Introduction
The economic growth and prosperity of the second half of the twentieth century depleted the world’s natural resources. A powerful scientific response ensued in an effort to stem the decline of biodiversity (Soulé Reference Soulé1985, Sarkar Reference Sarkar1998). A proliferation of programmes and policies emerged, which culminated in the negotiation of the Convention on Biological Diversity (CBD) in Rio de Janeiro in 1992. Yet, biodiversity continues to decline, especially in the rapidly changing landscapes of tropical developing countries. Conservation biology, as a discipline, emerged with the explicit aim of tackling that decline. However, by focusing on biology and ecology, it has tended to overlook the complex social, economic and political context within which conservation actions take place. Pressey et al. (Reference Pressey, Weeks and Gurney2017) identified this problem, accusing conservation biologists of a collective form of displacement activity, advocating irrelevant responses to what on the face of it seem like the incompatible goals of protecting biodiversity and supporting human development. More attention to the human behaviours and attitudes behind drivers of environmental decline might lead to a broader, more inclusive conservation science (Byerly et al. Reference Byerly, Balmford, Ferraro, Hammond Wagner, Palchak and Polasky2018).
Successful conservation needs to influence the behaviour of the people who live in, and depend upon, the resources of the landscapes and seascapes where conservation actions take place (McNie Reference McNie2007, Knight et al. Reference Knight, Cowling, Rouget, Balmford, Lombard and Campbell2008). Conservation efforts have frequently sought to restrict income-generating activities, portraying them as negative and harmful (Kamoto et al. Reference Kamoto, Clarkson, Dorward and Shepherd2013). When conservationists seek to lock up nature, conservation has too often been seen by local people as locking up vital assets in ways that obstruct their entry into the cash economy, hinder their access to education and health services and disrupt the provision of roads and other infrastructure (Feintrenie et al. Reference Feintrenie, Chong and Levang2010, Colfer Reference Colfer2011, Levang et al. Reference Levang, Riva, Orth, Cramb and McCarthy2016, Sayer & Margules Reference Sayer and Margules2017, Boedhihartono et al. Reference Boedhihartono, Bongers, Boot, van Dijk, Jeans and van Kuijk2018). Conservation has struggled to link biological fundamentals with the messy reality of human behaviour, economic growth and politics (Boedhihartono et al. Reference Boedhihartono, Bongers, Boot, van Dijk, Jeans and van Kuijk2018). Land-use conflicts, civil and military disorder, donor fatigue and narrow sectoral approaches to governance – what Indonesians colourfully call ‘egosectoral’ approaches – all conspire to thwart conservation actions. Here, we reflect upon broader conservation science and endorse a set of practices that embrace multiple disciplines to explicitly address the complex human activity that conservation has become.
Arguments for integrating conservation and development were articulated in the World Conservation Strategy 40 years ago (IUCN & WWF 1980), but the logic of sustainable development has still not been mainstreamed into the conservation research agenda. A recent response has been the development of an updated framework for conservation science (Kareiva & Marvier Reference Kareiva and Marvier2012) that acknowledges the role people must play if the protection of biodiversity is to succeed. The publication of that framework generated unnecessary controversy between its proponents and many traditional conservation biologists (Kloor Reference Kloor2015). Williams et al. (Reference Williams, Balmford and Wilcove2020) have documented the persistent failure of much conservation science to deliver conservation outcomes. Recognition by practitioners of the need for people to own and drive conservation is not new. Integrated conservation and development projects (ICDPs) have been widely implemented since the late 1960s (Sayer & Campbell Reference Sayer and Campbell2004). ICDPs have generally not succeeded for a range of reasons, notably the assumption that global environmental values had primacy over local development needs (Sayer & Wells Reference Sayer, Wells and To McShane2004, Sunderland et al. Reference Sunderland, Ehringhaus and Campbell2007). ICDPs were often just attempts by aid agencies and international non-governmental organizations to reduce the impact of conservation interventions on local people living in poverty. Many conservation interventions from this period suffered from a failure to recognize that the geographical locations in which they were taking place were governed in complex, polycentric, multi-sectoral ways, and that changes occurring outside of the project area would determine outcomes (e.g., climate change, fluctuating global markets, etc.) (Sayer et al. Reference Sayer, Endamana, Boedhihartono, Ruiz Pérez and Breuer2016a).
Conservation actions take place on the ground, in geographically defined areas consisting of landscapes and seascapes. Landscapes are linked social-ecological systems (Walker & Salt Reference Walker and Salt2006). They are shaped by interactions between biological, ecological and physical processes mediated by the interventions of people (Opdam et al. Reference Opdam, Luque, Nassauer, Verburg and Wu2018). Scientists and conservation practitioners often seek transformational change in landscapes. They want problems solved – preferably within the short time frames of donor funding. Our experience is that change is usually gradual and incremental, but is punctuated by transformational events often triggered by major external drivers, such as infrastructure and other major investments. Understanding and influencing change requires long-term commitments in situ, drawing on and linking different sources of knowledge from different disciplines.
Conservation problems are societal constructions; they cannot be separated from broader aspects of well-being, relational values, politics and global change. Conservation science would benefit from gathering evidence on how these complex systems work and identifying processes that might influence change. Evidence in this sense is not only social and environmental data, but the knowledge of multiple actors all with different ‘ways of knowing’, political interests and sources of information (Evans et al. Reference Evans, Davila, Toomey and Wyborn2017). Conflicts are inevitable; epistemological differences between disciplines often result in incompatible methods and approaches, resulting in cultural and practical barriers (Hicks et al. Reference Hicks, Fitzsimmons and Polunin2010, Filer Reference Filer2011). For example, social-ecological systems approaches have been seen to offer a neutral template for contextualized problem-solving in sustainability science, but they are criticized for not addressing more socially interpretive aspects of dynamically coupled human–natural systems (Agrawal Reference Agrawal2001, Kates et al. Reference Kates, Clark, Corell, Hall, Jaeger and Lowe2001, Ostrom Reference Ostrom2009). Science deployed in landscapes to solve conservation problems can draw from the expansive interpretations of systems approaches to ensure that the generation of knowledge is meeting the diversity of values and users. Stronger emphasis could be placed on understanding ‘predictably irrational’ human behaviour and on how to influence the decisions of those who shape the landscape (Thaler & Sunstein Reference Thaler and Sunstein2003, Osbaldiston & Schott Reference Osbaldiston and Schott2012, Veríssimo Reference Veríssimo2013). Many conservation scientists have argued for this agenda (Brown Reference Brown2002, Ghazoul Reference Ghazoul2007, Lynam et al. Reference Lynam, De Jong, Sheil, Kusumanto and Evans2007, Milner-Gulland Reference Milner-Gulland2012, Bennett et al. Reference Bennett, Roth, Klain, Chan, Christie and Clark2017), but few have been able to engage for the long term in carrying out these activities in situ, with the people whose decisions conservation scientists wish to influence.
We report here on lessons learned for achieving conservation outcomes from our engagement in long-term landscape-scale projects, which have tried to balance local aspirations with global environmental values. We draw on attempts by the Center for International Forestry Research (CIFOR), the International Union for Conservation of Nature (IUCN), World Wildlife Fund (WWF) and Conservation International (CI) to establish ‘learning’ or ‘sentinel’ landscapes (IUCN 2008). We have monitored changes in these landscapes over periods of 5–25 years (Sayer et al. Reference Sayer, Endamana, Boedhihartono, Ruiz Pérez and Breuer2016a), and have developed approaches for assessing the performance of landscapes in delivering conservation and development benefits (Endamana et al. Reference Endamana, Boedhihartono, Bokoto, Defo, Eyebe and Ndikumagenge2010, Sayer et al. Reference Sayer, Margules, Boedhihartono, Sunderland, Langston and Reed2016b). We have not conducted controlled experiments or randomized controlled trials at landscape scale and doubt the feasibility of doing so. We positioned ourselves as ‘peripheral agents’ in these landscapes, seeking to understand change and influence – nudge – the decision-making process to achieve continuous learning and adaptation (Thaler & Sunstein Reference Thaler and Sunstein2009). Our intention was to practice and learn from the different ways science can be organized to solve conservation problems. We document attempts to balance conservation and development in tropical landscapes where iconic biodiversity overlaps with urgent issues of human poverty (Campbell et al. Reference Campbell, Sayer and Walker2010, Langston et al. Reference Langston, Riggs, Boedhihartono, Kastanya and Sayer2020).
We drew from a broad range of methods (Lynam et al. Reference Lynam, De Jong, Sheil, Kusumanto and Evans2007), but five practices were particularly effective in gathering evidence and improving decision-making. These are: understanding the social-ecological context; theories of change; network analysis; scenario development and simulation modelling; and multi-criteria analysis (MCA). Our experience with projects applying these practices shares many of the features of what has been described as transdisciplinary research and sustainability science (Clark Reference Clark2007) – what we call ‘embedded science’ (Langston et al. Reference Langston, Riggs, Kastanya, Sayer, Margules and Boedhihartono2019). Conservation actions have to align with and work through governance arrangements in the places where conservation is needed. The practices identified in this paper were agreed upon by the authors because of their useful attributes for situating conservation actions within a broader process of transdisciplinary science and for meeting the principles of knowledge co-production practiced in sustainability science (Norström et al. Reference Norström, Cvitanovic, Löf, West, Wyborn and Balvanera2020).
Background to the review
Information on the landscapes that inspired this review is given in Table 1. We were never in control of conservation in these landscapes and we did not have the resources to apply all five practices in all of the landscapes. In each of the landscapes, the deployment of a subset of these practices led to better shared understanding of change processes amongst scientists from different disciplines and between the scientists and resource users.
Table 1. Long-term landscape conservation initiatives that inspired the ideas expressed in this paper.
NGO = non-governmental organization.
We review recent empirical and conceptual advances in each of the five practices, focusing on the contribution of each practice to a broader process of collaboration and change. It is not our intention to provide full methodological details of these practices, but rather to demonstrate how conservation science can use them to engage more effectively with the broader social context in which conservation actions take place. We sought to find practical applications for the theoretical and philosophical tenets of transdisciplinary science and to adhere to the principles of knowledge co-production in sustainability research. As researchers we sought to experience and learn from the consequences of our actions in pursuing conservation goals (Reyers et al. Reference Reyers, Roux, Cowling, Ginsburg, Nel and Farrell2010, Lang et al. Reference Lang, Wiek, Bergmann, Stauffacher, Martens and Moll2012, Taleb Reference Taleb2018).
Social-ecological context and inductive research
Understanding the social-ecological context within which conservation actions take place can be achieved through inductive research (Goddard & Melville Reference Goddard and Melville2004), which involves researchers engaging with end users in a highly flexible manner to shape hypotheses about how decision-making might be influenced. Inductive methods encourage what has been called ‘appreciative enquiry’ (Cooperrider & Srivastva Reference Cooperrider and Srivastva1987), in which scientists engage with and learn from other people and organizations active in the landscape seeking to solve problems. Transdisciplinary problem-framing provides a starting point whereby science targets the existing strengths and capabilities within the system instead of imposing externally designed solutions (Brondizio Reference Brondizio2017). For example, deciding what natural components that make up biodiversity merit conservation, and to what extent, reflects contextual cultural values (Boedhihartono et al. Reference Boedhihartono, Gunarso, Levang and Sayer2007, Sarkar Reference Sarkar2019).
Inductive research draws from a diversity of anthropological perspectives that have contributed to biodiversity conservation, including cultural ecology, ethnobiology, political economy and interpretative anthropology (Orlove & Brush Reference Orlove and Brush1996). Anthropological techniques can be used to explore local knowledge and to ensure collective understandings and embedded norms and values amongst resource users and governance agencies are not underestimated (Boedhihartono Reference Boedhihartono2012, Bernard Reference Bernard2017). More pertinently, these techniques can help to determine how knowledge is produced and who is empowered to produce it, how knowledge is shared and the impacts of power asymmetries (Brosius Reference Brosius2006). Historical trends analysis, debates concerning strengths, weaknesses, opportunities and threats (SWOT) confronting a landscape and a broad range of other participatory activities can help to engage local stakeholders in knowledge-sharing and scenario exploration. Drawing rich pictures can elicit information on the values that people attach to landscape attributes (Bell & Morse Reference Bell and Morse2013).
All of the above issues determine the context in which conservation research is conducted and the ways in which resource-use decisions are made. Unstructured, inductive research methods are often more effective at unravelling the complexity of these issues than rigid, pre-planned conceptual and analytical frameworks. Social scientists have led our engagement in each of the landscapes in Table 1. The task of the initial social science engagement has been to build relationships by living in the communities and sharing their lives (Boedhihartono Reference Boedhihartono2004).
Theories of change
Theories of change consist of models that show how initiatives (e.g., a policy, strategy or project) might contribute through a sequence of outcomes to an intended goal (Serrat Reference Serrat2017). Such theories help us to navigate through the complexity of social change. Theories of change are conceptually different from the linear logical frameworks that have guided conservation projects in the past (Prinsen & Nijhof Reference Prinsen and Nijhof2015). Rather than accounting for inputs and outputs, they are designed to account for broader systems processes along non-linear pathways that enable continual learning to achieve accepted goals. Theories of change are transparent and account for trade-offs and make assumptions explicit, thus enabling buy-in from a wide range of stakeholders (Biggs et al. Reference Biggs, Cooney, Roe, Dublin, Allan, Challender and Skinner2016). There is no single definition of, or fixed methodology for, a theory of change; rather, the approach allows flexibility according to the needs of the implementers (Vogel Reference Vogel2012). The utility of a theory of change is therefore limited by its application, by how rigorous and genuinely collaborative it is and by how it impacts on decisions that determine outcomes. Reinforced by Forsyth (Reference Forsyth2018), theories of change risk perpetuating exclusionary practices if they fail to make assumptions explicit and include a diversity of perspectives.
Theories of change can provide conservation scientists with a method for mainstreaming conservation into the development process. We have built theories of change in landscapes with local partners by collaboratively framing why goals and desired futures are what they are, what the details of those goals are, who enables and hinders the achievement of those goals and how they might be achieved (Fig. 1). Doing this effectively means laying all change logic assumptions bare. Theories of change built with local partners enable conservation scientists to contribute to developing a common understanding of the challenges and building a common commitment to deciding how to allocate resources and how and when to trigger interventions. Theories of change built at the start of interventions or retrospectively can also form the basis of evaluating the effectiveness of interventions, serving as a focus for learning and adaptive management (Belcher et al. Reference Belcher, Davel and Claus2020). In our experience, theories of change are most useful as consensus-building and societal learning tools (Brondizio et al. Reference Brondizio, O’Brien, Bai, Biermann, Steffen and Berkhout2016). As part of an inductive research process, historical trends analysis, scenario visualization, simulation analysis, network analysis and discourse analysis have all proven useful in contributing to a theory of change (Boedhihartono Reference Boedhihartono2012, van Noordwijk et al. Reference van Noordwijk, Minang, Freeman, Mbow, de Leeuw, Minang, van Noordwijk, Freeman, Mbow, de Leeuw and Catacutan2015, Amaruzaman et al. Reference Amaruzaman, Leimona, van Noordwijk and Lusiana2017). Agreeing on a conceptual understanding of an issue and a way forward can be a significant step towards building commitment to conservation partnerships.
Fig. 1. Generic theory of change for social-ecological systems in landscapes, adapted from Sayer et al. (Reference Sayer, Endamana, Boedhihartono, Ruiz Pérez and Breuer2016a). A management coalition drives progress towards improved landscape performance. Arrows show the direction of progress in changing the system. Competing claims provide the justification for the process. Metrics for tracking progress correspond to the critical processes, which are shown as numbered boxes: (1) negotiation and communication of clear goals; (2) a clear and agreed theory of change; (3) a rigorous and equitable process for continuing stakeholder engagement; (4) connection to policy processes and key actors; (5) effectiveness of governance; and (6) transparency.
Theories of change are increasingly being used by agencies working to steer complex systems towards more sustainable trajectories (Thornton et al. Reference Thornton, Schuetz, Förch, Cramer, Abreu, Vermeulen and Campbell2017). The rate of change in social-ecological systems is accelerating to the point where the past is not necessarily a good indicator of the future (Steffen et al. Reference Steffen, Broadgate, Deutsch, Gaffney and Ludwig2015). Theories of change must account for temporal and spatial feedback throughout the entire system. Local knowledge, judgement and quantitative data are all required to build theories of change. In developing a theory of change, key milestones and processes to achieve goals have to be identified. Process and outcome metrics should be identified and interventions adapted based on progress towards those milestones and goals (Sayer et al. Reference Sayer, Margules, Boedhihartono, Sunderland, Langston and Reed2016b).
Network analysis
Conservation policy and conservation management are embedded in networks. Emerging from network science, network analysis describes the relationships between societal actors within networks, including individuals, organizations and government agencies. The analysis can identify who and what has influence, the degree of influence and where the linkages between actors can be improved for better overall management of complex systems (Prell et al. Reference Prell, Hubacek and Reed2009, Gallemore et al. Reference Gallemore, Di Gregorio, Moeliono, Brockhaus and Prasti2015). Network science is prominent in sociology and political sciences, where social network analysis (SNA) and policy network analysis (PNA) have been applied to natural resource management in a variety of circumstances (Marin & Mayntz Reference Marin and Mayntz1991, Bodin & Prell Reference Bodin, Prell, Bodin and Prell2011, Scott & Carrington Reference Scott and Carrington2011). Network analysis is distinct from actor network theory, which emphasizes relationality among entities rather than relationships between individual actors (Latour Reference Latour1996). For the purposes of conservation, network analysis can help us to identify who to work with and where to leverage support for building consensus. An accompanying discourse analysis applied to those networks enables a better understanding of the subjective and political vantage points of influential people and institutions (Carmenta et al. Reference Carmenta, Zabala, Daeli and Phelps2017).
Network analysis can uncover hidden alliances and conflicts. Paired network and discourse analysis can show where knowledge might be co-generated to influence the narratives upon which decisions are made (Berkes Reference Berkes2009, Huitema & Turnhout Reference Huitema and Turnhout2009, Nel et al. Reference Nel, Roux, Driver, Hill, Maherry and Snaddon2016). Because governance of natural resources requires cross-sectoral collaboration, bridging institutions are necessary to coordinate the actions of the multiple actors that exist in landscapes (Berkes Reference Berkes2009, Kowalski & Jenkins Reference Kowalski and Jenkins2015). Network analysis can help identify where those bridging institutions should apply their relationship-building efforts (Clark et al. Reference Clark, Tomich, van Noordwijk, Guston, Catacutan, Dickson and McNie2011, Nel et al. Reference Nel, Roux, Driver, Hill, Maherry and Snaddon2016). Network analyses can be used as a diagnostic tool, but its real strength lies in enabling a living and evolving representation of actors and their relationships so that interventions can be adapted to deal with changing realities. Network analyses should not be undertaken by scientists alone; co-produced network analysis, with local people and practitioners, enables a shared understanding of the networks of influence and knowledge sharing, to better manage conservation and development trade-offs (Langston et al. Reference Langston, Riggs, Kastanya, Sayer, Margules and Boedhihartono2019, Sari et al. Reference Sari, Sayer, Margules and Boedhihartono2019). Various methods can be employed to create networks, including participatory activities (Schiffer & Hauck Reference Schiffer and Hauck2010), interview-based data collection (Hogan et al. Reference Hogan, Carrasco and Wellman2007) and content-based analysis.
We used participatory activities at a locally held workshop to gather information on network dynamics in the landscape at the district level in East Lombok, Indonesia (Riggs et al. Reference Riggs, Langston, Margules, Boedhihartono, Lim and Sari2018). The intent of the workshop was to connect local forestry officials and researchers in order to develop an understanding of the challenges and opportunities of integrated landscape management in the district. The network was first mapped on paper, with participants identifying the actors that affect, or are affected by, forest management decisions, including power, influence and relationships. As the network activity was conducted with the local forest authorities, it created an opportunity to discuss power distribution within the landscape and the diversity of objectives of different actors. The network activity incentivized local actors that participated in the workshop to give more attention to stakeholder needs and relationships and to seek partnership opportunities for improved landscape management.
Scenario development and simulation modelling
We have developed scenarios employing simple visual techniques or simulation modelling using the methods of Sandker et al. (Reference Sandker, Campbell, Ruiz-Perez, Sayer, Cowling, Kassa and Knight2010) and Boedhihartono (Reference Boedhihartono2012). Scenarios based initially upon drawings and later formalized with models were used as a basis for a long-term initiative to monitor the performance of a Congo Basin forest landscape (Endamana et al. Reference Endamana, Boedhihartono, Bokoto, Defo, Eyebe and Ndikumagenge2010). Scenarios for best-case and worst-case future situations were developed in all of the landscapes where we worked. Scenario exercises encouraged debate about conservation decision-making and built consensus on desirable future situations. For example, to visualize two possible future scenarios seen by local people and other stakeholders in La Paloma (Uruguay), we facilitated negotiations in Paloma between local Gaucho communities and outside investors who proposed to construct a pulp mill (Boedhihartono & Sayer Reference Boedhihartono, Sayer, Stanturf, Lamb and Madsen2012). Participants were asked to illustrate their preferred balance between mixed crops, plantations, cattle grazing and forest. These visualizations helped with the exchange of information, as well as helping to reduce tensions and move towards an acceptable compromise at the landscape scale.
Simulation modelling can be an effective way of creating and analysing a formal abstraction of an actual social-ecological system to predict its performance in the real world (Sandker et al. Reference Sandker, Campbell, Ruiz-Perez, Sayer, Cowling, Kassa and Knight2010). Simulation modelling can be used to assess alternative performance outcomes by varying the parameters of the model, its stocks and flows and the weights they are given. It can help users understand whether, under what conditions and in which ways desired or perverse outcomes might be achieved (Collier et al. Reference Collier, Campbell, Sandker, Garnett, Sayer and Boedhihartono2011). For example, in Malinau District (North Kalimantan, Indonesia), government officials, local communities and researchers from CIFOR conducted participatory modelling to simulate potential scenarios for the impact of oil palm on the landscape (Sandker et al. Reference Sandker, Suwarno and Campbell2007). We also used these techniques in East Lombok (Indonesia) to work with local forestry officials in order to explore scenarios for restoring degraded land (Riggs et al. Reference Riggs, Langston, Margules, Boedhihartono, Lim and Sari2018). The purpose of the model was not to make accurate predictions of restoration, but to generate discussion on the long-term impact of different restoration options for the land and for the local community. Models have limited power to make accurate predictions about complex outcomes of land-use change, but in these cases they were used to promote dialogue on different possible trajectories of change.
Multi-criteria analysis
The allocation of areas of land to different uses requires the simultaneous consideration of multiple goals and choices on the trade-offs between those goals (Lai & Hopkins Reference Lai and Hopkins1989, Malczewski Reference Malczewski2006, Moffett & Sarkar Reference Moffett and Sarkar2006, Sarkar & Illoldi-Rangel Reference Sarkar and Illoldi-Rangel2010). Measuring suitability for agriculture is a function of climate, soil type, market access, etc. Identifying areas for the protection of biodiversity often involves incorporating opportunities and constraints such as economic and social benefits and costs, as well as species distribution patterns and the viability of populations.
Incorporating multiple criteria into land-use decisions is challenging. It requires explicit attention to values that may be in conflict and the trade-offs between them. In formal MCAs, each value is modelled by a criterion, and the potential performance of each alternative outcome is assessed against each criterion. These assessments can then be integrated by various MCA methods to provide a final ranking of the outcomes that takes into account the trade-offs between the criteria. We used the MCA methodology of multi-attribute value theory (MAVT) to explore land-use options in an area allocated for estate crops near Merauke in Indonesian Papua (Sarkar et al. Reference Sarkar, Dyer, Margules, Ciarleglio, Kemp and Wong2016). This methodology required the explicit formulation of the problem to be solved, what the available options or alternatives were, what socioeconomic, biological and other factors or criteria needed to be taken into account, how they should be measured and a step-by-step protocol for making choices. The core of this methodology was to elucidate all of the relevant factors and to order them by perceived importance in an objectives hierarchy of values modelled by the criteria (Keeney Reference Keeney1996). We constructed such hierarchies and assigned weights to the relevant criteria through consultations with all stakeholders. The processes of inductive research and the development of theories of change described earlier enriched this preference elicitation stage.
The structured explicit approach of the MCA has four important advantages (Sarkar et al. Reference Sarkar, Dyer, Margules, Ciarleglio, Kemp and Wong2016):
-
(1) The analysis is transparent, with all objectives and methods available for scrutiny.
-
(2) The analysis, repeated by others with the same data, will yield the same results.
-
(3) Because it is explicit, the method of analysis is easily transported from one scenario to another.
-
(4) The structured protocol enables the development of computer-based decision support tools, making the rapid exploration of the costs and benefits of a large number of decision choices feasible.
Typically, a number of alternative solutions will be found based on the different weightings and preferences of stakeholders. For example, one solution might show the cost in foregone production opportunities of achieving a high-biodiversity conservation goal. By contrast, another might show the cost to biodiversity protection of reaching a high-production target. The idea is to find a compromise balancing both production and protection that all stakeholders can live with, even though it may not satisfy their highest aspirations.
A large number of methods have been applied to MCAs (e.g., Figueira et al. Reference Figueira, Greco and Ehrgott2005, Moffett & Sarkar Reference Moffett and Sarkar2006). MAVT (Dyer Reference Dyer2005) was the methodology used for our work in Merauke because, in the absence of explicit consideration of uncertainties in outcomes, it is the simplest form of analysis consistent with standard microeconomic theory (Moffett & Sarkar Reference Moffett and Sarkar2006).
MCA was used to identify areas that might be protected for their biodiversity value within the concession for growing wood for pulp near Merauke (Sarkar et al. Reference Sarkar, Dyer, Margules, Ciarleglio, Kemp and Wong2016). The result also simultaneously incorporates areas that protect hydrological processes and access for local communities to practice traditional uses such as hunting and gathering food and medicines, extracting sago from sago palm groves and maintaining important cultural sites.
A very simple solution and a more complex solution are contrasted here to illustrate how MCA was used (Table 2). In each solution, the following question was asked: ‘What is the set of grid cells to be protected that …?’
Table 2. Two solutions, shown in Fig. 2 as Scenario A and Scenario B, and the criteria used in trading off biodiversity protection with production opportunities in a concession for growing wood pulp near Merauke, Papua Province, Indonesia.
More total area is required to satisfy the goals of the second, more complex solution. Figure 2 illustrates the impact on potential wood pulp production.
Fig. 2. The number of grid cells rated 5, 4 and 3 (the highest ratings, respectively) for growing wood pulp (grey area) and the number required to satisfy the two solutions outlined in the main text. Red is the first, simpler solution and blue the second.
Both solutions require that some of the grid cells rated as most suitable for growing wood pulp be allocated to biodiversity protection, ongoing community use and maintaining hydrological connectivity. What is important here is that all stakeholders can see just what the opportunity costs for production are when a different goal – in this case, biodiversity protection – is satisfied. The analysis provides evidence upon which to base decisions. Stakeholders may well want to see the results of other solutions, such as the impact on biodiversity of higher-production goals, before deciding on one or another, and this can be done quickly with computer-based decision support tools of this kind.
Process integration
The five practices described above can be used interactively and with continual feedback. They complement and do not replace other scientific studies that may be needed to understand specific conservation and development challenges. For example, the first stage in Seram, Lombok and Malinau consisted of understanding landscape context through inductive research. Then, with a sufficiently rich understanding of the decision context, actors and values available, it became possible to formulate and use theories of change that make stage-wise predictions of outcomes leading towards goals. Network analysis complemented these stages by identifying the relevant actors in the system in which conservation decisions were to be implemented and how these actors interacted with each other. Established partnerships created an opportunity to use scenarios and simulation modelling to predict and explore possible outcomes. MCA was used before the other tasks were completed. Once the societal values, constraints and opportunities had been identified, it became possible to identify feasible decisions that could potentially be implemented. Effectively integrating these practices and tools required a multidisciplinary team of people with strong ‘epistemological agility’ (Haider et al. Reference Haider, Hentati-Sundberg, Giusti, Goodness, Hamann and Masterson2018), who are sensitive to their positionality, value inclusivity and have complementary skillsets. The difficulties of knowledge co-production through transdisciplinary research, such as managing power relations or working with limited time and funding, can be prohibitive to many organizations and academic researchers (Langston et al. Reference Langston, Riggs, Kastanya, Sayer, Margules and Boedhihartono2019, Turnhout et al. Reference Turnhout, Metze, Wyborn, Klenk and Louder2020). However, with minimal resources, the practices we outlined here have been integrated and incorporated into long-term processes of understanding and influencing change. These approaches often fail to meet the criteria against which the performance of academic researchers is assessed, and this rather open-ended research rarely meets the requirements of research funding agencies. We argue that more scientific resources should be deployed to support long-term place-based transdisciplinary science as a foundation within which targeted studies can identify technical solutions.
Discussion and conclusions
Knowledge of the biology and ecology of species and habitats is clearly important for conservation, but this knowledge increases enormously in value if it is embedded within a broader understanding of social-ecological systems and cultural context. What can be done in the biophysical realm is constrained by social, economic and political processes. Methods are needed to analyse complex multi-sectoral situations, drawing from empirical knowledge and from diverse disciplines to explore pathways for change (Kidd et al. Reference Kidd, Garrard, Bekessy, Mills, Camilleri and Fidler2019). Most importantly, these methods should embrace complexity and support the co-generation of knowledge by practicing science within landscapes, working with people, not on or for them (Boedhihartono et al. Reference Boedhihartono, Bongers, Boot, van Dijk, Jeans and van Kuijk2018).
We have deployed the tools described in this paper opportunistically in landscapes where our organizations were seeking to achieve conservation outcomes. The tools were not used in a planned, sequential manner. The use of the tools helped coalesce a community of practice within the landscapes; it fostered the emergence of a conservation narrative that was shared by the networks of actors who had a stake in the use of resources in the landscape. The benefits that came from the use of these techniques were incremental rather than transformational. Impact could not be measured against short-term biophysical indicators, but rather in improvements in governance of the social-ecological system. The approaches advocated here are by their nature intrusive. They assume the legitimacy of an outside agent to seek to influence conservation and development outcomes. In our actions, we are deliberately reflexive – we are influenced by local actors and we seek to influence them. There is increasing resistance in some quarters to what is perceived to be conservation colonialism (Agrawal Reference Agrawal1997). The practice of protecting biodiversity has often created and reinforced social hierarchies adverse to marginalized groups. Any outside agent seeking to pursue an agenda of the kind outlined above would need to establish its acceptance and legitimacy with local stakeholders. This might appear to be self-evident, yet conventional conservation biology often lacks local legitimacy. Recognition of diverse understandings of what conservation and environment mean to different people, and of approaches to bridging these understandings, is creating new pathways for legitimate and inclusive conservation (Sheil Reference Sheil, Sanz, Lewis, Mata and Connaughton2017, Norström et al. Reference Norström, Cvitanovic, Löf, West, Wyborn and Balvanera2020). The phase of inductive research that we propose is in line with these approaches, but even before this, any conservation organization should reflect upon what it is that gives it the right to seek to influence people’s lives. In many tropical situations, the people concerned are amongst the world’s poorest. All conservationists have to reflect upon the ethics of attempting to modify the lives of people who already live in very precarious conditions.
Transdisciplinary science is a collaborative process bringing together scientists from different disciplines with practitioners and local people to co-generate knowledge for solving practical problems (Hadorn et al. Reference Hadorn, Biber-Klemm, Grossenbacher-Mansuy, Hoffmann-Riem, Joye and Pohl2008, Popa et al. Reference Popa, Guillermin and Dedeurwaerdere2015). Societal challenges require integrated approaches, drawing from different types of knowledge and disciplines to solve complex problems. Transdisciplinary science emphasizes participation and process; it is a commitment to respecting and learning from contrasting perspectives and values (Stokols Reference Stokols2006). The co-generation of knowledge extends from collaborative problem-framing to impact, encompassing a diverse set of activities and understandings (Norström et al. Reference Norström, Cvitanovic, Löf, West, Wyborn and Balvanera2020). There is now strong recognition that knowledge co-generation should be attentive to inclusivity, power, process and reflexivity (Miller & Wyborn Reference Miller and Wybornin press). Transdisciplinary science is not just about creating new knowledge with new people, but embraces a deeper philosophy of how knowledge is held, created, shared and used.
Conservation scientists should be exposed to the consequences of the solutions that they advocate. Taleb (Reference Taleb2018) argues that scientists have a moral responsibility for the implications of their research findings. They have to be actors engaged in the broader processes of societal change and assert their influence from within those processes. The practices and their tools and methods that are outlined above enable scientists to understand and engage with broader development processes and seek to influence those processes to achieve better outcomes for conservation and for people. We do not seek to impose any particular model of development, but rather to enrich the processes through which local actors define their own futures. At present, conservation science is not investing enough in engaging with the people who will experience the immediate consequences of conservation interventions. It is not sufficient to theorize about conceptual frameworks for transdisciplinary research, nor to understand how social systems function, who influences that functionality and what alternative scenarios might exist, as useful as this is. A further step is necessary, which is to learn from practices and to share the understanding gained with the people whose livelihoods are impacted. When conservation scientists engage actively with the processes of change, better outcomes for biodiversity have been achieved.
Acknowledgements
Analyses of the kind illustrated here require a great deal of collaboration and cooperation. We acknowledge the engagement and support of many people from communities, local research and development organizations and government agencies in the locations in Africa, South America and South-East Asia where we have experimented with the approaches described here. Funding for some of the work in Eastern Indonesia was provided by the Critical Ecosystems Partnership Fund through Conservation International and the Centre for Tropical Environmental and Sustainability Science at James Cook University, Cairns, Australia. For the MCA, Daniel Juhn, Grace Wong and their colleagues at the Conservation International GIS lab derived the production suitability map and produced all of the other maps. Neville Kemp consulted local communities on their uses of local habitats.
Financial support
This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Conflict of interests
None.
Ethical standards
None.
