On “success” in applied environmental research – What is it, how can it be achieved, and how does one know when it has been achieved?

1 On “success” in applied environmental research – What is it, how can it be achieved, and how does 2 one know when it has been achieved? 3 4 In Prep for Environmental Reviews as a Perspective Article 5 6 Steven J. Cooke1,*, Trina Rytwinski1, Jessica J. Taylor1, Elizabeth Nyboer1, Vivian M. Nguyen1, Joseph R. 7 Bennett1, Nathan Young2, Susan Aitken3, Graeme Auld4, John-Francis Lane1, Kent A. Prior5, Karen E. 8 Smokorowski6, Paul A. Smith7, Aerin L. Jacob8, David R. Browne9, Jules M. Blais10, Jeremy T. Kerr10, Banu 9 Ormeci11, Steven M. Alexander12, Christopher R. Burn13, Rachel T. Buxton1, Diane M. Orihel14, Jesse C. 10 Vermaire3,13, Dennis L. Murray15, Patrice Simon7, Kate A. Edwards16, John Clarke16, Marguerite A. 11 Xenopoulos15, Irene Gregory-Eaves17, Elena M. Bennett18 and John P. Smol19. 12 13 1 Canadian Centre for Evidence-Based Conservation, Department of Biology and Institute of 14 Environmental and Interdisciplinary Science, Carleton University, 1125 Colonel By Dr., Ottawa, ON, K1S 15 5B6, Canada 16 2 School of Sociological and Anthropological Studies, University of Ottawa, 120 University Private, 17 Ottawa, ON, K1N 6N5, Canada 18 3 Institute of Environmental and Interdisciplinary Science, Carleton University, 1125 Colonel By Dr., 19 Ottawa, ON, K1S 5B6, Canada 20 4 School of Public Policy and Administration, Carleton University, 1125 Colonel By Dr., Ottawa, ON, K1S 21 5B6, Canada 22 5 Parks Canada, 30 Victoria (PC-03-C), 3-84, Gatineau, QC, J8X 0B3, Canada 23 6 Great Lakes Laboratory for Fisheries and Aquatic Sciences. Fisheries and Oceans Canada. 1219 Queen 24 St. E., Sault Ste. Marie, ON, P6A 2E5, Canada 25 7 Wildlife Research Division, Environment and Climate Change Canada, National Wildlife Research 26 Centre, 1125 Colonel By Dr., Ottawa, ON, K1S 5B6, Canada 27 8 Yellowstone to Yukon Conservation Initiative, 200-1350 Railway Ave, Canmore, AB, T1W 1P6, Canada 28 9 Canadian Wildlife Federation, 350 Michael Cowpland Drive, Kanata, ON, K2M 2W1, Canada 29 10Department of Biology, University of Ottawa, 30 Marie Curie Pte, Ottawa ON, K1N 6N5, Canada 30 11 Department of Civil and Environmental Engineering, Carleton University, 1125 Colonel By Drive, 31 Ottawa ON K1S 5B6 32 12 Environment and Biodiversity Science, Fisheries and Oceans Canada. 200 Kent Street, Ottawa, ON, 33 K1A 0E6, Canada Page 1 of 35


74
Applied environmental research is critical for understanding and solving the complex environmental 75 problems of the Anthropocene (Crutzen 2006). From reducing carbon emissions to developing 76 sustainable fish harvesting methods to restoring degraded habitats, decision-makers and environmental 77 practitioners can struggle to make good decisions (Costanza and Jorgensen 2002). Some challenges are 78 truly global (consider the UN Sustainable Development Goals), while others are specific to a taxon (e.g., 79 how to recover an endangered frog population), issue (e.g., where to site wind turbines to reduce 80 impacts on wildlife), or location (e.g., what is the trajectory of permafrost near a given mine site). No 81 matter the scale of the problem, it is difficult to make effective decisions without using evidence (Dicks 82 et al. 2004). To that end, researchers in academia, government, industry, and non-profit organizations 83 conduct studies that aim to generate new knowledge and deepen our understanding of environmental 84 issues and problems.

85
Given that financial and logistical resources for conducting research are normally limited, applied 86 research must generate information and knowledge that is truly relevant and useful to decision-makers 87 and practitioners, and thus leads to more effective outcomes (Milner-Gulland et al. 2012). Moreover, 88 because evidence demonstrates that many current environmental issues are crises (e.g., climate change, 89 biodiversity loss, plastic pollution -see Ripple et al. 2017), delivering actionable science is urgent. 90 Although fundamental research plays a crucial role in understanding environmental problems and 91 identifying solutions and can sometimes be applied in unexpected, immediate ways (Lederman 1994;92 see Littlewood et al. 2012 for an example of fundamental science contributing to improved 93 management of grasslands or Burnett et al. 2017 for an example of how the fundamental concept of 94 carryover effects was used in an adaptive management study to benefit salmon passage at a dam), 95 actionable findings generally arise from research with the explicit goal of informing policy and practice. 96 As such, applied environmental research that aims to inform policy or practice but fails to do so is a 97 waste of resources, and may put species, ecosystems, or people at risk. Quite simply, environmental 98 decision-makers and practitioners need and deserve high quality, applicable environmental evidence. 99 Evidence can take many forms, but for the purpose of this paper, we focus on scientific knowledge, 100 using the term broadly to span the social and natural sciences. We acknowledge and respect the role of 101 other ways of knowing (e.g., Indigenous knowledge, local knowledge), but here we focus on knowledge 102 generation that uses the scientific method or other forms of "western" scholarship -whether 103 qualitative or quantitative, experimental or observational, empirical, or modeling. Indeed, this may 104 involve social science or ethnographic studies of other knowledge holders. Other ways of knowing are 105 beyond the scope of our collective expertise; we encourage decision-makers to engage with experts in 106 those fields.

107
Although most environmental researchers spend many years in university, formal training for 108 environmental researchers to apply their skills in applied ways are scarce (Touval and  is also a growing number of funding opportunities for those working on applied environmental issues, 114 and certainly one of the best ways to master this skill is simply to practice and learn from trial and error 115 (Cooke 2019). Such practices can be cumbersome and delays caused by this more complex way of 5 116 working compound the challenges of addressing already difficult environmental problems (Martin et al. 117 2012). Moreover, a research project that fails to deliver useful information (note -failure can arise from 118 issues with the research itself or that the knowledge generated was not used or a combination of the 119 two) can jeopardize future funding, negatively affect environmental outcomes for policy, and cause 120 stakeholders to limit future, partnered research.

121
There are some resources for applied environmental researchers that share perspectives on how to be 122 successful in applied environmental research, but rarely have they been collated in a peer reviewed 123 paper. Laurance et al. (2012) highlight some strategies for scientists to design and undertake research 124 that should help conservation practitioners. Specifically, they identify the importance of producing 125 time-critical research, attacking "wicked" problems, using multidisciplinary approaches, and better 126 communicating their findings. Moore  practical steps intended to enhance the impact of environmental science on decision-making (i.e., (1) 131 identify and understand your audience (or partners); (2) clarify the need for evidence; (3) gather "just 132 enough" evidence; and (4) share and discuss the evidence).

133
There is also extensive literature on how to bridge the knowledge-action divide more broadly. Cash et 134 al. (2003) and Cook et al. (2013) suggest that for evidence to be actionable it needs to be salient 135 (relevant and timely), credible (authoritative, believable, and trusted), and legitimate (developed via a 136 process that considers the values and perspectives of all relevant actors). It is also intuitive that the 137 evidence needs to be "correct", reproducible, and repeatable (Baker and Penny 2016 intuitive that success is measured in terms of the ultimate outcomes. That is, did the research address 203 key science needed to inform action to resolve or ameliorate an environmental problem so as to reduce 204 its negative environmental, cultural, health, economic, and/or social impacts (Wall et al. 2017) (including 205 tangential benefits such as raising public awareness)? Yet, it is also clear that regardless of whether the 206 ultimate outcome is achieved, the process by which the science is conducted and how it is shared is also 207 important (Nel et al. 2016). Certainly the research needs to be rigorous and deemed to be of high 208 quality but that alone is insufficient for success. When one thinks of success in terms of the broader 209 research ecosystem, it can be achieved (or not) in various components that are visualized in Figure 1 Failure to recognize the importance of engagement in the research process will mean that even the 219 most rigorous science has a strong likelihood of being ignored (Young et al. 2016b). Relatedly, when 220 success is viewed solely from the perspective of the knowledge generator, there can be a disconnect 221 with socio-political issues.

222
Between the research processes and the proximate and ultimate measures of impact is the so-called 223 knowledge-action gap (Cook et al. 2013) where there exist many barriers to uptake even when new 224 knowledge is in the hands of decision-makers and practitioners ( Figure 1). As such, our discussions 225 elicited the idea that the minimum/lowest threshold for success is that the research findings contribute 226 to the knowledge base by being accessible, understandable, and shared, thus creating the potential for 227 change. Assuming that the science was done in a way that is respectful (Shackeroff and Campbell 2007), 228 it must then be clearly communicated such that findings are delivered to relevant parties in ways that 229 can be useful and understandable. Accessibility is important. If findings are communicated, but end 230 users can not find the data or peer reviewed papers, findings may be ignored (Cook et al. 2013 The creation of knowledge-based products can inform policy and practice (Possingham et al. 2001). 235 Collectively these actions can establish the potential for outcome/change and result in project-specific 236 proximate and/or ultimate outcomes. For some projects, impact may be viewed in broad, almost vague 237 terms, such as training the next generation of scientists, publishing papers, or changing policy (e.g., 238 stronger science in impact assessment). In others, impact may be highly focused, often site, organism, 239 or sector specific, such as recovering an at-risk species in Banff National Park or identifying regulatory 240 thresholds for a novel substance arising from the electroplating industry. To the extent that improving 241 environmental outcomes depends on changing human behaviours, success is likely to take far longer 242 ( A clear theme throughout our discussions was that success is defined differently according to scale 250 (temporal, spatial, institutional) and context. As such, success is presumably viewed and prioritized 251 differently (or even conflictingly) by various actors. For example, an academic may define success as 252 graduating students or publishing papers in high impact journals, while a decision-maker may view 253 success as a new tool, a completed decision, or reduced conflict. Nonetheless, as noted above, the 254 focus is often on the degree to which an ultimate goal (or goals) is addressed while, in reality, there are 255 more proximate successes that may collectively contribute to ultimate successes over longer time 256 scales. For those reasons, a singular definition of success is challenging to identify but for the purpose of 257 this paper we suggest that success in applied environmental science is respectfully-conducted, partner-258 relevant research that is accessible, understandable, and shared, with the potential to contribute to 259 change. Change could be in the context of improved decision-making or changes in behaviour or 260 attitudes but could also reinforce the status quo (i.e., continuation of good practices). We acknowledge 261 that others have attempted to define success. For example, Lubchenco (1998) considered success to be 262 when knowledge generators provided the "best possible science that is useful". In providing such a 263 definition we also recognize that failure and incomplete successes both have immense value (see Box 2). 264 Creating the potential to contribute to change rather than change itself, not unlike how Palmer defines 265 "actionable" science (Palmer 2012), is a necessary distinction between scientists' knowledge creation 266 and dissemination for policy purposes and policy-makers' work on policy creation and evolution. 267 Indeed, there are many socio-political and economic reasons why change may not occur that are 268 entirely beyond the sphere of influence of a researcher. We also recognize that it is possible to conduct 269 successful environmental research independent of partners, but doing so omits co-production in 270 knowledge creation and uptake (e.g., Matson et al. 2016). It is also possible that the partner on a given 271 project may not use research findings directly, but that research may still have broader impacts within 272 the environmental community. It goes without saying that the research also needs to be unbiased and 273 high-quality such that it contributes meaningfully to the evidence base (Roche et al. 2019). 274

275
What are the ingredients for success (best practices) in applied environmental research?

276
Here we provide a series of strategies that collectively create a recipe for designing and delivering 277 applied environmental research that is more likely to be successful. At the core of this recipe are actions 278 that foster a strong and respectful partnership throughout the entire research process, supported by 279 actions that can be embraced at one or more stages in the process ( Figure 2). These strategies are 280 intentionally short and punchy in the hopes that they resonate with and are retained by readers. To 281 facilitate uptake of these strategies a video summary was also created and can be viewed at 282 https://youtu.be/JdMVueunJfo. We acknowledge that it was impossible to cite all relevant literature; we 283 encourage readers to use the cited references as a starting point to find other materials rather than 284 assuming that these are the only or definitive references for a given topic. 285

Build a network and fill the gaps 286
A collegial team with mutually defined goals that are agreed-upon in advance is key to research that is 287 relevant to all (partners and researchers) and has the potential to create change (Caviglia-Harris et al. 288 Submitted). It follows that researchers who want to engage in co-designed work must invest time and 9 289 effort into networking and building and maintaining this team (Ansell and Gash 2008). A great place to 290 start is building a 'network map' or visual image of all known or potentially interested or affected 291 parties, including rights-and stake-holders, researchers, and decision-makers. This is useful not only to 292 see the connections among players, but also to identify important gaps that should be filled. It is also 293 useful to understand the variety of roles played by people and groups in the network, including the roles 294 of science and scientists, rights-and stakeholders, partners, industry, and different levels of 295 government. Researchers may want to rely on those more established in their study area to help 296 identify gaps in the network and make introductions when necessary. Here, boundary organizations can 297 be especially helpful (Safford et al. 2017). Boundary organizations are those groups that can effectively 298 help bridge divides between groups with different norms and goals; for example, a non-profit with a 299 long history of working with academics might understand both typical academic goals (e.g., publications, 300 student training) and typical non-profit goals (e.g., mission-driven change). It is worth noting here that 301 there are broader networks and social spheres in which applied environmental research is embedded 302 increasingly fueled by connectivity of humans around the globe. Although this can be a force for good, it 303 can also lead to misinformation and disinformation activities (e.g., climate change denial; Dunlap and 304 McCright 2010) and the notion of living in a post-truth world (Gross 2017). The scientific community will 305 have to become more savvy in using various networks to spread evidence based knowledge and stand 306 up for environmental science by demonstrating its value and relevance (see Lubchenco 2017). 307 308 309 310 Maintain frequent and respectful two-way communication with partners and stakeholders 311 Success in environmental research demands on-going communication with partners and stakeholders 312 from before the project is designed through to after it is completed, which is inherent in co-production 313 ( reported that ongoing engagement and co-production yielded more actionable science than science 316 done independent of partners or with more limited communication. Related to the need for frequent 317 communication is the need for such communication to be respectful. It is not uncommon for 318 researchers to reach out within days of a grant deadline to seek a letter of support -and then the letter 319 writer may not ever hear back from the researcher after such a letter is supplied. Similarly, some 320 researchers have adopted a "parachute model" where they drop in to do research in a given region or 321 community for a short period of time, and then take off and are never heard from again. Such practices 322 are detrimental to the development of meaningful partnerships and respect (Chapman et al. 2015). 323 Respect also means understanding cultural differences among individuals, organizations, and regions. 324 Active listening is regarded as an effective strategy to understand the needs of partners and end-users 325 to ultimately achieve success in collaborative environmental research (Toomey et al. 2017). Moreover, 326 early and frequent communication that is participatory (rather than uni-directional) has been shown to 327 improve project outcomes (Evely et al. 2011). 328 329 Don't rush relationships 330 There is a growing momentum to co-create projects amongst academic and non-academic partners. This 331 can increase the likelihood of success for applied environment research by: 1) ensuring that the 332 questions and experimental design of the project are relevant for the real world; 2) providing learning 333 opportunities for all members of the project to share knowledge and perspectives on a wide range of 334 issues; and 3) potentially enhancing the chance that recommendations based on project outcomes are 335 adopted. decision-makers to help them interpret scientific claims (e.g., bias is rife; no measurement is exact; 393 extrapolating beyond the data is risky). We suggest that this list should be mandatory reading for 394 scientists; anything that can be done to minimize those issues in the research process or communication 395 of scientific evidence will contribute to project success. Uncertainty and limitations should be 396 considered from project inception through to application of findings. Similarly, it is important to 397 recognize that there are often disconnects between perceived risk and actual risk which can lead to poor 398 policy choices (Gilbert 2011). Transparency is a key concept that can be incorporated during all phases 399 of environmental research as a mechanism for clarifying and overcoming aspects of uncertainty -both 400 with current research but also in contextualizing that work relative to existing and future evidence 401 (Ellison 2010). 402 403 Consider the sphere of influence 404 The overall goal that applied environmental scientists should aim for is to answer research questions 405 that are not only interesting and important to science, but also have the potential to help solve 406 environmental problems. Boundary spanning, defined as 'work to enable exchange between the 407 production and use of knowledge to support evidence-informed decision-making in a specific context' 408 (Posner and Cvitanovic 2019), is essential to the success of solutions-oriented environmental research. 409 Intrinsic to boundary spanning is assessment of the societal context of the environmental issue being 410 addressed and consideration of its 'sphere of influence', particularly during project planning stages. 411 412 Before initiating a research project, we recommend environmental scientists first identify a timely 413 environmental issue of interest and begin to understand it through multiple lenses (e.g., economic, 414 social, political) and at different scales (e.g., time, space) to appreciate the most pressing and timely 415 science needs. This is best achieved through conversations with people who care about the 416 environmental issue, often outside of academia -such as landowners, elders, government scientists, 417 resource managers, and stewardship groups -and by collaborations between social and biophysical 418 scientists. Integrating social science into research planning helps put the environmental problem in a 419 societal context and better understand stakeholder perspectives (Maxwell et al. 2019). It is vital to 420 consider the political jurisdiction relevant to the environmental issue(s) at hand -specifically, whether it 421 falls within the purview of municipal, regional, federal, or Indigenous authorities -and the applicable 422 statutes, laws, and regulations. 423 424 We also recommend creating a conceptual diagram of the "sphere of influence" of a proposed project 425 including likely 'influencers'. The sphere of influence of a project describes, hypothetically, the various 426 pathways that research outcomes could lead to valued impacts (e.g., change in environmental policy, 427 creation of a protected area, or enhanced protection of a threatened species As in other scientific endeavours, students often play a crucial role in applied environmental research. 449 Training students so that they are equipped with the skills required to build and maintain partnerships 450 can yield current and future benefits for all involved. Embedding students in partnerships has clear 451 benefits for the students themselves. They are exposed to varied training environments and 452 perspectives, which can make them better prepared for non-academic environmental careers (Cid and  453 Brunson 2020) and broaden their experience and professional networks. In academic environments, 454 they learn to value creativity and originality while also pursuing general understanding of natural 455 phenomena. Since partners may ascribe higher value to pragmatism and locally relevant information, 456 the tension between researchers' pursuit of the general and stakeholders' desire of the specific is a 457 frequently cited conflict in applied environmental research partnerships (e.g., Podesta et al. 2013). 458 Navigating these and other conflicts to see a project through to application is an incredibly valuable 459 experience for young scientists. 460 461 The onus should be on the supervisor to be the role model (Filstrup 2019) so that students learn to be 462 good citizens in partnerships, with skills that include collaboration, critical thought, creativity, patience, 463 respect and effective communication. Yet, academics should avoid sharing their responsibility for 464 students with partners without having discussions about it. For some it may create an unwelcome 465 burden but other partners may want to embrace such responsibility as long as it is not an abdication of 466 responsibility by the academic mentor. Engagement early and often ensures that students are prepared 467 with the skills and habits needed to work with the partner organization. It further strengthens student 468 access to supportive relationships and resources, and it helps identify shared skills for capacity building. 469 Because students often need to meet certain institutional deadlines towards their degree, clear 470 expectations and time commitments should be mutually discussed and revisited as needed. When this 471 training is done successfully, the student can bring demonstrable benefits to the partnership. Students 472 can bring a level of focused dedication to a project that is difficult for later-stage professionals to 473 sustain, and their inclusion elevates scientific productivity (e.g., Kyvik and Smeby 1994). The potential 474 for co-learning increases as students and partners participate in joint field and/or lab experiences. 475 Moreover, these students will become the leaders of tomorrow's partnerships, having learned the 476 challenges and rewards, and applying these skills to solve so-called "wicked problems" through 477 collaborative and interdisciplinary approaches. 478 Be flexible and responsive to partner needs 480 To foster successful partnerships, we recommend that applied environmental research have a degree of 481 adaptive capacity embedded in all stages. There is no single path or process that works for every 482 partnership and pathways to mutually-defined success can be mapped out at early stages of the project 483 (see Define questions and consider pathways together). That said, all parties will need to remain nimble 484 and responsive over the course of the project as situations change. It is important for researchers to 485 recognize that there may be areas where partners can be flexible (e.g., academic partners may be able 486 to take on new tasks, government partners may be able to make resources available for emerging 487 problems) and areas where there is little to no flexibility (e.g., government partners with fiscal 488 deadlines, academic deadlines for students (highly qualified personnel), non-profit organizational 489 mission with partners, reading annual reports, and reviewing the user's context (e.g., organizational mission; 529 past, present, and future projects; community), will not only inform research design but help both 530 parties to find common language and may increase the relevance and applicability of the research itself. 531 Creating a knowledge mobilization plan that includes participation of knowledge users is essential but 532 needs to be underpinned by the philosophy that emphasizes empowerment, equity, trust and learning. 533 The iterative exchanges should be considered in the knowledge mobilization plan as well as time to 534 understand user needs.

536
Acknowledge partners 537 Token partner engagement abounds. We recommend wrapping up projects by 'closing the loop' with 538 partners (for long-term projects, occasional reviews during the research process may also be useful). 539 Such reflection reminds one to consider the context of our research and to ask for feedback about 540 whether we have been effective partners. It goes beyond the project's exit strategy, which outlines how 541 it ends or is transitioned (e.g., to a future project), and reminds us that the collaboration must meet the 542 needs of our research partners. This challenges us to think outside of what is valued in the academic 543 context, to ask what would benefit our partners and then not only to build that into the project plan but 544 into what we do to maintain the relationship. Closing the loop also means thinking about power and 545 recognizing the privilege inherent in the access to funding, research support and compensation for 546 research activities that is the norm for academic researchers (Higginson 2018; Wallerstein 2019).

547
Awareness of power relationships may lead to recognition rather than reward for the contributions of 548 research partners. Rewarding carries with it an implicit imbalance of power between the partners 549 receiving and distributing rewards. Both academic and non-academic partners reap the rewards of a 550 well-designed collaboration, but this does not necessarily mean the rewards flow from one to the other. 551 The nature of the recognition and of the support that academic partners can provide is specific to the 552 research partner and project. Examples include: advocate for non-academic partners to funding bodies 553 with respect to valuing in-kind contributions of data and expertise; include funding in grant applications 554 to cover the time contributed to the project by non-academic partners, especially not-for-profit and 555 community groups; be aware of paywalls and provide access to library resources; find opportunities to 556 use university resources to showcase stories featuring the work and contributions of research partners 557 (e.g., quotes and contact information in university press releases about research partnership, feature 558 partners alongside researchers in university-created videos about the project; invite partners to give 559 solo or joint seminars about the research and its applications). The importance of returning to a 560 community to share results (in-person or by video conference) and to hear feedback cannot be 561 understated; it can help to interpret and validate results, adapt analysis or communication as necessary, 562 and increase uptake and implementation.

563
In some instances, acknowledging involvement in the research may include considering co-authorship 564 on peer-reviewed journal articles. This ought to be discussed early in the research process, particularly 565 for partners who are unfamiliar with academic publishing or for interdisciplinary work where publishing 566 norms may differ (Cooke et al. 2020). The extent to which a given partner values -and is able to engage 567 in -co-authoring publications will depend on the interests, resources, organizational culture, and 568 contributions of research partners. For example, partners may not be compensated for time spent to 569 publish or have publications factor into career advancement; some groups might have limited internet 570 access, familiarity with or access to specialized computer programs, or seasonal activities that make it 571 difficult to turn around manuscripts. Where individual authorship is not appropriate (e.g., community 572 science, degree of involvement), authorship that recognizes the contributions of a group can provide an 573 alternative (Ward-Fear 2019).

574
Change the narrative 575 Successful uptake of environmental research is about more than getting information to the desk or 576 email inbox of a decision-maker -evidence must be framed in clear, persuasive, and, in some instances, 577 politically salient ways (Rose 2015;Rose et al. 2017). This means that successful applied research does 578 not end at peer-reviewed publications or even policy briefs, which few busy decision-makers or 579 practitioners have time to read and may not be successful instruments of change. Effectively 580 communicating the weight of evidence and implications of research findings is a process that involves 581 understanding the audience that will use the evidence -where listening and connecting are just as 582 important as talking (Smith et al. 2013). Successful communication that engages with decision-makers is 583 one of the most potent catalysts for action (Baron 2010; see 'Consider the sphere of influence' above). 584 Written communications have been shown consistently to be less effective or persuasive than face-to-585 face meetings (Roghanizad and Bohns 2017), and in many cases, telephone conversations may be more 586 productive than email communication. Moreover, do not assume that partners have reliable internet 587 access or that an unanswered email is a lack of interest. Do not be afraid to pick up the phone, which 588 can often be more effective in forging meaningful connections than email. Never trade-off scientific rigour 608 Sound environment-related decisions need sound environment-related information. One of the most 609 powerful aspects of applied research is that it can increase our understanding and lead to new, reliable 610 knowledge. Poorly designed and/or executed applied research runs the risk of providing incomplete or 611 incorrect information that could lead to ineffective or even harmful decisions (Sells et al. 2018). As 612 noted by Hofself (2018), "If science isn't rigorous, it's reckless"; although this message was coming from 613 the field of human health, it is just as relevant to environment-related management. Providing practical 614 and applied information that has relevance -ideally beyond the immediate question or problem -615 requires a genuine collaboration among researchers and partners (see Define questions and consider 616 pathways together, above); however, scientific rigor should never be compromised in the process. For 617 some collaborative environmental projects there can be a sense of urgency to deliver actionable science 618 to end users, and in this process, there may be temptation or pressure to disregard proper scientific 619 rigour. Indeed, some decisions, or the implications of those decisions, may require less detail or 620 specificity than others. However, researchers must ensure that the integrity and credibility of research 621 findings and interpretations are not sacrificed. The underlying goal of researchers should be to produce 622 rigorous, unbiased, and reproducible science that is conducted in a way that is ethically minded, 623 regardless of speed (Roche et al. 2019) and/or pressure from partners, and to be willing to maintain 624 scientific independence (i.e., walk away from the partnership) if/when asked to do anything less. The 625 researcher will need to navigate the need to get the science done while working diligently to ensure that 626 stakeholder and partner engagement occurs throughout the process. 627 628 Protect individual research integrity 629 Partnership and co-production can create dynamics and tensions that may be unfamiliar to researchers 630 trained in traditional scientific norms and methods. Partners may have particular outcomes or 631 applications in mind, including socio-political aims which may or may not be shared by researchers. 632 Partners may also have ideas about what types of data and findings are most useful to them (Young et  633 al. 2016a), and may exert pressure on researchers (intentionally or not) to focus their efforts on 634 research questions and data collection of potential high utility. Some researchers are comfortable 635 accommodating such interests, while others are concerned with maintaining distance from policy or 636 political considerations (e.g., Lackey 2016; Donner 2017). In all cases, however, it is important for 637 researchers to reflect on such possibilities prior to engaging in partnerships and co-production, and to 638 take measures to anticipate and clarify how research will be conducted and results communicated. For 639 instance, it is advisable to develop explicit agreements with partners about access to raw data and 640 metadata, analysis of findings, and communication of results (noting, however, the imbalance in 641 familiarity with what these things mean and the implications). Such agreements may include 642 commitments to publicize information that is useful to the broader scientific community, even if they 643 run contrary to partner priorities or expectations. Information about methods, study limitations, and 644 null findings (if applicable) are important for our global understanding of phenomena and should be 645 communicated transparently. More generally, these issues relate to questions of research integrity and 646 credibility. Partnership and co-production imply that all parties contribute to the research process for 647 mutual gain. Prior reflection and agreement on questions of integrity can help structure these 648 collaborations to ensure that research is both useful to partners and credible in the eyes of the broader 649 scientific community. 650 651 Balance the short and long game 652 Most environmental problems are sufficiently complex that they are best approached by tackling 653 focused questions that can be addressed in a short time-frame (months to a few years) while 654 simultaneously collecting data that will feed into larger, often longer-term studies (decades or more). 655 This means having outcomes that are project specific and achievable, yet contain the vision and 656 forethought for the long term, recognizing that sustained funding for long-term research is challenging 657 to secure (Parr et  What are specific indicators of success?

665
Every project is unique, so each project requires unique criteria to assess success. To that end, scientists 666 should work with their partners at the very start of a project to co-develop relevant metrics/indicators 667 to gauge success. Importantly, if this is done from the beginning (i.e., during application phase) then 668 efforts to track success can be incorporated all along the project and not be relegated to a disconnected 669 post-hoc activity. Such an approach is intuitively more effective than simply providing researchers (and 670 partners) with a generic survey after-the-fact into which they have to try and fit their successes. 671 Moreover, although we focus here on the idea of success, learning from failure is also a form of success; 672 when strategic and timely, sharing hard-won lessons can be an act of generosity to both the research 673 and non-research communities. Accordingly, researchers and partners should be encouraged to reflect 674 and comment on what worked, what did not, and why. Ideally these reflections would be shared with 675 the broader communities if they enabled others to avoid the same pitfalls. We stress that funders tend 676 to see these exercises as part of the learning process; put another way, the willingness to "fail forward" 677 should not be used to overtly or inadvertently punish people who take time to reflect on and to share 678 lessons learned. learning and sharing opportunities than simply having the researchers and partners reflect on this in a 690 written final report. We also note that the topic of research evaluation (whether specific to applied 691 research or more broadly) remains an area where there is much ongoing discussion (Penfield et al. 692 2014). Moving beyond counting papers and using peer-reviewed journal Impact Factor to assess paper 693 quality to better gauge influence of research remains a fundamental challenge (Donaldson and Cooke 694 2014). To that end, some of the ideas raised here are inherently subjective. In the future it is hoped 695 that more robust and reliable indicators will be available to gauge success in applied environmental 696 research. Although altmetrics can assess reach across media platforms (Erdt et al. 2016), it is a limited 697 reflection of whether and how research results had meaningful impact and is subject to manipulation 698 (Bornmann 2014). How to assess societal impact (broadly) in a quantitative manner remains difficult 699 (Bornmann 2012(Bornmann , 2013

703
Given the number of researchers that self-identify as environmental scientists, applied ecologists, 704 conservation scientists, sustainability scientists and so on, one would expect that policy-makers and 705 decision-makers would be drowning in the knowledge needed to make good evidence-based 706 environmental decisions. Yet, despite being called for more than a decade ago (i.e., Sutherland et al. 707 2004), evidence-based management still faces many challenges. The reasons why this idea has not been 708 fully realized into action are many and complex (Cook et al. 2010), yet what is apparent is that there is 709 much that the scientist or researcher can do to increase the likelihood that their work will be used and 710 create the potential for change (i.e., a core aspect of our first objective of identifying what we mean by 711 success). The recipe for success that we share here is driven both by peer reviewed literature and the 712 methods or strategies that our team members, or their partners, use to achieve success. A number of 713 key themes emerged, notably: acknowledging limitations, the need for extensive partner engagement 714 (ideally in a co-production framework), sharing outputs via diverse channels, and ensuring that there are 715 opportunities to train early career researchers in applied partnership science. What is clear from our 716 discussions and writing is that there is not a single path to success nor a singular action that will ensure 717 success.

718
The best advice we can provide to those embarking on applied environmental research is to embrace 719 the strategies that we outline here and to continuously reflect on progress toward shared research goals 720 and to remain open to adjusting course where necessary. Failures happen and present researchers with 721 opportunities to learn and share lessons. We emphasize the importance of fundamental science and the 722 need to balance it with applied, mission-oriented research. Although this paper is focused on how to 723 deliver applied research with impact, we cannot predict what the future will hold nor how fundamental 724 research will inform policy and practice. Governments and other funding bodies fund applied 725 environmental research to inform policy and practice, so we suggest that it is incumbent on researchers 726 to adopt strategies that make such outcomes as effective as possible. The strategies we outline here 727 (also see companion video; https://www.youtube.com/watch?v=JdMVueunJfo&t=0s) are intended to 728 support that effort and to improve transparency so that partners and funders are better able to assess 729 the success of the work and where new efforts might then be needed.

Indicator Timeline Responsible party
Relative ease Quality and quantity of scientific outputs Is the science of sufficient rigour that it could be defended in legal proceedings or used in evidence synthesis (e.g., meta-analysis and/or systematic review)?

During, Conclusion
Researchers, Broader scientific community Easy Does the work ascribe to best scientific practices (e.g., disciplinary norms such as use of blinding) with respect to methodological rigour and reporting (e.g., sufficient detail that it could be replicated)? For personal use only. This Just-IN manuscript is the accepted manuscript prior to copy editing and page composition. It may differ from the final official version of record.
How many presentations were made to nonscientific audiences (and number of attendees)?

Conclusion Researchers Easy
Were any alternative forms of engagement delivered to non-scientific audiences (e.g., story boards, training to conduct ongoing monitoring)?

Conclusion Researchers Easy
Were there changes in literacy, numeracy, or human behaviour? For personal use only. This Just-IN manuscript is the accepted manuscript prior to copy editing and page composition. It may differ from the final official version of record.