di sc ar ds t o u nd er pi n r eq ue st f or e xe m pt io n o n h ig h su rv iv al u nd er t he E ur op ea n L an di ng O bl ig at io n. In ve st ig at e e ff ec ts o f c at ch in g a nd p ro ce ss in g m od ifi -ca tio ns o n c ha nc es o f s ur vi va l o f d is ca rd s. 20 14 201 5 In dus tr y In du st ry w ith g ra nt f ro m E FF (Co nti nue s)

The industry was convinced that, with a co-management system in place (Hoefnagel & De Vos, 2017), fishing mortality (F) was overes-timated in stock assessments. Managers, industry representatives and scientists jointly decided to start a SIRC project. This so-called F-project aimed at: (a) improving ICES stock assessments, (b) ensur-ing better use of fisheries data in ICES stock assessments, and (c) improving communication and collaboration between the govern-ment, industry and science (Quirijns, Keeken, & Densen, 2007). The latter included capacity building within the industry to understand stock assessments and improving transparency about the process of generating catch advice and setting the catch quota.

This first SIRC, the F-project, was fundamental in establishing the current cooperative relationships between fishers and scientists. Also, it showed fishers that they could make a real contribution to filling in data gaps and providing valuable knowledge for manage-ment. Finally, the F-project increased the understanding of stock as-sessments and the role of science amongst industry representatives and managers. As a consequence, the industry realized that discards estimates for plaice and sole from the scientific observer programme were based on very low sample sizes. Thus in 2004, the industry initiated a discards self-sampling programme (SSDis) with assistance from the fisheries institute. The collaboration in the F-project and SSDis helped scientists to overcome concerns that data collected by fishers would be biased (See also Johannes et al., 2000). Following thorough scientific evaluation, it became clear that the SSDis data contributed to addressing knowledge gaps in a cost-effective way and were an appropriate alternative to the established observer programme (Kraan et al., 2013). Subsequently, the SSDis was inte-grated in the official Dutch discards monitoring programme under the European Union's (EU) Data Collection Framework. By this time, an essential element for the further development of SIRC had been established: trust in each other's contributions (see also Dörner et al., 2015; Mangi et al., 2018).

The F-project in a way became the archetype for SIRC projects in the Netherlands (Table 1). Whereas the F-project and SSDis were mainly aimed at improving data for stock assessments, they opened the door for SIRC projects with different goals. For instance, from 2007 onwards, the observers on-board scientific surveys (OBSS) project has annually taken fishers on board as observers in the scientific fisher-ies-independent beam trawl survey (BTS; see ICES, 2019a). Their partic-ipation convinced the industry that the BTS is generally carried out well. Feedback from fisher observers in the first years resulted in the inclu-sion of three additional sampling stations in the coastal area. However, some changes requested by the industry, such as gear adaptations and changes in existing sampling locations, were not made because they would affect time series of abundance estimates. Whilst fishers under-stood the rationale for not changing the existing survey, they wondered whether a new survey, more in line with the practice of commercial fishing, would lead to different outcomes. As a result they initiated an industry survey for plaice and sole (ISPS). The outcomes from the ISPS data collected from 2011–2015 were compared with the results from research vessel surveys. It became clear that whilst the ISPS caught larger numbers of fish and, on average, fish of a larger length, this did

SI RC p ro jec t Ty pe C om mo n o bj ec tiv es a St ar t En d In iti at or Fu ndin g Pi lo t i nd us tr y s ur ve y d at a-lim ite d s pe ci es (IS D LS) Su rv ey f or s to ck ass ess m en t Te st w he th er i nd us tr y s ur ve y c an i m pr ov e c at ch a nd bi ol og ic al d at a f or s to ck a ss es sm en ts f or d at a-lim ite d fla tf is h s pe ci es ( tu rb ot , b ril l, l em on s ol e) . 20 14 201 5 Fi sh ing com pan y C om pa ny w ith g ra nt f ro m EF F N et i nn ov at io n d em er sa l I I ( N IK O -I I) Fi she rie s te ch no lo gy In cr ea se s el ec tiv ity o f d em er sa l f is he rie s t hr ou gh n et in no va tio ns . 20 16 20 18 In dus tr y In du st ry w ith g ra nt f ro m EM FF D is ca rd s ur vi va l I I ( SU RV -I I) Fi sh s ur vi va l s tu di es D et er m in e c ha nc es o f s ur vi va l o f p la ic e, s ol e, t ur bo t, br ill , t ho rn ba ck a nd b lo nd e r ay d is ca rd s t o u nd er pi n re qu es t f or e xe m pt io n o n h ig h s ur vi va l u nd er t he Eu ro pe an L an di ng O bl ig at io n. I nv es tig at e e ff ec ts o f im pr ov em en ts i n c at ch a nd p ro ce ss in g p ro ce ss o n su rv iv or sh ip . 20 16 20 18 In dus tr y In du st ry w ith g ra nt f ro m EM FF B es t p ra ct ic es I I ( B P-II) C at ch m on ito rin g Ev al ua te s oc io -e co no m ic c on se qu en ce s o f t he Eu ro pe an L an di ng O bl ig at io n. 20 16 20 18 In dus tr y In du st ry w ith g ra nt f ro m EM FF A bb re vi at io ns : D C F, E ur op ea n D at a C ol le ct io n F ra m ew or k R eg ul at io n; E FF , E ur op ea n F is he rie s F un d; E M FF , E ur op ea n M ar iti m e a nd F is he rie s F un d. aIn a dd iti on t o c om m on o bj ec tiv es , o rg an iz at io ns o r i nd iv id ua ls i nv ol ve d i n S IR C s m ay a ls o h av e o w n o bj ec tiv es . T he se a re n ot i nc lu de d i n t hi s t ab le . T A B LE 1  (Co nti nue d)

not affect the inter-annual trends observed in Catch per Unit of Effort time series (Van der Reijden, Poos, Trapman, & Verkempynck, 2016). Consequently, the effect of including the ISPS results in the stock as-sessments was negligible. This built confidence in the adequateness of existing surveys for collecting stock assessment data for plaice and sole, and hence, the industry saw no further need to continue the ISPS.

3.2 | Broadening scope

In the first 10 years, SIRC focused on evaluating and improving data for the plaice and sole stock assessments, and on increasing trans-parency and understanding of the data collection methodologies for these assessments. In its second decade, its scope broadened. Under societal pressure, reducing gear impacts on fish stocks and the seabed became the two priorities. This increased interest in gear selectivity improvements, catch monitoring and discards survival (Table 1).

The changed mode of collaboration also altered the approach taken in fisheries technology research. Until the early 2000s, fish-eries technologists generally came up with solutions that were tested on board research vessels and were then communicated to the fleet. Occasionally, mesh selectivity experiments were done on board chartered fishing vessels with crew assisting in data collection (e.g. Van Beek, Beek, Rijnsdorp, Van, & Leeuwen, 1981), but fishers had limited input in project objectives and execution. Later, research ideas were discussed with fisheries representatives with respect to practicality, acceptability and economy of fisheries. However, actual testing was still done on board research vessels without involve-ment of fishers (e.g. Van Marlen, 2003). Promising fisheries technol-ogy developed by scientists hardly resulted in uptake by the fishing fleets, which is also observed by Eayrs and Pol (2018).

In the new SIRC projects on net innovations (NIKO I and II), fish-ers had input in the complete scientific process, from developing research objectives and designing protocols to interpretation and presentation of results (Table 2). Whilst it is too early to conclude whether or not this inclusive approach to fisheries technology leads to an increase in uptake, first results look promising.

3.3 | Increasing degree of integration of fishers'

(experiential) knowledge

Three more changes relating to the use of fishers' knowledge hap-pened along with the SIRC's broadening scope. Projects crossed horizontal and vertical “boundaries” on the FKR spectrum by Stephenson et al. (2016).

First, 5 years into the start of SIRC, the degree of integration of fishers' knowledge evolved towards more collaborative arrange-ments and participatory governance (Figure 3). With the exception of one (CodMon), none of the new projects limited fishers' roles to data collection only. In some cases (F-project, SSDis, CodMon), active projects crossed boundaries in the horizontal spectrum (Figure 3). Some evolved towards greater integration (F-project,

SSDis). Conversely, in the CodMon SIRC, fishers' involvement ceased altogether. In this project, fishers were involved in a self-sampling scheme as part of the North Sea cod (Gadus morhua, Gadidae) avoid-ance plan. From working together, the joint conclusion was that the cumbersome scheme did not result in useful additional information (Kraan, Machiels, Van der Reijden, & Paijmans, 2014).

Second, fishers' knowledge use shifted vertically on the FKR spectrum. Fishers’ experiential knowledge (Stephenson et al., 2016) became an additional source of information in SIRC. For example, when developing a pilot for an industry survey for data-limited flatfish stocks (ISDLS), the SIRC drew heavily on fishers' intimate knowledge of fishing grounds and fish behaviour to establish the temporal and spatial distribution of turbot (Scophthalmus maximus, Scophthalmidae), brill (Scophthalmus rhombus, Scophthalmidae) and lemon sole (Microstomus kitt, Pleuronectidae). FEXK thus signifi-cantly contributed to survey design. Likewise, the fishing technology SIRC NIKO-I and NIKO-II relied fully on FEXK to generate ideas and to develop and test net adaptations.

Third, interactions between the two types of fishers' knowledge (FOK, FEXK) used in SIRC started to take place. Scientists started to draw on FEXK as a qualitative source to interpret findings. An exam-ple is a SIRC project on fully documented fisheries (CCTV). Fishers received a 30% cod quota uplift in return for participation. Scientists hypothesized that his uplift would result in changes in fishing be-haviour for all participating vessels. However, these changes were not observed for fishers using small vessels. Subsequent interviews with participating fishers helped to interpret the results and gave insights in the decision-making process and reasoning of fishers tar-geting cod (Van Helmond, Chen, Trapman, Kraan, & Poos, 2016).

3.4 | Funding and communication issues

After the first decade of SIRC, funding increasingly became an issue. This was partly caused by institutional changes. In 2011, the national fish product board was dissolved. As a result, there was no longer a collective levy system to raise funds for research. To save costs, the amount of direct contact between scientists and fishers was re-duced. Industry representatives took over coordination of research trips and communicating preliminary findings. This had conse-quences for fishers' involvement in the analyses and joint interpre-tation of results. Table 2 shows an overview of the involvement of different parties in the selected SIRC projects in each project stage identified by Johnson and Van Densen (2007). Whilst fishers were still involved in most stages, the extent of their involvement in data analysis (stage 4) was decreasing. Together with reduced direct con-tacts during data collection (stage 3), this affected data quality in some cases (e.g. self-sampling data during initial gear trials in NIKO-I, CodMon) as well as the input of FEXK during stages 3 and 4. Also, communication of the results to the fleet as a whole came under pressure due to budget constraints.

Communication in SIRC should be a joint responsibility and should have a "clear demarcation between results (neutral) and the

management implications (value-laden)" (Johnson & Van Densen, 2007). However, particularly in the first set of projects (BP-I, NIKO-I, SURV-I) that were set-up in the context of the EU Landing Obligation (LO; see Salomon, Markus, & Dross, 2014), industry representatives tended to take control of communication to their constituents, gov-ernment and the media. They sometimes selectively interpreted SIRC results to support their opinion that the LO was unworkable. In one case, they produced a brochure with “selected results” of SIRCs with-out informing the scientists involved. Likewise, preliminary results of the CCTV SIRC were interpreted by industry representatives to sup-port their case for not using CCTV for compliance purposes under

the LO. In response, in all successive SIRC projects, clear communica-tion agreements were made that have precluded these events. In one case, when the discards survival project was extended (SURV-II), the industry even allocated additional budget for scientists to commu-nicate results to the wider fleet, including a video, infographics and factsheets (www.wur.eu/fishs urvival). This was well received and offered a point of entry for direct discussions with fishers who had not participated in this SIRC, but would be affected by the results. The fishing industry also used these communication materials in the discussions with managers and NGOs as part of (successful) requests for (temporary) exemptions for plaice and sole from the LO.

TA B L E 2   Stakeholder involvement for selected SIRC projects

SIRC project Type of SIRC Stage 1a _{Stage 2}a _{Stage 3}a _{Stage 4}a _{Stage 5}a

F-project - WP2 Better use of fisheries data (F-project)

Catch monitoring for stock assessment

F, FR, G FR, F, G F, FR, G, SU FR, F, G, SU, WF FR, F, G, SU, WFb Self-sampling of discards in

demer-sal fisheries (SSDis)

Discards monitoring for stock assessment

FR F, FR, SU F, FR FR. SU F, FR, G,

SU, WFb Observers on-board scientific

surveys (OBSS)

Survey for stock assessment

F, FR, SU F, FR, SU F, SU SU F, FRa_{, SU,}

WFb Self-sampling of discards in

demer-sal fisheries (SSDis-DCF) Discards monitoring for stock assessment F, FR, G SU F, FR F F, FR

a_{, SU}

Industry survey plaice and sole (ISPS)

Survey for stock assessment

FR, SU F, FR, SU F, SU F, FR, SU F, FR, SU,

WFb

Pulse monitoring (PulsMon) Catch monitoring FR, SU F, SU F, FR FR, SU Fa_{, FR, G}b_,

SU, WFb Catch monitoring cod using CCTV

(CCTV)

Catch monitoring G, SU FR, G, SU F, SU F, FR, SU FR, F, G,

SU, WFb Net innovation demersal I (NIKO-I) Fisheries technology FR, SU F, FR, SU F, FR, SU F, FR, SU FR, F, G,

SU, WFb

Cod monitoring (CodMon) Catch monitoring FR, G FR F G, FR G, FR, WFb

Best practices I (BP-I) Catch monitoring FR FR F, FR FR, G, SU F, FR, G,

NGOa_{, SU,} WFb

Discard survival I (SURV-I) Fish survival studies FR F, FR F, FR F, FR, SU F, FR, Gb_,

NGOb_{, SU,} WFb Pilot industry survey data-limited

species (ISDLS)

Survey for stock assessment FC, NGO F, FC, SU, NGO F, FC, NGO F, FC, NGO, SU F, FC, G, NGO, SU, WFb Net innovation demersal II (NIKO-II) Fisheries technology FR, SU F, FR, SU F, FR, SU F, FR, SU FR, F, G,

SU, WF*

Discard survival II (SURV-II) Fish survival studies FR F, FR F, FR F, FR, SU F, FR, Gb_,

NGOb_{, SU,} WFb

Best practices II (BP-II) Catch monitoring FR FR F, FR FR, G, SU F, FR, G,

NGOb_{, SU,} WFb

Note: In all stages for each project, the scientific project leader was involved and not separately included in the table.

Abbreviations: F, fisher; FC, fishing company; FR, fishers' representatives; G, government; NGO, environmental non-governmental organization; SU, scientific end-User; WF, whole fleet.

a_{5 project stages of SIRC: (1) problem definition, (2) project design, (3) data collection, (4) data analysis and (5) communication or results (Johnson &} Van Densen, 2007).

In relation to funding, another challenge emerged when the government introduced an obligation to include budget offers from three scientific institutions as part of industry project grant pro-posals. Consequently, researchers involved in SIRC could not work together with industry in formulating exact research questions and an appropriate research approach (project stages 1 and 2, Johnson & Van Densen, 2007). Such collaboration would imply having “prior knowledge” and thus a not allowed competitive advantage over others in the tender procedure. This government decision violated the very principles of successful SIRC. As a result, SIRC projects from those proposals were not necessarily fully fit for purpose at the start, whilst research design changes were difficult due to grant regulations. Following discussions, the “three scientific tenders” rule was abolished in the consecutive grant schemes.

3.5 | Arrival of industry scientists

In 2014, a new dimension was added to the demersal SIRC. Following the example of the pelagic fleet, one of the national

demersal fisheries associations hired two scientists as part of their staff. Both were former employees of the national fisheries institute. These “industry scientists” engaged with the “institute scientists.” They took a visible role in writing research proposals for fisheries innovation funds, structuring data collected by fishers to support fleet-wide analyses, carrying out own research, interpreting (inter-mediate) results of SIRC projects, and communicating with fishers. “Institute scientists” were initially concerned by this development. The industry scientists were seen as potential competitors in the day-to-day communication with fishers, potentially even affecting the relationship and trust that had been carefully built. However, the good relationship between the industry representatives, industry scientists and institute scientists enabled an open discussion about these concerns.

Having scientists working in the industry, as opposed to working with the industry, was considered as, and also proved to be, an added value in the translation of fishers' knowledge into the scientific do-main and discussing results with fishers and industry representa-tives. It was agreed that the industry scientists could participate in ICES working groups as national representatives. This led to some

F I G U R E 3   Dutch SIRC projects crossing boundaries in the fishers' knowledge spectrum (Stephenson et al., 2016). Arrows show shifts in the

positioning in the spectrum between start and completion of the project. Numbers between brackets are years. Key to project abbreviations, in alphabetical order (See Table 1 for project overview): BP, best practices (for implementation of EU landing obligation); CCTV, catch monitoring of cod using closed-circuit television; CodMon, cod catch monitoring; F-project, better use of fisheries data in stock assessments; ISDLS, industry survey data-limited species; ISPS, industry survey plaice and sole; NIKO, net innovation demersal fisheries; OBSS = (fisher) observers on-board scientific surveys; PulsMon, pulse fishery catch monitoring; SSDis, sampling of discards in demersal fisheries; SSDis-DCF, self-sampling of discards in demersal fisheries integrated in Dutch EU Data Collection Framework programme; and SURV, fish survival studies

discussion within ICES. ICES now has a code of conduct for all ex-perts aimed at minimising any risk of actual, potential or perceived conflict of interest (ICES, 2018a).

3.6 | Participation of NGOS

A final development in the evolution of the Dutch SIRC is the ac-tive participation of NGOs. Whilst generally, Dutch NGOs and the industry are on speaking terms and sometimes cooperate in the policy domain, having the NGOs actively involved in SIRC was considered one step too far by the industry. There was one excep-tion, a pilot industry survey for data-limited stocks (ISDLS). This was initiated by a private fishing company in partnership with the North Sea Foundation as part of the fishers' aspiration of achiev-ing Marine Stewardship Council certification for turbot and brill. This project yielded results which were supported by both the in-dustry and the NGO, but was not continued due to lack of financial support. Five years later, in 2018, this ISDLS project formed the basis for the start of a full (ongoing) industry survey for turbot and brill.

On request of the government, discussion of progress and pre-liminary results of all SIRC projects in relation to the landing obliga-tion (NIKO, BP and SURV) became a part of regular meetings with industry representatives, NGOs and scientists. The NGOs, through this government initiative, thus became involved in SIRC, albeit in the last stage only. This paved the path for further integration of NGO representation in SIRC. Furthermore, cooperation with NGOs was encouraged by the government by allocating more money to proj-ects that included NGOs. In most SIRC projproj-ects from 2018 onwards (hereafter referred to as “SIRC 2.0”), NGOs are represented from stage 1. However, NGO involvement remains a balancing act. Whilst a sufficient level of trust between industry and science has been de-veloped over the years, this is neither yet the case between science and NGOs nor between industry and NGOs in relation to access to data and communication. In addition, the history of cooperation be-tween industry and research means both parties developed "inter-actional expertise" (Collins & Evans, 2008; Stange et al., 2014). With the NGOs now joining the SIRC table, this interactional expertise will need to be built up again within this new group composition (Stange et al., 2014). This is why for all SIRC 2.0 projects a contract is signed between the consortium partners (industry organizations, scien-tific institute(s), NGOs). This contract excludes the government that provides research grants, but is not a project partner. The contract comprises rules on data sharing and a communications protocol. For example, in the case of industry surveys, data will be uploaded to the ICES DATRAS database. In other cases, vessel owners mandate the research institute to store commercial catch data and video im-ages and to make them available for research. Vessel owners can withdraw this consent at all times. Furthermore, all public communi-cation about projects by individual partners have to be signed off by all project partners. Whilst it is too early to conclude how well the contract works in relation to data sharing, fishers feel that they are

“in control” of data collected by them. The agreed communications protocol has so far been followed without issues.

4 | OUR LESSONS FROM SIRC

We can identify clear stages in over 15 years of evolution of Dutch SIRC (2002–2018) (Figure 4). SIRC started with a focus on using fish-ers as platform providfish-ers for data collection and on improving their understanding of research methodology and interpretation of find-ings. This evolved to collaborative arrangements in setting up and carrying out research, with a focus on filling data gaps. Through the SIRC projects, awareness of the value of fishers’ experiential knowl-edge developed, and industry became aware about how (fishers') information feeds into science and management advice. This led to an increased call for additional research, and the realization that em-ploying scientists would improve the knowledge base for manage-ment. The involvement of industry scientists and NGOs marks the transition from “SIRC 1.0” to “SIRC 2.0.” We identify four key (inter-related) lessons from these developments:

1. Capacity building in SIRC is multidirectional, leading to shifts in the FKR spectrum.

2. Embedding end-users of data is crucial for successful SIRC. 3. Equity questions and selection bias need to be considered. 4. The governance system impacts structure and integration of

SIRC.

Each lesson is elaborated below.

4.1 | Lesson 1: Capacity building is multidirectional,

leading to shifts in the FKR spectrum

In literature on collaborative fisheries research, capacity building is seen as a benefit to the industry (Hartley & Robertson, 2006; Johnson & Van Densen, 2007; Kaplan & McCay, 2004; Mackinson et al., 2011). Fishers participating in collaborative projects are enabled to get new roles and agency as knowledge actors, progressing from “knowledge holders” to “knowledge agents,” and resulting in what has been coined as the "scientific fisherman" (Dubois, Hadjimichael, & Raakjaer, 2016). Fishers subsequently are also able to use this interactional expertise (Stange et al., 2014) in the management arena (Holm et al., in press). Our SIRC analysis confirms this. Examples include the industry's deci-sion to initiate a self-sampling programme for discards (SSDis), and their request to develop popular communications materials on the results of discards survival studies (SURV-II) to assist discussions amongst fishers and with managers and NGOs. But our analysis also shows that ca-pacity building is not unidirectional. Scientists equally benefit as their understanding of how fishers observe and interpret changes in the fishery and the ecosystem deepens. Examples include the co-design of industry surveys (ISDLS and its follow-up), and the evaluation of socio-economic consequences of the LO (BP-I, BP-II). The positive role

collaborative research plays in capacity building of scientists was also observed by Thompson et al. (2019) in their evaluation of the Canadian Research Fisheries Network. This two-way capacity building trans-lates into increased integration of fishers' knowledge when developing SIRC projects, and a growing use of FEXK when validating results and research design. Integration of FEXK follows when trust and respect are established and when scientists acknowledge that fishers can play a valuable role in interpreting results and optimizing research design. However, the use of FEXK in fish stock assessments and impact stud-ies remains a challenge (See also lesson 4, and Johannes et al., 2000; Mangi et al., 2018; Holm et al., in press).

4.2 | Lesson 2: Embedding end-users is crucial for

successful SIRC

Data collected by the industry for management purposes is often met with suspicion; there is concern about potential bias in data col-lected by those who have a vested interest in them (Kraan et al., 2013; Mangi et al., 2018). SIRC projects provide a forum to discuss this concern and to develop checks and balances to ensure that qual-ity of data collected by fishers can pass scrutiny in scientific peer review. In addition to adding these checks and balances, scientists involved in Dutch SIRC also engaged in discussions with their peers about the use and added value of fisher's information. However, whilst data generated in SIRC may have been collected in accord-ance with scientific standards, this does not automatically mean that these data are actually used (See also Mangi et al., 2018). For exam-ple, the Landings per Unit of Effort series for plaice and sole gener-ated in the F-project never made it into the ICES stock assessment as was intended. The main reason was the limited commitment in the stock assessment working group to use industry data. This may have been reinforced by the (at that time) lack of clear processes to deal with this data.

In our SIRC 2.0 projects, the need for early involvement of sci-entific users of the data was a lesson that was put into action. As part of the SIRC process, draft designs for industry surveys and catch monitoring schemes are presented by industry scientists as full members of ICES stock assessment working groups. Feedback

from the groups is incorporated in research protocols in order to meet expectations of its eventual end-users. In the meantime, ICES also realized that a process for evaluating available data (irrespective of the source) is needed before they are used in assessments: the benchmark process. By being clear about data standards, in combi-nation with an evaluation process, “data trust” now seems less of an issue (See also Mangi et al., 2018). Data from Dutch SIRC projects have, however, not yet been supplied in benchmarks, and effective-ness of the benchmark process for the inclusion of these data cannot be assessed at present.

The International Council for the Exploration of the Sea has plans for the improvement of the data quality assurance system within the benchmark process. We recommend that as part of these develop-ments, particular attention is paid to data from SIRC. Here, the joint science-industry data collection protocol developed by Mangi et al. (2018) in the context of SIRC in the United Kingdom can provide useful guidance. Similarly, the discussions in ICES about the use of quality controlled industry data could benefit from a review of ex-periences elsewhere. New Zealand, for example, has implemented a Research and Science Information Standard (Ministry of Primaries Industry, 2011), defining best practice in delivering quality-assured research irrespective of who provided the information. As long as the standard is met, knowledge generated by the fishing industry (or other stakeholders) can become part of the evidence base informing management (Mackinson & Middleton, 2018).

With an increasing number of SIRC initiatives in Europe, the challenge for ICES is now to ensure that data from these initiatives find their way into assessments and other ICES science. If data from SIRCs that pass peer review are not embraced by the ICES commu-nity or other science organizations, this will undermine SIRC and carefully built trust (See also Mangi et al., 2018). Furthermore, it may affect the position of institutes and science organizations that strive to give advice on sustainable marine resource use based on the best available scientific knowledge. In response to these inter-national SIRC developments, ICES organized a first Workshop on Science with Industry Initiatives (WKSCINDI) in 2019. The workshop produced an overview of the potential roles of industry in providing scientific information relevant to ICES advice and marine research, and a roadmap for a stepwise incorporation and application of

quality-assured scientific data from industry. WKSCINDI stressed that careful attention should be paid to the social and practical pro-cesses involved. It emphasized that co-design between science, in-dustry and managers in the planning stage is fundamental to ensure science and advisory needs are recognized and matched with prac-tical plans to meet them (ICES, 2019). The potential contribution of SIRC to improve stock assessments was also acknowledged in the ICES Workshop on Research Roadmap For Mackerel (WKRRMAC). In particular, WKRRMAC recommended to "build mechanisms to in-corporate industry sampling of biological information into the formal stock assessment process and develop approaches for formalising the flow of information of industry perceptions of the state of the stock into the assessment process" and "develop credible methods for industry surveys" (ICES, 2019). In this context, the ICES code of conduct for all experts, no matter their affiliation (ICES, 2018a), is another important signal in acknowledging that fisheries and marine science are changing to include more collaborative approaches.

As part of all efforts towards the inclusion and application of industry data in fisheries and marine science, it is essential that ex-pectation management towards the industry remains key. This as-pect needs continuous attention. In our SIRC projects, we see that even whilst fishers are well informed about data standards and the need to collect data over longer time periods, they often find that the translation of data into management measures is going too slow. In this light, it is important that scientists and managers realize that SIRC projects often cover relatively short time spans due to funding restrictions. This is at odds with the requirements for long time se-ries in stock assessments (Mangi et al., 2018).

4.3 | Lesson 3: Equity questions and selection BIAS

need to be considered

Those involved in SIRC should be aware that fishers' knowledge research needs to consider the potential risks of selection bias and often-overlooked equity questions (Reed, 2008; Steins & Edwards, 1999). We tend to talk about “fishers,” whilst in reality, the group ac-tively participating in SIRC is small compared with the actual number of fishers. We found that participating fishers are usually selected by fisheries representatives or by scientists based on experiences in previous SIRCs. The rationale is often that these fishers have proven to be serious and reliable and like to cooperate with scientists. Whilst this rationale is understandable, it results in selection bias and in a large number of fishers who miss out on actively contribut-ing to knowledge development and capacity buildcontribut-ing. Selection bias is an issue for any data collection programme.

Equity questions arise because participation in SIRC projects is as-sociated with benefits. These benefits include a better understanding of science and fisheries management, easier access to scientists who can answer ad hoc questions, financial or quota incentives, opportuni-ties to go on study trips abroad and access to new projects or side-jobs. Fishers included in collaborative projects become “scientific fishers” (Dubois et al., 2016). For example, one participating fisher from various

Dutch SIRC projects is now a regular speaker at conferences and has side-jobs in the maritime industry. Participating in SIRC projects allowed him to develop a network and new skills, and opportunities arose. Fishers are well aware of potential benefits generated by partic-ipation in projects. This may lead to envy in the fleet and sometimes to tensions in small fishing communities. We should therefore ensure that access to SIRC is not limited to an exclusive club.

4.4 | Lesson 4: The governance system impacts

structure and integration of SIRC

The governance context in which SIRC takes place impacts the struc-ture of SIRC (Röckmann, Leeuwen, Goldsborough, Kraan, & Piet, 2015; Steins & Edwards, 1999) and the integration of SIRC knowledge (Holm et al., in press; Stephenson et al., 2016). Examples of the im-pact of the governance context on Dutch SIRCs include the long-term availability of financial support for SIRC projects aimed at fostering a culture of collaboration, the redesign of funding rules to encourage NGO participation in SIRCs, and the government's decision to formally include the industry's discards self-sampling in the national research programme. Science and management of Dutch demersal fisheries is, furthermore, embedded in the EU's Common Fisheries Policy (CFP), a multilevel governance system (Piattoni, 2009). In Europe, collabo-rative research between science and fishing industry from different member states has been facilitated in recent years as part of a shift-ing policy discourse (Holm et al., in press), and is gainshift-ing momentum. The challenges associated with large multilevel governance systems in relation to collaborative research, however, remain to be addressed.

Large multilevel governance systems for fisheries, such as the CFP, are typically characterized by a considerable degree of standardization of scientific procedures, feeding information into a separate central decision-making process on management. Holm and Nielsen (2004) metaphorically use the term "Total Allowable Catch (TAC) machine" to describe this institutionalised relationship. This “TAC machine” strongly limits the possibility to incorporate context-specific informa-tion, both in the “science pillar” (for example, data provided by industry to ICES) and the “management pillar.” However, collaboration between fishers and scientists (and often also managers) typically takes place at lower (national) levels, tends to highlight specificities and detailed local knowledge and creates appreciation for heterogeneity in the local con-text (of habitats, species behaviour and fisher behaviour).

Furthermore, policy and regulations are shaped through inter-actions between multiple institutional levels. These interinter-actions are often driven by contradictory goals held by different actors, whilst final decision-making is supra-national (Burns & Stöhr, 2011; Steenbergen, Trapman, Steins, & Poos, 2017). To illustrate, in order to alleviate the impacts of the CFP's LO, the Dutch government, in-dustry, and NGOs agreed on a pragmatic implementation. SIRC proj-ects have been set-up in support of this objective (BP-I and II, NIKO-I and II, SURV-I and II). The fishing industry and fishers expect the results to be used in the policy process. However, the government operates at the interface between national interests and European

objectives. It exercises its own "room for manoeuvre" (Long, 2001). That means that if the national government sees an opportunity to use SIRC outcomes to alleviate impacts of the LO, they will use them to get European agreement on specific management measures (e.g. an exemption on the basis of high discards survival). However, if the national government expects SIRC outcomes are associated with high political risk, they will remain passive (Van Hoof et al., 2019). This means that expectations raised at the outset of SIRC projects in support of policy implementation may not be met, potentially risking industry and NGO support of SIRC or management measures.

Finally, the European multilevel governance system and its frame-work of co-decision between the European Parliament, Council and Commission, brings challenges as to whom are the end-users of col-laborative projects and how to ensure their support and commit-ment. This means that expectation management about outcomes towards fishers involved in collaborative research in a multilevel governance context becomes more complex.

5 | RECOMMENDATIONS FOR

COLL ABOR ATIVE MARINE SCIENCE AND

MANAGEMENT

For successful collaborative research between scientists and stakeholders, context is important: Is there already a history of research cooperation? Is the management system receptive or participatory? Is traditional (fisheries or marine) science open to new types of knowledge, including qualitative knowledge? What is the level of involvement of non-active stakeholders who are af-fected by results of collaborative research projects? Is collabora-tive research embedded solely in a local context or in a multilevel governance system? These questions all play a role in the set-up of collaborative research projects and the uptake of their results. We have three recommendations that partly address the four lessons from our evaluation of SIRC in Dutch demersal fisheries. Our first recommendation is to ensure that participation and collaboration in research is embedded in institutional structures for science and resource management. Our second recommendation is that those involved in collaborative research should continuously reflect on selection bias and equity questions. Our final recommendation is that communication should not simply be an afterthought; in-tensive communication with everyone involved is necessary and should be an integral part of SIRC (See also Johnson & Van Densen, 2007; Mangi et al., 2018).

5.1 | Recommendation 1: Embed collaborative

research in institutional framework

It is essential that collaborative research projects are embedded in the institutional framework within which they take place. If they take place in isolation, then they will not likely move beyond being a tool for improving relationships amongst involved parties (Holm

et al., in press; Kaplan & McCay, 2004; Steins & Edwards, 1999). If outcomes are not used in science or policy, there is a risk that these relationships erode.

Sometimes existing science frameworks need to be adjusted to ensure that SIRC data can be taken up and meet agreed qual-ity standards and formats. For scientists involved in collaborative projects that intend to deliver results to a science community, this means acting as knowledge brokers (Pielke, 2007) and communica-tors or gatekeepers (Mangi et al., 2018). This finding is supported by an ICES workshop that identified factors influencing the uptake of science into advice (ICES, 2018b). The knowledge broker and com-municator or gatekeeper roles are not traditional roles of scientists. However, in a context of wicked problems, individual scientists and marine science communities must reflect on the interface between society (including industry) and science. In the case of ICES, im-portant first steps have already been taken in this respect, includ-ing the adoption of a benchmark process for the inclusion of new data, a code of conduct for all experts involved in ICES work and dedicated workshops involving science, industry and managers on the co-creation of data to inform the scientific and advisory process (ICES WKSCINDI, ICES WKRRMAC). Experiences outside Europe could provide useful guidance in setting up more collaborative ap-proaches for evidence-based marine resource management. For example, Mackinson and Middleton (2018) propose that the New Zealand governance system, which in contrast to the EU framework is characterized by short pathways, fewer people and a unilateral de-cision-making process, presents “a good case to learn about inclusive governance of fisheries.”

The governance system also has to embrace these new forms of collaborative research. Making sure that managers are at least aware and at best supportive of such research is important. Otherwise projects run the risk of “hitting the management wall” (Holm et al., in press). In situations of multilevel governance char-acterized by a “TAC machine” setting (Holm & Nielsen, 2004), such as in the EU, this is not an easy task. In such governance settings, the ability to identify the relevant people (managers, NGOs, in-dustries), set-up a dialogue and align these different stakeholders with the implementation of collaborative research is hindered sig-nificantly by increasing numbers of actors, policy processes and interactions between them (See also Haasnoot, Kraan, & Bush, 2016; Steenbergen et al., 2017; Steins & Edwards, 1999; Van Hoof et al., 2019). Furthermore, policy reforms may be needed to in-clude these new forms of collaboration (Holm et al., in press). Such policy reforms may be impossible, or are time-consuming and not necessarily lead to objectives held at the onset (Wakefield, 2016). However, if "TAC machine" systems fail to adapt to include knowl-edge from collaborative research in science and management, this may eventually lead to diminishing support for collaborative research and eroding trust (See also Holm et al., in press). This would be at odds with the benefits associated with SIRC, such as cost-effective, improved data collection, increases transparency and communication, capacity building amongst fishers and scien-tists and improved societal relevance of research.

5.2 | Recommendation 2: Keep reflecting on

selection BIAS and equity questions

During the course of collaborative research, it is important that all involved keep reflecting on who is participating and what this par-ticipation means for project design, communication and outcomes. When research collaboration between science and stakeholders is in its early days, it makes sense to select fishers or other stakehold-ers who are open to new forms of collaboration and are seen as forerunners. These stakeholders can act as ambassadors amongst their peers and are crucial in developing a culture of collabora-tion. However, particularly when research collaboration becomes established, it is important to ensure that participation is not by default limited to "the usual suspects," even though this may be attractive (e.g. established relationship of trust, reliable partici-pant). Going for the "safe route" has the inherent risks of selection bias and of (unintentionally) fostering inequity. This implies, first, that when developing new collaborative projects, joint discussions on research design should include transparent criteria for select-ing participants (e.g. Reed et al., 2008). Second, it requires that in data analysis and reporting, sufficient attention is paid to potential impacts of stakeholder selection on results. Finally, it means that those involved in collaborative research should at all times be open and willing to discuss potential equity issues arising from collabora-tive projects.

5.3 | Recommendation 3: Make communication an

integral part of SIRC

Our final recommendation is that intensive communication with all parties involved should be an integral part of SIRC from the outset. Communication in all project stages is also core in the guidelines for SIRC by Johnson and Van Densen (2007). We found, however, that even if fishers and end-users are only involved in a few stages, fruitful SIRCs are possible. For example, in most of our SIRC projects, fishers were not involved in problem definition and data analysis (Table 2). Yet, project outcomes were supported by them. Meanwhile, we saw that in projects where fishers were only involved in research design or data interpretation, they were making a real difference. Hence, it is not the case that "more com-munication and more involvement is always better;” the context defines how well SIRCs work. Johnson & Van Densen's guidelines should therefore not be used as a blueprint, but as guidance. In using this guidance, we advise that communication should ad-dress the following points:

1. Be clear about objectives, including agreement on joint objec-tives and mutual awareness of individual goals.

2. Jointly agree on a research approach and be clear about why this approach and use of (fisher) knowledge was chosen, including its limitations.

3. Be clear about and respect the different roles of the different parties in the project and in subsequent translation of science to advice.

4. Manage expectations in terms of uptake of results, particularly when time series are required.

5. Allocate sufficient time for communication in projects, both with the "core participant group" in the project as with the wider group (through dedicated (social) media and at annual meetings). 6. Develop a communications protocol.

7. Address "elephants in the room" by making dilemmas explicit and discussing them.

8. In cases where the project takes place in a multilevel governance context, communication should not be limited to the local or op-erational level, but should involve end-users from the outset.

6 | CONCLUSION

The management of marine resources is becoming increasingly complex. This calls for stronger stakeholder participation in advanc-ing the knowledge base that underpins management approaches. Here, much can be learnt from experiences with collaborative re-search in fisheries. We re-iterate that collaboration between scien-tists and stakeholders takes hard work. A sustainable, continuous process of collaboration including meaningful uptake of stakehold-ers' knowledge in marine resources research, requires continuous maintenance of the relationship between scientists and stakehold-ers in the context of management developments that typically take place at multiple levels. The lessons and recommendations from our evaluation of SIRC in Dutch demersal fisheries can assist the broader integration of stakeholder knowledge in marine science and management where stakeholder involvement in the co-crea-tion of knowledge is necessary but not yet common.

ACKNOWLEDGEMENTS

This article could not have been written without those involved in science-industry research collaboration. First and foremost, we thank the fishers, their representatives from Coöperatieve Visserij Organisatie, Nederlandse Vissersbond and VisNed (formerly: Federatie van Visserijverenigingen), the fishing company Ekofish Group and all our colleagues who are involved in SIRCs on a daily basis. In particular, we thank our (former) colleagues Jurgen Batsleer, Ingeborg de Booijs, Pieke Molenaar, Mascha Rasenberg, Edward Schram, Ruben Verkempynck, Edwin Van Helmond and Harriët Van Overzee. The Dutch Ministry of Agriculture, Nature and Food Quality and the aforementioned fisheries organizations and com-pany (co)funded some of the SIRCs, whilst others received grants from the former European Fisheries Fund or the current European Maritime and Fisheries Fund. Finally, we thank the reviewers for their helpful comments.