Biosphere Reserves’ Management E ﬀ ectiveness—A Systematic Literature Review and a Research Agenda

: Research about biosphere reserves’ management e ﬀ ectiveness can contribute to better understanding of the existing gap between the biosphere reserve concept and its implementation. However, there is a limited understanding about where and how research about biosphere reserves’ management e ﬀ ectiveness has been conducted, what topics are investigated, and which are the main ﬁndings. This study addresses these gaps in the ﬁeld, building on a systematic literature review of scientiﬁc papers. To this end, we investigated characteristics of publications, scope, status and location of biosphere reserves, research methods and management e ﬀ ectiveness. The results indicate that research is conceptually and methodologically diverse, but unevenly distributed. Three groups of papers associated with di ﬀ erent goals of biosphere reserves were identiﬁed: capacity building, biodiversity conservation and sustainable development. In general, each group is associated with di ﬀ erent methodological approaches and di ﬀ erent regions of the world. The results indicate the importance of scale dynamics and trade-o ﬀ s between goals, which are advanced as important leverage points for the success of biosphere reserves. Building on the gaps identiﬁed in the literature, a research agenda is proposed, focusing on the need to investigate mechanisms for holistic research, outcomes and trade-o ﬀ s, transformations for social-ecological ﬁt and institutions for integrated management across scales.


Introduction
Biosphere reserves are privileged places to understand how to sustainably manage and govern social-ecological systems [1] and to advance sustainability science [2,3]. The World Network of Biosphere Reserves (WNBR) currently contains 701 designated sites, distributed over 124 countries [4]. The relevance and broad interest in the biosphere reserve enterprise does not translate, however, into a successful implementation, as there is a considerable gap between the concept and its practical realization [5][6][7]. Research about biosphere reserves' management effectiveness can provide a better understanding of why there is this gap [1, 8,9], what is its extension [10] and how it can be closed [11][12][13]. However, there is a limited understanding about where and how the research about biosphere reserves' management effectiveness has been conducted, what topics have been investigated and which are the main findings. This study addresses these gaps in the literature.
The selection of the papers used in this study results from the systematic literature review conducted by Ferreira et al. [1], to develop a conceptual framework that summarizes which factors are important to biosphere reserves' management effectiveness. An overview of how the papers were selected is given here, however, a more profound description can be found in Ferreira et al. [1]. A literature search was conducted in the Scopus database on 10 March 2017 (search string in Appendix A).
Only peer-reviewed papers published in English were included. Papers published before 1996 were excluded in order to focus the analysis on the more integrative approach which biosphere reserves have adopted after the Statutory Framework [16]. The abstract, title and keywords of 2286 potentially relevant papers were screened against the following inclusion criteria: (i) engagement with the biosphere reserve concept; (ii) useful to understand management and governance of biosphere reserves; and (iii) is an empirical study. Another reviewer evaluated 10% of the papers to identify disagreements in the paper selection process. From the 177 papers downloaded (9 papers were not accessible), those that performed comparative analysis [10,21] were excluded, to obtain only the studies that were developed in one biosphere reserve. Research conducted in biosphere reserves that were not present in the UNESCO databases in June 2017 [22,23] were also excluded, such as the study of Schmidt et al. [24] that was developed in a biosphere reserve yet to be designated. The references of the 66 publications obtained are disclosed in Table A1. The search string used, and the selection process, ensured a high specificity for peer-reviewed literature related to management and governance of biosphere reserves. This explains why only a small part of the existing scientific literature was included. Similar results were obtained in a bibliometric analysis of biosphere reserves' research [25]: most of the existing research is developed in the biosphere reserves, but not necessarily about them.

Definition of Subcategories
To analyze the literature, a set of categories and subcategories were defined, related to four main topics: (i) features of the publication; (ii) scope, status and location of the biosphere reserve where the study was developed; (iii) methods used in the research; and (iv) management effectiveness (Table 1). Scale * * Context, inputs, process and outcomes subcategories were coded for their relevance at international or national/regional scales.
A total of 147 subcategories were used to review the papers (Table A2). They were adapted from existing classifications, such as the classification of countries according to the UNESCO MAB regions [22], or inductively developed, e.g., the subcategories of research methods. To analyze the main findings concerning management effectiveness, the framework developed by Ferreira et al. [1] was used. This framework describes 53 general factors, grouped in four main categories-context, Sustainability 2020, 12, 5497 4 of 31 inputs, processes and outcomes-which were identified as important for understanding biosphere reserves' management effectiveness. An overview of the framework subcategories is given in Table A3.
Data about the publication was retrieved from ELSEVIER [26], and data about the biosphere reserves was retrieved from UNESCO databases [22,23]. Data about the research and biosphere reserves' management effectiveness was coded in the 66 papers using MAXQDA Plus ver. 12 (VERBI Software, Berlin, Germany) [27]. To accommodate information that did not fit in the defined subcategories, "other" options were included. Coding was only performed in the Results section of each paper. Text from other parts of the paper was coded, if necessary, to understand the results. Multiple codes could be assigned to the same text segment. The text was interpreted in order to associate text chunks to the codes, guided by the definitions of each of the subcategories. All aspects of the social-ecological systems where biosphere reserves are implemented, and the management and governance systems in place, were coded.

Data Analysis
To access the main patterns in the data, descriptive statistics was used in R ver. 3.4.3 (R Foundation for Statistical Computing, Vienna, Austria) [28]; plots were developed using the ggplot2 package for R [29]. A cluster analysis [30] was performed to identify groups of publications that address biosphere reserves' management effectiveness in a similar way. The variables, the clustering method and the number of clusters were determined in a back-and-forward procedure. A database with the presence/absence of context, processes, inputs and outcomes subcategories (n = 53) in the 66 papers was used (Table A2). A distance matrix was developed using the Jaccard Index, as implemented in the vegdist function of the vegan package for R [31]. After testing different clustering methods, the ward.D was selected to continue the cluster analysis because of its interpretability and the strong clustering structure, as revealed by the agglomerative coefficient (Table A4).
The optimal number of clusters was determined by evaluating and interpreting different cluster solutions in relation to the generality and specificity of the results. Multidimensional scaling (MDS) was also used to determine if groups in the data can be visually identified. Vectors of external variables significantly correlated with the dissimilarities among papers were fitted in the MDS, as implemented in the envfit function of the vegan package for R, in order to explore the influence of: (i) the methods used for data analysis, (ii) the MAB region where the study took place and (iii) if the study was conducted in a biosphere reserve designated before or after the Statutory Framework.
The dissimilarities among the groups of papers obtained from the cluster analysis were investigated by conducting a permutational multivariate analysis of variance (PERMANOVA), using the Jaccard distance matrix and 999 permutations, as implemented in the adonis function of the vegan package for R [31]. The analysis was repeated for each pairwise comparison among clusters. The regression coefficients from each PERMANOVA were used to identify the subcategories that most contributed for the dissimilarities among the clusters tested.
The proportion of papers that refer each subcategory in each cluster was computed, to identify the common subcategories that are very frequently referred (in more than 70% of the papers included in each cluster).
To analyze the outcomes, the subcategories of social benefits, empowerment and learning were merged in "positive social outcomes"; and the subcategories of social impacts and inequality were merged in "negative social outcomes". Then, the number of papers that refer a given positive or negative outcome (environmental, social, cultural or economic) in each biosphere reserve was summed.
To evaluate the importance of scales in management effectiveness, the proportion of papers that refer each subcategory at international or national/regional scales was calculated, in relation to the total number of papers that refer each subcategory.

Characteristics of the Publications
From 1998 to March 2017, the number of publications related with management and governance of biosphere reserves have generally increased, despite annual variations ( Figure A1). The number of studies published in journals related to environmental or social sciences is higher than in other fields of research ( Figure A2).
The first authors of the analyzed papers have affiliations in Europe and North America (57.6%), Asia and the Pacific (25.7%), and Latin America and the Caribbean (16.7%). The relationship between the author's affiliations and the region where the study was developed is represented in Figure 1. Authors from Europe and the USA and Canada have developed studies in a diversity of MAB regions. About 87% of the research from lead authors from USA or Canada is developed in the Latin America and the Caribbean, and none in their own biosphere reserves. Authors from Europe, Asia and the Pacific, and Latin America and the Caribbean have developed studies mainly in their own respective regions. From 1998 to March 2017, the number of publications related with management and governance of biosphere reserves have generally increased, despite annual variations ( Figure A1). The number of studies published in journals related to environmental or social sciences is higher than in other fields of research ( Figure A2).
The first authors of the analyzed papers have affiliations in Europe and North America (57.6%), Asia and the Pacific (25.7%), and Latin America and the Caribbean (16.7%). The relationship between the author's affiliations and the region where the study was developed is represented in Figure 1. Authors from Europe and the USA and Canada have developed studies in a diversity of MAB regions. About 87% of the research from lead authors from USA or Canada is developed in the Latin America and the Caribbean, and none in their own biosphere reserves. Authors from Europe, Asia and the Pacific, and Latin America and the Caribbean have developed studies mainly in their own respective regions. where the studies were developed (inferior row). The MAB region "Europe and North America" is divided in "Europe" and "USA and Canada".

The Biosphere Reserves Studied
The papers analyzed performed their research in a total of 38 different biosphere reserves (Table  A5). The higher number of studies was conducted in the Maya (Guatemala), Nanda Devi (India), Wolong (China), Danube Delta (Romania/Ukraine) and El Vizcaíno (Mexico) biosphere reserves. Two transboundary biosphere reserves were analyzed-Gerês/Xurés, in Portugal and Spain, and the Danube Delta, between Romania and Ukraine-however, only the study in Gerês/Xurés was performed for the entire transboundary biosphere reserve. At the time the data was analyzed, none of the investigated biosphere reserves have withdrawn the network. The biosphere reserves studied were designated between 1977 and 2012; 42% before and 68% after the adoption of the Seville Strategy.
In total, single case studies about management/governance were performed in about 6% of the designated biosphere reserves. The countries where more than three studies were performed are: Mexico (n = 21), Guatemala (n = 9), India (n = 8) and China (n = 7). Among the countries with a higher number of sites designated, only Mexico and China have studies developed in more than 10% of their biosphere reserves (Figure 2). . The MAB region "Europe and North America" is divided in "Europe" and "USA and Canada".

The Biosphere Reserves Studied
The papers analyzed performed their research in a total of 38 different biosphere reserves (Table A5). The higher number of studies was conducted in the Maya (Guatemala), Nanda Devi (India), Wolong (China), Danube Delta (Romania/Ukraine) and El Vizcaíno (Mexico) biosphere reserves. Two transboundary biosphere reserves were analyzed-Gerês/Xurés, in Portugal and Spain, and the Danube Delta, between Romania and Ukraine-however, only the study in Gerês/Xurés was performed for the entire transboundary biosphere reserve. At the time the data was analyzed, none of the investigated biosphere reserves have withdrawn the network. The biosphere reserves studied were designated between 1977 and 2012; 42% before and 68% after the adoption of the Seville Strategy.
In total, single case studies about management/governance were performed in about 6% of the designated biosphere reserves. The countries where more than three studies were performed are: Mexico (n = 21), Guatemala (n = 9), India (n = 8) and China (n = 7). Among the countries with a higher number of sites designated, only Mexico and China have studies developed in more than 10% of their biosphere reserves (Figure 2).

Research Methods
Studies related to biosphere reserves' management effectiveness have used a median number of three different methods for data collection ( Figure A3a), mostly interviews, document analysis and observation ( Figure 3a). Almost 91% of the studies involved actors in data collection. Half of the studies involved two different actors (median, Figure A3b), mainly local communities and governmental actors (Figure 3b). Considering the data analysis, qualitative methods were used in about 58% of the papers alone; in about 29% of the papers, mixed qualitative and quantitative methods were used; and exclusive quantitative methods were used in only 13% of the papers.

Research Methods
Studies related to biosphere reserves' management effectiveness have used a median number of three different methods for data collection ( Figure A3a), mostly interviews, document analysis and observation ( Figure 3a). Almost 91% of the studies involved actors in data collection. Half of the studies involved two different actors (median, Figure A3b), mainly local communities and governmental actors ( Figure 3b).

Research Methods
Studies related to biosphere reserves' management effectiveness have used a median number of three different methods for data collection ( Figure A3a), mostly interviews, document analysis and observation ( Figure 3a). Almost 91% of the studies involved actors in data collection. Half of the studies involved two different actors (median, Figure A3b), mainly local communities and governmental actors (Figure 3b). Considering the data analysis, qualitative methods were used in about 58% of the papers alone; in about 29% of the papers, mixed qualitative and quantitative methods were used; and exclusive quantitative methods were used in only 13% of the papers.  Considering the data analysis, qualitative methods were used in about 58% of the papers alone; in about 29% of the papers, mixed qualitative and quantitative methods were used; and exclusive quantitative methods were used in only 13% of the papers.

Cluster Analysis
A cluster analysis was developed to assess how the scientific literature is related according to the subcategories of biosphere reserves' management effectiveness [1] (Figure 4). A partitioning with three clusters provided the ideal trade-off between specificity and generality of the results ( Figure A4). This solution achieved a high agglomerative coefficient (0.87; maximum of 1), however, according to the cluster evaluation statistics (Table A6), there is some uncertainty about which papers should be clustered together. The MDS (Figure 4b) also show some overlap between the groups, in particular between clusters #1 and #2.  There are, however, significant differences in the composition of the subcategories across all clusters (P < 0.001). The first 20 subcategories (i.e., coefficients) that most contribute to the dissimilarities between clusters are aggregated in Table 2. Comparing to the other clusters, the papers  22 25 14  60  63  9  36  2  1  12  45  55  58  56  35  57  50  41  15  39  42  61  27  43  46  51  52  65  23  53  18  62  19  44  16  59  54  20  37  5  31  7  29  10  49  17  3  13  6  8  21  4  11  34  38  47  64  66  32  40  33  48  24  26  There are, however, significant differences in the composition of the subcategories across all clusters (P < 0.001). The first 20 subcategories (i.e., coefficients) that most contribute to the dissimilarities between clusters are aggregated in Table 2. Comparing to the other clusters, the papers included in the cluster #1 (n = 32) are more associated with subcategories of governance (empowerment, participatory processes, management body) and learning (information, type of knowledge, learning). Papers included in the cluster #2 (n = 23) are more focused on subcategories related with biodiversity conservation (conservation and habitat management, economic and social impacts) and activities associated with it (cultural use of natural resources, material investments and infrastructure, cultural benefits). The subcategory that mostly contributes to the dissimilarities between papers included in the cluster #3 (n = 11) and the others is economic benefits. The subcategories mainly associated with the papers in each cluster are, therefore, related with three goals of the biosphere reserves: capacity building, conservation of biodiversity and sustainable development. Moreover, some of the subcategories that most contribute to the dissimilarities among the clusters are shared between cluster #1 and cluster #2 (Table 2). This suggests the existence of similar factors that influence the achievement of the goals of "capacity building" and "conservation of biodiversity", but not "sustainable development". Further analysis of the most frequent subcategories referred in each cluster revealed the existence of more similarities: socio-economic attributes of the context and the restrictions and incentives implemented in the biosphere reserve being studied are referred in more than 70% of the papers included in each cluster. The big majority of the publications in each cluster investigate the management/governance of a project in biosphere reserves.
We also assessed the contributions of other variables to the dissimilarities among papers, namely the methods used for data analysis, the MAB region where the study was conducted, and the study being developed in a biosphere reserve designated before or after the Seville Strategy. The methods used for data analysis are very strong predictors of the dissimilarities among papers (P < 0.001), as represented in Figure 4b. Quantitative methods are more correlated with papers in the third cluster (sustainable development), qualitative methods with papers in the first cluster (capacity building) and mixed methods with papers in the second cluster (biodiversity conservation). The MAB region where the study was performed is also correlated (P < 0.01), however, the predictors are weaker. Papers in cluster #3 are more correlated with studies developed in Asia and the Pacific and cluster #1 in Latin America and the Caribbean. The second cluster includes studies conducted in a diversity of regions. The study being developed in a biosphere reserve designated before or after the Seville Strategy is not significantly correlated with the dissimilarities among papers. These results are indicative of how studies developed using different methods of data analysis and in different regions of the world are also associated with different subcategories of biosphere reserves' management effectiveness.

Outcomes
From the 66 papers analyzed, 43 report at least one positive outcome; 49 at least one negative outcome, and 32 both positive and negative outcomes. The number of papers that report environmental, economic, cultural and social positive and negative outcomes in each biosphere reserve is represented in Figure 5. For most of the biosphere reserves, both positive and negative outcomes were reported.
regions. The study being developed in a biosphere reserve designated before or after the Seville Strategy is not significantly correlated with the dissimilarities among papers. These results are indicative of how studies developed using different methods of data analysis and in different regions of the world are also associated with different subcategories of biosphere reserves' management effectiveness.

Outcomes
From the 66 papers analyzed, 43 report at least one positive outcome; 49 at least one negative outcome, and 32 both positive and negative outcomes. The number of papers that report environmental, economic, cultural and social positive and negative outcomes in each biosphere reserve is represented in Figure 5. For most of the biosphere reserves, both positive and negative outcomes were reported.  Publications (n) Figure 5. Sum of the number of publications that report positive and negative outcomes regarding social, economic, environmental and cultural aspects, in each biosphere reserve. Positive outcomes include, e.g., the provision of jobs, decrease of conflicts, empowerment, motivation, learning, provision of recreation opportunities or increase of the population of an endangered species. Negative outcomes include, e.g., decrease of households' income, displacement of communities, inequality, frustration, erosion of traditions or the overexploitation of natural resources.

Scales
The subcategories most frequently identified at international or national/regional scales are represented in Table 3. The most frequently referred subcategories are related to the context and inputs to management/governance processes, namely funding and material support/opposition. Only cultural outcomes, such as the provision of opportunities for recreation, are frequently referred at international or national/regional scales, which reflect the benefits derived by tourists. Table 3. Most frequent referred subcategories at international or national/regional scales. The proportion of papers refers to the number of papers in which a subcategory is present.

General Patterns of the Literature
The scientific literature related to biosphere reserves' management effectiveness has increased in the last decades, following the general trend of the biosphere reserves' research [25]. However, the number of papers that conduct studies related with biosphere reserves' management effectiveness in one biosphere reserve is very limited [1]. Complementing these papers with studies that were excluded from the literature review, e.g., because they do not mention "biosphere reserve" in the abstract, title and keywords, and with gray literature, would be important to provide a more complete overview of biosphere reserves' management effectiveness. Moreover, more recent research [32][33][34], should also be included, since our review only includes papers published until March 2017.
As indicated by the journal subject area, environmental and social sciences are the main disciplines contributing to the research about biosphere reserves' management effectiveness. Therefore, despite the results of this study indicating a limited contribution of other disciplines, management effectiveness comprises a more heterogeneous field of study than biosphere reserves' research, which is manly related with natural and environmental sciences [25].
The results of this study also demonstrate that lead authors from Europe and North America have been responsible for a big part of the research about biosphere reserves' management effectiveness, including in other regions of the world. This result echoes the findings of other studies that examined the authors of sustainability-related research [35], demonstrating the need of greater geographic diversity.
Research about biosphere reserves' management effectiveness is methodologically diverse, using multiple methods for data collection and analysis. In contrast with large-scale studies about biosphere reserves' management effectiveness [8,19,20], local communities are the privileged actor included in the research. Most of the actors enrolled are, however, only consulted in interviews or surveys, and few studies have applied more profound methods of stakeholders' engagement.

The Biosphere Reserves Studied
According to the results, studies related with biosphere reserves' management effectiveness have been developed in only about 6% of the designated sites, and are mainly concentrated in four countries: Mexico, Guatemala, India and China. These countries are also amongst those that have developed more research, in general, in biosphere reserves [25].
In the literature analyzed, only one study covered the whole biosphere reserve-in the Gerês-Xurés between Portugal and Spain [36]. Within the WNBR, there are actually 21 transboundary biosphere reserves [4]. A better understanding of the management and governance of transboundary biosphere reserves is necessary, given their increased complexity.
Despite further information about biosphere reserves' management effectiveness which can be found in studies that compare biosphere reserves at a global scale [8,19,[37][38][39], among countries [40,41] or at a national scale [9,10], existing data is insufficient to provide a comprehensive understanding of the effectiveness of the WNBR. The generalized lack of research related with management effectiveness in biosphere reserves reinforces the claim that biosphere reserves have been underutilized in terms of their potential contribution to the theory and practice of sustainability science [3].

The Topics Investigated
The literature related with biosphere reserves' management effectiveness frequently addresses the implementation of restrictions to reduce environmental harms (e.g., regulation and surveillance of the use of marine resources [42]), incentives to promote more environmentally friendly behaviors (e.g., a conservation-oriented language school [43]), as well as the socio-economic characteristics of the settings where these processes are implemented (e.g., demography and sources of income in the community [44]). Moreover, three groups of papers were identified which investigate, more profoundly, topics related to: (1) governance and learning; (2) activities associated with biodiversity conservation; and (3) economic incentives to sustainable development. These groups are related to the goals of biosphere reserves: (1) capacity building, (2) biodiversity conservation and (3) sustainable development.
The literature analyzed does not cover, however, the four goals of biosphere reserves, according to the MAB Strategy 2015-2025 [2]. Some areas, in which more research seems to be important in each goal, and examples of papers found in the literature that addresses these topics are: (1) research activities, investigated in the study of Alonso-Yañez and Davidsen [45]; (2) environmental outcomes, as examined by Mehring and Stoll-Kleemann [46] and Steinberg et al. [47]; (3) equity, that was studied in the papers of Sundberg [48,49] and health, investigated in the study of Sylvester et al. [50]; and (4) climate change adaptation and mitigation, briefly addressed in Durand et al. [51].

Methods and Context
The results of this study indicate that the goals of biosphere reserves-capacity building, biodiversity conservation and sustainable development-have been mainly investigated using, respectively, qualitative, mixed and quantitative approaches. This result suggests that a holistic understanding of biosphere reserves' management effectiveness requires the use of multiple approaches. Other studies have highlighted that different lenses and perspectives are required for the understanding and management of complex [52] social-ecological systems [53]. Conceptual and methodological plurality may also increase the possibility of finding solutions for wicked problems [54]. Research about biosphere reserves' management effectiveness should, therefore, combine different methodological approaches and a diversity of actors, in order to include different perspectives about the complex social-ecological systems being managed.
The results of this study also indicate that research related with capacity building and sustainable development have been mainly conducted in, respectively, Latin America and the Caribbean, and in Asia and the Pacific; the literature related with the goal of biodiversity conservation is geographically more diverse. These results concur with previous works that underscore the importance of the context in biosphere reserves' management effectiveness [1] and in integrated conservation strategies [55]. The seminal work of Ostrom [56] highlights the need to move beyond panaceas, i.e., simple universal recipes, to resolve the problems of overuse of natural resources. Research about biosphere reserves' management effectiveness should focus, therefore, on co-creating and investigating management and governance processes that are embedded in the social-ecological contexts in which biosphere reserves are implemented. The criteria for the designation of a biosphere reserve should also be critically analyzed, in order to avoid the prescription of simple solutions (e.g., zoning or participatory management) to solve complex problems.

Goals
The cluster analysis conducted in this study revealed that the classification of the scientific literature according to subcategories of biosphere reserves' management effectiveness [1] reflect the goals of the MAB Program. Some of the subcategories associated with each goal are: (1) capacity building-information, knowledge, management body, participatory processes, empowerment and learning; (2) biodiversity conservation-cultural and extractive use of natural resources, conservation and habitat management, socio-economic impacts and cultural benefits; and (3) economic development-economic benefits. This pattern suggests that the goals of biosphere reserves influence which processes are developed, which inputs are needed, and, consequently, the outcomes of their management, in a given context. The goals are, therefore, of central importance to biosphere reserves' management effectiveness. This result concurs with research about complex systems that underscores the importance of the goals of the system in determining its behavior [52]. Due to this, the goals are among the most important leverage points to change systems [52,57]. This suggests that closing the gap between biosphere reserves' concept and practice [5][6][7] may be more effectively achieved by addressing the goals of biosphere reserves. This result provides a different perspective about key factors for the success of biosphere reserves, which have been mainly associated with the participation, designation or the availability of resources [8,10,19,20].
The focus of biosphere reserves in sustainable development [16] and in the Sustainable Development Goals (SDGs) agenda [2,58] may require, therefore, a critical analysis. These concepts have been criticized for promoting economic growth on a finite planet [59,60] and for resulting from a Western construct that ignores existing cultural alternatives and worldviews of human-nature relationships [61]. Therefore, it seems to be important to investigate alternative approaches that provide more fundamental and context-specific transformations in biosphere reserves, such as Buen Vivir (South America), Ubuntu (South Africa), Swaraj (India) and degrowth (Europe) [13,59].

Interdependencies between Goals and Across Scales
In this study, there were identified subcategories that are associated with specific goals of biosphere reserves and subcategories which seem to be important for multiple goals (e.g., the implementation of restrictions and incentives, and the socio-economic context). The results also indicate the presence of trade-offs among outcomes of biosphere reserves-in most of the biosphere reserves studied, both positive and negative outcomes were reported. Many factors that influence management, but which control lies outside of biosphere reserves, were reported in the literature: funding to develop its activities [62,63], goals of the organizations [45,64], economic crises [36], power issues [49] and formal rules [65]. These results are indicative of the interdependencies between goals of biosphere reserves, and between biosphere reserves and the larger systems in which they are contained. Managers of biosphere reserves, therefore, have to articulate different goals, in order to prevent that the achievement of one goal compromises others, or the purpose of the biosphere reserve, and also to consider factors that, despite originating outside of biosphere reserves, may influence its effectiveness. How biosphere reserves navigate these scale dynamics between the systems they contain and in which they are contained is, therefore, an important topic for future research.
The existence of trade-offs in biosphere reserves concurs with existing research about win-win strategies, i.e., initiatives that aim to achieve conservation and development goals. Win-win situations rarely materialize; instead, gains and losses are the norm [66,67]. While some authors suggest that the irreconcilability between conservation and development have to be recognized in order to adequately deal with trade-offs and "hard decisions" [67], others claim that the apparent incompatibility between environmental and economic activities is an artefact of neoliberal conservation approaches [68]. By not considering the unequal access to natural resources, and relying on economic growth to end poverty, neoliberal conservation instruments exacerbate the conservation-development conflicts they were meant to resolve [68]. Given the contested nature of this topic, and the importance of trade-offs to biosphere reserves' management effectiveness, more research about the causes of trade-offs in biosphere reserves, and how to overcome them, is necessary.

A Research Agenda
Building on the topics discussed above, a research agenda, and some recommendations, are proposed to advance inquiry about biosphere reserves' management effectiveness ( Table 4). The proposed research agenda is in accordance with existing suggestions to advance investigation in sustainability of social-ecological systems [69] or sustainability science [70], and also with the current action plan for biosphere reserves [58]. These similarities suggest that biosphere reserves can benefit from the advancement of these fields of research, and vice versa. Collaborative work between these research communities, and with practitioners in biosphere reserves, can, therefore, contribute to leverage theory and practice of sustainability. Investigate the contribution of biosphere reserves to the development of research, environmental outcomes, equity and health in the regions in which they are implemented; (ii) What changes are needed to assure that management/governance of biosphere reserves is orientated to achieve a more balanced mix of social, cultural, economic and environmental outcomes?
Investigate the causes of trade-offs in biosphere reserves and how to overcome them.

Social-ecological fit
What transformations are needed to assure that the goals of biosphere reserves, criteria for designation and management/governance processes, fit the social-ecological contexts in which they are implemented?
Investigate the fit between biosphere reserves' goals, criteria for designation and management/governance processes, and the social-ecological contexts in which they are implemented; Critically analyze the pursue of sustainable development and the SDGs in biosphere reserves; Investigate how context-orientated transformations can be incorporated in biosphere reserves.

Scales
What new institutional mechanisms, or changes in existing institutions, are required to facilitate the management and governance of scale dynamics in biosphere reserves?
Study what mechanisms can facilitate the integrated management of the multiple goals of biosphere reserves; Analyze how multi-scale and large-scale cooperation can be promoted to achieve social-ecological benefits in biosphere reserves, and the role of UNESCO in this regard.
Lastly, it is important to highlight the important role that UNESCO can play in potentiating research about biosphere reserves' management effectiveness. Existing databases containing information about biosphere reserves [22,23] should be improved, in order to provide a more complete source of data. Current shortcomings include unavailability of data (e.g., periodic reviews and spatial boundaries), data that is not updated, and lack of systematic information between biosphere reserves (e.g., information about the main ecosystems) and between both databases. Despite progress being made regarding the systematization of literature about biosphere reserves [71], further work is still necessary to disclose and better understand topics related to management effectiveness. The categories and subcategories analyzed in this study, including those of Ferreira et al.
[1], could be useful in this regard. Besides providing a characterization of the context, processes, inputs and outcomes associated with biosphere reserves' management/governance, the subcategories also allow to understand how and where data was collected in the first place. The systematization of such information would be useful, not only to biosphere reserves' managers and researchers, but also, to build theory about how to sustainably manage and govern social-ecological systems on a regional scale.

Conclusions
Using a systematic literature review of the scientific literature, this study aimed to contribute to a better understanding about where and how the research about biosphere reserves' management effectiveness have been conducted, which topics have been investigated and what the main findings are.
The results indicate that, in line with their multiple goals and complex processes of implementation, the research about biosphere reserves' management effectiveness is diverse-it investigates different topics in different locations-and plural, because it includes different conceptual perspectives and methodological approaches. Three groups of papers, that address different subcategories of the context, inputs, processes and outcomes of biosphere reserves, were identified. These groups are associated with different goals of the program: capacity building, biodiversity conservation and sustainable development. In general, the papers in each group use different methodological approaches and were developed in different regions of the world. Given the importance of the goals in structuring the scientific literature according to subcategories of management effectiveness, the goals of biosphere reserves are proposed as effective leverage points to increase their success. The results also suggest the importance of scale dynamics and interdependencies between goals in biosphere reserves' management effectiveness.
However, there were identified gaps and bias in the literature that prevent a more holistic understanding of biosphere reserves' management effectiveness. In order to advance inquiry in this important topic, a research agenda for the field, and some recommendations, are proposed, focusing on the need to investigate mechanisms for holistic research, outcomes and trade-offs, transformations for social-ecological fit and institutions for an integrated management across scales. The pursue of this research agenda may contribute to biosphere reserves becoming real laboratories for sustainable development, in all its dimensions and diversity. Moreover, collaborative work between different research communities, and practitioners in biosphere reserves, i.e., managers, local communities and other stakeholders [58], would be important to leverage theory and practice of sustainability.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.   Table A2. Categories and subcategories used to review the publications. Subcategories used in the cluster analysis are identified with "1" in the last column ("C"). The subcategory "health benefits" was not used in the cluster analysis because it was not coded in any paper. NA-Not applicable; BR-Biosphere reserve; MAB-UNESCO Man and Biosphere Program; NGO-Non-governmental organizations.       Table A4. Criteria used to select the clustering method. The interpretability was considered hard when the dendrograms form long chains or reversals [74]. The cluster analysis was performed using the hclust function of the stats package for R.   Table A4. Criteria used to select the clustering method. The interpretability was considered hard when the dendrograms form long chains or reversals [74]. The cluster analysis was performed using the hclust function of the stats package for R. Publications (no.)

Clustering method
Journal subject area Figure A2. Subject area of the journals where the studies about biosphere reserves' management effectiveness were published, according to ELSEVIER [26]. Journals can belong to more than one subject area.     Figure A4. Definition of the optimal number of clusters according to: (a) the silhouette index [74] and (b) a scree plot. The different solutions were also interpreted to decide which one provides a better relationship between the specificity and generality of the results. Table A6. Cluster evaluation statistics. To evaluate the internal quality of the clustering, the average silhouette width-s(i)-was used, as computed in the silhouette function of the cluster package for R. Values around 0 indicate that observations lie between two clusters; well-clustered solutions have an average s(i) close to 1. To evaluate the robustness of the clustering, the clusterwise Jaccard bootstrap mean was used, as computed in the clusterboot function of the fpc package for R with 100 resampling runs. Following Zumel and Mount [75], clusters with stability values lower than 0.6 are unstable and values of stability between 0.6 and 0.85 indicate patterns in the data, but there is a high uncertainty about which observations should be clustered together.  Average silhouette width Figure A4. Definition of the optimal number of clusters according to: (a) the silhouette index [74] and (b) a scree plot. The different solutions were also interpreted to decide which one provides a better relationship between the specificity and generality of the results.