Mapping social-ecological systems to understand the challenges underlying wildlife management

(1)

http://www.diva-portal.org

This is the published version of a paper published in Environmental Science and Policy.

Citation for the original published paper (version of record):

Dressel, S., Ericsson, G., Sandström, C. (2018)

Mapping social-ecological systems to understand the challenges underlying wildlife management

Environmental Science and Policy, 84: 105-112 https://doi.org/10.1016/j.envsci.2018.03.007

Access to the published version may require subscription.

N.B. When citing this work, cite the original published paper.

Permanent link to this version:

http://urn.kb.se/resolve?urn=urn:nbn:se:umu:diva-145893

(2)

Contents lists available atScienceDirect

Environmental Science and Policy

journal homepage:www.elsevier.com/locate/envsci

Mapping social-ecological systems to understand the challenges underlying wildlife management

Sabrina Dressel

^a,⁎

, Göran Ericsson

^a

, Camilla Sandström

^b

aDepartment of Wildlife, Fish & Environmental Studies, Swedish University of Agricultural Sciences, SE-901 83 Umeå, Sweden

bDepartment of Political Science, Umeå University, SE-901 87 Umeå, Sweden

A R T I C L E I N F O

Keywords:

Environmental governance Panacea trap

Social-ecologicalﬁt Adaptive learning

A B S T R A C T

A holistic understanding of the complex interactions between humans, wildlife, and habitats is essential for the design of sustainable wildlife policies. This challenging task requires innovative and interdisciplinary research approaches. Using the newly implemented ecosystem-based management of moose (Alces alces) in Sweden as a case, we applied Ostrom’s social-ecological system (SES) framework to analyse the challenges that wildlife management faces throughout the country. We combined data derived from natural and social science research to operationalize the framework in a quantitative way; an approach that enabled a spatially explicit analysis on the national and regional levels. This study aimed to discover patterns in the social-ecological context of Swedish moose management. Identifying these patterns can provide input for an in-depth evaluation of the institutional ﬁt of the current system and subsequently for national policy development. Our SES maps suggest that there are spatial variations in factors challenging moose management. In some areas, ecological aspects such as the co- occurrence of carnivores and other ungulate species burdens future management, while in other regions challenges are shaped by governance aspects, e.g. diverse property rights. Theseﬁndings demonstrate that the new management system must apply adaptive learning principles to respond to local context attributes in order to be successful. Our innovative approach provides a valuable tool for the assessment of other natural resource management issues and the avoidance of panacea traps, especially when repeated over time.

1. Introduction

Managing wildlife (i.e. the processes of dealing with or controlling wildlife for diﬀerent purposes) in a sustainable way is a key challenge around the globe. To balance societal needs and ecological functions, the complex interactions between humans, wildlife, and habitats must be fully understood (Apollonio et al., 2017). Previously, our understanding of these relationships was limited by the disciplinary boundaries that restricted complex analyses (Berkes et al., 2008;Liu et al., 2007;Ostrom, 2009;Schlüter et al., 2014). To bridge the gap between social and ecological sciences research and to foster a holistic understanding of how humans interact with the surrounding ecosystem, a number of frameworks, e.g. social-ecological systems (SES) and human- environment systems (HES), have been developed (Binder et al., 2013).

These analytical frameworks aim to avoid the tendency of prescribing certain governance solutions or policy instruments as a panacea for environmental conﬂicts (Brock and Carpenter, 2007; Ostrom et al., 2007). The use of such one-size-ﬁts-all approaches as a simple solution

to complex issues has been highly unsuccessful (Cox, 2011; Ostrom et al., 2007). It includes the obvious risk of falling into panacea traps due to incorrect assumptions; notably that all resource governance problems can be represented by a small set of simple models and that most resource users have the same preferences and perceptions (del Mar Delgado-Serrano and Ramos, 2015;Ostrom, 2007;Ostrom et al., 2007).

Previous research has highlighted that sustainable management of natural resources depends on a thorough diagnostic procedure, which produces a holistic understanding of the system and assists the design of suitable policies (Cox, 2011; Ostrom, 2007; Schlüter et al., 2014;

Young, 2011). A misfit between social institutions (i.e. rules and norms) and ecological attributes can lead to conflicts and the unsustainable use of resources (Cumming et al., 2006;Folke et al., 2007; Leslie et al., 2015;Ostrom, 2009). Thus, the unique attributes of SES must be understood and considered in order tofind the best policy solutions (Brock and Carpenter, 2007;Folke et al., 2007;Ostrom, 2009). Moreover, the policies that set the objectives and institutions should not be static, but

https://doi.org/10.1016/j.envsci.2018.03.007

Received 20 September 2017; Received in revised form 10 March 2018; Accepted 11 March 2018

⁎Corresponding author.

E-mail address:sabrina.dressel@slu.se(S. Dressel).

Abbreviations: SES, social-ecological system; I, interactions; A, actors; GS, governance system; RS, resource system; RU, recourse units; MMA, moose management areas; MMU, moose management units; PCA, principal component analysis

Available online 20 March 2018

T

(3)

rather adapt over time and diversify according to diﬀerent spatial scales when needed (Brock and Carpenter, 2007; Olsson et al., 2007). Re- gional and local adaptations may be necessary to maximize adaptive capacity and ensure resilience within the framework of national policies (Berkes et al., 2008;Liu et al., 2007).

However, policies are often designed on an overarching and national perspective to enhance uniform solutions that promote predictability and rule of law, which might leave little or no room for local variation orflexibility (Ebbesson, 2010). This poses a dilemma from a sustainability perspective, as these types of policies include a prominent risk of creating a misfit between social institutions and ecological attributes, which will undermine the long-term governability and resilience of the system (Galaz et al., 2008;Young, 2011). This risk of a misfit further increases when the system changes over time. In the case of wildlife management temporal changes can likely happen due to increasing and/or spreading wildlife populations. Thus, any policy or institution that is designed to manage natural resources, such as wildlife, must include the capacity to handle diverse and changing ecological systems (Galaz et al., 2008;Levin et al., 2012;Young, 2011). Still most management strategies focus on a single natural resource and specific groups of users instead of adopting a system perspective. This overlooks the fact that most SES are complex, since the same resource system may contain several resources that compete with each other and are of varying importance to certain stakeholders (Dwyer and Hodge, 2016; Hinkel et al., 2015). One classic example of such a system are forest ecosystems that host valuable timber resources for forestry stakeholders, game species that are cherished by hunters, and a recrea- tional value for the public. Hence, the governance model for forest ecosystems should include policies and a mix of policy instruments that may contribute to the sustainable use of all competing resources and take different objectives into consideration. However, forest governance models rarely do so. For example, the Swedish forestry model has been struggling with the task of balancing the diverse policy objectives due, among other things, to the lack of a system perspective (Lindahl et al., 2015).

Being linked to the forest ecosystem Swedish moose (Alces alces) management is an example of a continuous strive to adjust and adapt to changes in the moose population and environment. Sweden has one of the world’s densest moose populations, but this was not always the case (Kardell, 2016). Both the population and corresponding management approaches have undergone major changes during the past century (Edenius et al., 2002;Sandström et al., 2013). Rationalization in agricultural practices in combination with the shift in forestry from single tree harvesting to large-scale clear cutting have opened up the forest structure and created large areas of land that provide moose with a suitable diet and habitat (Edenius et al., 2011;Kardell, 2016). This led to a rapid increase in the moose population and heavy browsing on certain tree species, causing not only economic losses for forest owners but also threatening the natural regeneration of these species and thereby biodiversity (Ericsson et al., 2001; Jaren et al., 2003). The previous management attempts were criticized for not being able to handle increasing conﬂicts, creating a mismatch between ecological and social scales, and disregarding the importance of a system perspective (Sandström et al., 2013).

Thus, as a response to increasing browsing pressure and conﬂicts among stakeholders (Sandström et al., 2013), Sweden established a local, ecosystem-based management system for moose in 2012 (Prop.

2009/10:239, NFS 2011:7; for more information seeAppendix A. De- tailed description of the current moose management system). Following the Malawi principles for the ecosystem approach (UNEP/CBD/COP/4/

INF/9;Jaren et al., 2003), its central components are decentralization of decision-making to local levels, the involvement of relevant actors to find a balance between different societal interests, as well as adaptive and knowledge-based management. A range of structures and institutions had to be created at various levels before the new system could take effect. For example, so-called moose management areas (MMA)

were established, with each area comprising a distinct moose population and requiring that the local landowner and hunter representatives come to an agreement about management plans and population goals (Bjärstig et al., 2014). These changes led to the introduction of a more comprehensive multi-level governance system, which aims to create a better match between social and ecological aspects.

The new ecosystem-based approach is therefore a promising starting point, but to be sustainable in the long run the management system must be carefully adapted to the local ecological and social circum- stances. For this reason, the present study aims to analyse the spatial patterns in social and ecological attributes that the newly designed institutions have to accommodate.

We use the SES framework developed by Ostrom and colleagues (McGinnis and Ostrom, 2014;Ostrom, 2007,2009) for our diagnostic procedure and apply it in a quantitative and spatially explicit way. The framework enables the integration of social and ecological aspects with equal analytical depth, multi-layered diagnostic procedures of a system andﬂexibility in choosing relevant variables (Binder et al., 2013;del Mar Delgado-Serrano and Ramos, 2015). As it is derived from the In- stitutional analysis and development (IAD) framework, the SES framework contains action situations in which interactions lead to certain outcomes in terms of sustainability (McGinnis and Ostrom, 2014). In- teractions (I) and social-ecological performance (outcomes, O) are directly shaped by involved actors (A), the governance system (GS) in place, the ecological resource system (RS) and attributes of the natural recourse units (RU), with each of these components providing feedback to the others (McGinnis and Ostrom, 2014).

So far, studies focused mainly on the action situation and took the system context less into consideration. Most attempts to operationalize the SES framework in a comparative way have either been rather de- scriptive in nature, limited in scale (with a primary focus on the local scale), or restricted in the choice and operationalization of variables (del Mar Delgado-Serrano and Ramos, 2015; Hamann et al., 2015;

Leslie et al., 2015; Thiel et al., 2015). These restrictions limit the transferability of case study results to diﬀerent systems. In contrast, we focus on the social-ecological context in which the action situations take place. Furthermore, the system presented in this paper oﬀers the possibility to apply the SES framework across a whole country, based on high-quality quantitative data derived from ecological monitoring, GIS or nationwide surveys on human-nature interactions. The variables within the SES framework have proven to play an important role in predicting outcome in terms of sustainability and therefore provide a clear picture of social and ecological attributes that the governance system has to acknowledge (Hinkel et al., 2015;Ostrom et al., 2007).

Consequently, we apply the SES framework to examine the current social-ecological context for moose management in Sweden to provide input for in-depth evaluations of the institutional ﬁt of the current system as well as further studies on action situations. We use an interdisciplinary approach to map spatial variations in the SES, which has rarely been done before (Hamann et al., 2015). Thus, the objectives of our study were to elucidate spatial variations in relevant context variables, which can provide a tool for national policy development and show a type of SES mapping that can be applied to other systems. In this way we respond to the recent call to use the framework to reach a place-based understanding of SES and contribute to the development of new methodological approaches for applying the framework in practice (Hamann et al., 2015;Karimi et al., 2015;Leslie et al., 2015).

2. Methods 2.1. Variable selection

The starting point for the diagnostic approach was the framework for analysing SES proposed by Ellinor Ostrom and colleagues (McGinnis and Ostrom, 2014; Ostrom, 2007, 2009). We used elements of an adapted version of the framework proposed byVogt et al. (2015), to

S. Dressel et al. Environmental Science and Policy 84 (2018) 105–112

106

(4)

enrich our understanding of the ecological attributes underlying SES (Fig. 1).

Ostrom (2009)highlighted that the selection of variables should be mainly driven by the research question at hand. Since we aimed to reach a place-based understanding of the challenges underlying wildlife management, we incorporated a stakeholder-based problem description into our variable selection and mapping. Based on a literature review, previous research (seeAppendix B. Material and Methods), and con- sultation with experts in the field, we selected first- and second-tier variables that are important in the context of moose management. We decided to focus on stakeholders who are directly involved in management and conflicts that shape their interactions. During this process, we used a participatory approach (i.e. a Q method to elicit critical social-ecological variables with an expert group of 35 wildlife managers from 20 county administrative boards; see Appendix B. Material and Methods for further details) to refine and validate our set of selected variables.Fig. 1 gives an overview of the 15 chosen variables (highlighted in bold). Each variable was operationalized in the context of moose management and suitable indicators were selected. We used iterative dialogues within the research group to identify data sources for indicators and how to calculate indicators in a meaningful way. We identified and collected data for 19 indicators representing the chosen social-ecological variables. Furthermore, for each variable/indicator, the situation that creates more challenges for management was speci- fied (Table 1; see Appendix B. Material and Methods for a detailed description of all indicators and our reasoning on challenges).

2.2. Data collection and spatial scale

The data used in this mapping approach originated from a variety of tools, ranging from ecological monitoring and GIS to quantitative nationwide surveys on human-nature interactions. Oﬃcial statistics and databases were used for most of the data retrieval. The Swedish National Forest Inventory provides detailed and freely available information on forest condition (e.g. tree species, side conditions, forest

damage). Statistic Sweden (SCB) coordinates and certifies official statistics from 26 agencies across the country (e.g. Swedish Forestry Agency, Swedish Board of Agriculture) and collects data on population statistics. The Swedish Association for Hunting and Wildlife Management provided us with annual bag statistics for all ungulate species. Data for two of the social variables (i.e. I4 Conflicts and A8 Importance of resource) were based on surveys that had been carried out by the authors during previous research (further details regarding the indicators can be found inAppendix B. Material and Methods. The underlying data is available via [dataset]).

Even though ecosystem-based management is applied, pragmatic political boundaries (i.e. county borders) are used to a certain extend within the management, and the system is applied autonomously in each county. Based on this, along with limitations in the spatial resolution of some of the indicators we have chosen counties (Swedish:

län) as the most appropriate level for our analysis. Averages for the county were calculated for variables that included data on a ﬁner spatial resolution. One of Sweden’s 21 counties, namely Gotland (Sweden’s largest island), was excluded from the mapping as there are no moose and thereby no management system implemented.

2.3. Data transformation, analyses and interpretation

Min-max normalization was applied to the data to produce indicators in the range [0, 1]. This transformation helped to combine primary data and enabled the comparison of variables across maps. For each indicator, the end of the data range that signiﬁed a more challenging situation for management was assigned a value of 1.

A principal component analysis (PCA) was performed on 19 continuous indicators measured across 20 counties. Principal component analysis (PCA) is a statistical method that uses orthogonal transformation to convert a group of potentially correlated variables into a set of linearly uncorrelated variables that capture variability in the underlying data (McGarigal et al., 2000). This procedure is vital to re- ducing the dimensions of multivariate problems. The set of linearly Resource systems (RS)

RS1 Sector (e.g., water, forests, pasture, fish) RS2 Clarity of system boundaries

RS3 Size of resource system RS4 Human-constructed facilities RS5 Productivity of system

RS5-b Community/species composition*

RS6 Equilibrium properties

RS7 Predictability of system dynamics RS8 Storage characteristics

RS9 Location

Actors (A) A1 Number of relevant actors

A1-a Diversity of relevant actors A1-b Relative number of relevant actors A2 Socioeconomic attributes

A3 History or past experiences A4 Location

A5 Leadership/entrepreneurship A6 Norms/social capital

A7 Knowledge of SES/mental models A8 Importance of resource A9 Technologies available Resource units (RU)

RU1 Resource unit mobility RU2 Growth or replacement rate RU3 Interaction among resource units

RU3-b Competition between species*

RU3-c Predation*

RU4 Economic value RU5 Number of units

RU5-b Absolute size*

RU5-c Relative size*

RU6 Distinctive characteristics RU7 Spatial and temporal distribution

Governance systems (GS) GS1 Government organizations GS2 Nongovernment organizations GS3 Network structure GS4 Property-rights systems GS5 Operational-choice rules GS6 Collective-choice rules GS7 Constitutional-choice rules GS8 Monitoring and sanctioning rules Interactions (I)

I1 Harvesting I2 Information sharing I3 Deliberation processes I4 Conflicts

I5 Investment activities I6 Lobbying activities I7 Self-organizing activities I8 Networking activities I9 Monitoring activities I10 Evaluative activities

Outcomes (O) Related ecosystems (ECO) Social, economic, and political settings (S)

Fig. 1. Social-ecological system framework adapted from Ostrom (McGinnis and Ostrom, 2014;Ostrom, 2007,2009). We focused onﬁve ﬁrst-level subsystems (RS, RU, I, GS, A);

therefore, only variables related to these subsystems are listed on subsequent levels. * The initial framework has been supplemented with variables fromVogt et al. (2015). Bold font colour indicates which second- and third-tier variables have been selected and operationalized for the diagnostic procedure of moose management in Sweden.

(5)

Table1 OverviewofselectedSESvariablesandtheindicatorsusedtooperationalizethem. 1sttiervariable2ndand3rdtiervariableOperationalizationIndicatorChallenge RSRS1SectorLandusediversityDiversityoflandcovertype(e.g.forestry,agriculture)Morechallengingwhenmultipleland-usetypeshavetobeintegrated RS3SizeofresourcesystemSizeofmoosepopulationrangeSizeofmoosemanagementareaMorechallengingwhenawidegeographicextentneedstobe considered(thisalsocorrespondstomoosemigration) RS5-bSpeciescompositionForageavailabilityIndexofmooseforageavailabilityMorechallengingwhenspeciescompositionprovideslittleformoose toforage RS7Predictabilityofsystem dynamicsFluctuationinforageavailabilityVariationinmooseforageavailabilityover10yearsMorechallengingwhenforageavailabilityfluctuatesconsiderablyover time RURU3-bCompetitionbetween speciesPresenceofotherungulatesPresenceofotherungulatespeciesMorechallengingwhenmultipleungulatespeciescompeteoverhabitat andforage RU3-cPredation(1)PredationbybearsPresenceofbears(Ursusarctos)Morechallengingwhenlargecarnivorespreyonmoose (2)PredationbywolvesPresenceofwolves(Canislupus) RU5-bAbsolutesizeMoosedensityNumberofshotmoosepersquarekilometerMorechallengingwhenmoosedensityishigh RU5-cRelativesizeProportionofungulatepopulationRatioofmoosetootherungulatespeciesMorechallengingwhenmooserepresentasmallportionoftheoverall ungulatecommunity II4Conflicts(1)TrafficaccidentsNumberofmoose-car-collisionsMorechallengingwhenmoreaccidentsoccur (2)BrowsingdamageFreshbrowsingdamageonScotspine(Pinussylvestris)Morechallengingwhenbrowsingdamageishigh (3)DisagreementonpopulationgoalsPotentialforconflictindexonmoosemanagersevaluationofmoose populationMorechallengingwhenthedegreeofdisagreementbetweenmanagers ishigh I7Self-organizingactivitiesLevelofself-organizationintoMMUGeographiccoverageofmoosemanagementunits(MMU)Morechallengingwhenthelevelofself-organizationislow GSGS3NetworkstructuresSub-unitsperMMANumberofsub-units(i.e.licenseareas,MMU’s)permoosemanagement area(MMA)Morechallengingwhenthenumberofsub-unitsishigh GS4Property-rightssystems(1)DiversityofforestryownershipDiversityindexofforestownershiptypesMorechallengingwhenthepropertyrightssystemishighlydiverse (2)DiversityofagriculturalownershipDiversityindexofagricultureownershiptypes AA1-aDiversityofrelevantactorsForestownerdiversityPropertysizeclassesofprivateforestownersMorechallengingwhenthereisamixtureofsmallandlargeprivate forestowners A1-bRelativenumberofrelevant actorsNumberofrelevantactorsProportionofgeneralpublicthatarerelevantactors(i.e.holdersof huntinglicenses,forestowners,agriculturalbusinesses)Morechallengingwhenalargepartofthepublicisdirectlyinvolved A8ImportanceofresourceImportanceofmoosemeatFrequencyofmoosemeatconsumptionMorechallengingwhenthemooseisconsideredtobeanimportant resource

108

(6)

uncorrelated variables identiﬁed through PCA comprises principal components, with the ﬁrst principal component accounting for the largest possible variance in the original data (McGarigal et al., 2000).

Given that PCA was used for exploring and describing patterns in the complex data set rather than conﬁrmatory testing of hypotheses, we assumed the multivariate normality to be suﬃcient. We based the PCA on the correlation matrix to give equal weight to all variables.

For the mapping the PCA results were visualized across counties, using a linear combination of the centred and scaled observations with the eigenvectors as coefficients. The created maps illustrate spatial patterns in social and ecological attributes that could influence in- stitutionalfit. Previous research as well as discussions with involved stakeholders helped us interpret the generated SES maps (seeAppendix B. Material and Methods). All data analyses were performed in JMP 10.0.2 (SAS Institute Inc. Cary, NC, 1989–2007) and all maps were generated in ESRI ArcGIS Desktop 10.4.1.

3. Findings 3.1. Component scores

We retained four principal components that explained 78% of the variance in our data, with the first, second, third and fourth components accounting for 38.8, 17.7, 13.3 and 7.8% of the variation, re- spectively (Table 2). Thefirst component can be classified as a gradient progressing from social importance to ecological diversity and contains variables from allfive of the first-level subsystems (RS, RU, I, GS, A).

This component had high positive loadings from importance of moose meat (A8), number of relevant actors (A1b), size of moose population range (RS3), predation by bears (RU3c_1) and traffic accidents (I4_1). Land use diversity (RS1), presence of other ungulates (RU3b), diversity of agricultural ownership (GS4_2) and forest owner diversity (A1a) showed negative loadings on thefirst component (Table 2). Component two was loaded by forage availability (RS5b), level of self-organization into MMU (I7) and moose density (RU5b), the latter demonstrated a negative loading (Table 2). The number of sub-units per MMA (GS3) had positive loadings on component one and two. The third component yielded negative factor loadings for browsing damage (I4_2) and fluctuation in forage

availability (RS7), but a positive loading for diversity of forestry ownership (GS4_1). Component four was deﬁned by two variables, with a negative loading by disagreement on population goals (I4_3) and a positive loading by predation by wolves (Ru3c_2) explaining the relationship between these two variables (Table 2).

3.2. Spatial patterns

The mapping of each county onto the PCA eigenvectors revealed several distinct sub-regions (Fig. 2). Theﬁrst principal component can be perceived as a north-south gradient describing the transition from social importance to ecological diversity (Fig. 2). Management challenges in northern Sweden are driven by the involvement of many di- rect stakeholders, who value the moose as an important resource, with large resource systems leading to many vertical partners in management. Higher numbers of bears and a relatively large amount of moose- car collisions are typical for these areas. Conversely, the region is also characterized by rather homogenous land use (dominated by forestry), little competition from other ungulate species, private forest owners with similar property sizes and low diversity in agricultural ownership.

In the south of Sweden, challenges in management occur due to diversity in land use, agricultural ownership and individual forest property sizes. Furthermore, from an ecological perspective, management is challenged by the common occurrence of multispecies-systems in which other ungulate species are more abundant than moose.

The second principal component was deﬁned by low forage availability for moose and low levels of self-organization of actors in moose management units, both of these attributes were predominant in the northern- and southern-most counties of Norrbotten and Skåne. The opposite end of the scale is encountered in central parts of Sweden and characterized by high moose density (Fig. 2).

The projection of the data onto the third principal component di- vides Sweden into north, middle and south sections. While northern and southern counties are characterized by high browsing damage and highﬂuctuation in forage availability, moose management in central counties is challenged by diverse forestry ownership structures (i.e. a diverse mix of forest companies, state-owned forests and private forest owners) (Fig. 2).

The fourth principal component was dominated by only two variables, high wolf predation and high disagreement over population goals among local mangers. There was no clear pattern throughout Sweden, yet central areas showed higher degrees of wolf predation whereas disagreement over population goals was prevalent in northern counties (Fig. 2).

4. Discussion

The results demonstrate that our approach can dismantle distinct spatial patterns and provide a deeper understanding of the social-ecological context of natural resource management. Thus, our approach can contribute a tool for designing locally-adapted institutions in order to avoid issues resulting from a misconception of the initial model (Brock and Carpenter, 2007;Levin et al., 2012). The disregarding of spatial heterogeneity can lead even highly advanced management approaches, such as adaptive co-management, into panacea traps (Plummer and Hashimoto, 2011;Young, 2011).

The studied case of Swedish moose management showed that there may be a need for diﬀerent adaptation strategies to avoid a growing

“problem of ﬁt” (Folke et al., 2007;Galaz et al., 2008) between social and ecological aspects. A spatial misﬁt might result from strong variation in the size of the resource system and number of involved actors (component 1). Moose management areas in northern Sweden cover an average of up to 17.000 km²and the management is coordinated with up to 150 sub-units. On the other hand, southern moose management areas can comprise less than 750 km²and have as little as six sub-units.

Nevertheless, the same institutional design is applied to both regions. In Table 2

Factor loadings of principal component analysis.

Operationalization Variable Component

1 2 3 4

Land use diversity RS1 −0.760 – – –

Size of moose population range RS3 0.809 – – –

Presence of other ungulates RU3b −0.869 – – –

Predation by bears RU3c_1 0.829 – – –

Proportion of ungulate population

RU5c −0.568 – – –

Traﬃc accidents I4_1 0.770 – – –

Sub-units per MMA GS3 0.644 0.605 – –

Diversity of agricultural ownership

GS4_2 −0.788 – – –

Forest owner diversity A1a −0.749 – – –

Number of relevant actors A1b 0.922 – – –

Importance of moose meat A8 0.789 – – –

Forage availability RS5b – 0.804 – –

Moose density RU5b – −0.797 – –

Level of self-organization into MMU

I7 – 0.808 – –

Fluctuation in forage availability

RS7 – – −0.779 –

Browsing damage I4_2 – – −0.723 –

Diversity of forestry ownership GS4_1 – – 0.835 –

Predation by wolves RU3c_2 – – – 0.622

Disagreement on population goals

I4_3 – – – −0.685

(7)

practice this means that independent of the size of the management area the same number of actors (i.e. three hunter and three landowner representatives) design management plans, coordinate management actions, and are responsible for collaboration among units. Besides this spatial misfit a functional misfit might originate from the emergence of multi-species systems in which up tofive other ungulate species co-exist and compete with moose. In northern counties, the moose is by far the most common ungulate species whereas in southern counties up to 63 other ungulates are shot for each culled moose. Regardless of this dif- ference, the current institutional design takes only moose into consideration. To avoid a functional misfit some counties should actively incorporate other species into the management plan.

The presented research suggests that Swedish moose management could be an emerging panacea trap. While adaptive co-management often has been prescribed for the stewardship of natural resources within complex and changing SES (Armitage et al., 2009), attention has been drawn to the risks of applying uniform solutions that ignore the context (Plummer and Hashimoto, 2011). In our case, it seems that instead of tailoring the management to social and ecological context, a strive for more uniform regulations, pragmatic administrative deci- sions, and the strong historical focus on the moose influenced the design of the management system. As a result, the institutions in place have a different fit with the discovered subsystems and their characteristics. Interpreting our results through the lens of multi-level learning principles (Pahl-Wostl, 2009), we suggest that it is time to consider transformations of the system for governing wildlife in Sweden to include spatial variations in its context. Therefore, the moose management system should be seen as a step in adaptive learning rather than afinal solution to initially identified conflicts.

So far, most of the adaptation is focused on environmental objectives and thereby reﬂects classical single-loop learning. Current management plans allow for a yearly adjustment of culling rates to mini- mize browsing damage and promote a healthy moose population. Quick adaptations of institutions and the governance system cannot be rea- lized to the same extent. Laws and policies focus on robustness, stability and uniformity on a national level, which may hamper the spatial,

temporal and functional adaptation of the system. True adaptive governance, including adaptive institutions and multi-level learning principles (Berkes, 2009;Pahl-Wostl, 2009), should however be applied if ﬂexibility and improvements in the governing structures are desired.

While these theoretical concepts are often highlighted as the key to success (Boyd and Folke, 2011;Koontz et al., 2015), real-world multi- scale applications are rare. Close feedback between multi-level governance structures is needed to allow not only for top-down adaptation to social-ecological variations but to also enable societal and institutional learning (Berkes, 2009;Koontz et al., 2015).

Local actors and their place-based collective actions can contribute to innovative institutional solutions and the successful design of management structures on higher levels (Ratner et al., 2013). Institution building needs to be understood as an evolutionary process in which stakeholders should be included and empowered to provide valuable local knowledge and develop best-practices that may contribute to the overall legitimacy of newly-shaped policies (Berkes, 2009; Rammel et al., 2007). This idea also guided our mapping approach, as we actively tried to integrate stakeholder-based problem assessment into our variable selection and interpretations (seeAppendix B. Material and Methods for more details on our participatory approach). While the current management system might work perfectly in some places, it may struggle in regions where ecological attributes (e.g. the presence of other ungulate species or large carnivores) and/or social attributes (e.g.

the number or diversity of actors) add additional challenges. Thus, our results highlight that policies designed from a national perspective, which predominantly try to enhance uniform solutions that will promote predictability and rule of law, bear the risk of overlooking important spatial variations.

5. Conclusions

Applying our comparative mapping approach on a regional scale provided insights into the full spectrum of governance and management challenges, which would not have been possible by evaluating Sweden as a whole. The identiﬁed spatial patterns can serve as a starting point Fig. 2. Social-ecological system maps of the challenges underlying Swedish moose management. Unique patterns and sub-regions were identiﬁed when county data was projected onto the four PCA eigenvectors, shown by (1) a gradient with social importance on the positive side of the scale and ecological diversity on the negative side; (2) a gradient with low self- organization on the positive side of the scale and high moose density on the negative side; (3) a gradient with diverse forest owners on the positive side of the scale and browsing damage on the negative side; (4) a gradient with wolf predation on the positive side of the scale and actor disagreement on the negative side.

110

(8)

for an in-depth analysis of how suitable the existing policy design is, which can provide recommendations for necessary policy adaptations.

Our study delivered a multi-faceted and heterogeneous snapshot picture of the current context of moose management in Sweden.

Repeated runs of the presented SES mapping approach over time could help researchers and policy-makers to identify changes in social-ecological systems, which is vital to the timely adaptation of policies and institutions.

Throughout the mapping process we experienced several hurdles in operationalizing the SES framework that had previously been noted by others (del Mar Delgado-Serrano & Ramos, 2015;Leslie et al., 2015).

Some of the variables are rather ambiguous or abstract, and thereby diﬃcult to operationalize on a local level (Hinkel et al., 2014; Thiel et al., 2015). Furthermore, we suspect that some of the variables include inherent thresholds; for example, there may be an optimum number of involved actors (Ostrom, 2009). As there is currently a lack of knowledge on the individual and collective responses and interactions in the system, we could not account for all social attributes. Social heterogeneity, such as diﬀerent interests within the same stakeholder group, should be integrated to a higher extent in future studies.

Notwithstanding these limitations, we see our approach as a bene- ﬁcial tool for analysing and illustrating multi-dimensional environmental issues and helping to tailor policies to the social-ecological context. Therefore we want to encourage future development of social- ecological mapping. We are aware that we had a data rich case, but given the growing availability of national and international datasets on social and environmental measures we see a good potential for the application of this approach in diverse settings.

The fact that the patterns we discovered were shaped by a combination of social and ecological factors, with thefirst principal component containing variables from allfive of the first-level subsystems (RS, RU, I, GS, A), highlights the need for interdisciplinary research. An analysis of the system from a purely social or ecological approach would most probably neglect certain drivers of complexity. Integrating participatory aspects into variable selection and thereby allowing a stakeholder-based framing of the issue enriched our quantitative approach and extended our understanding beyond patterns and structures. Thus, using a portfolio of methods can contribute to a better understanding of complex social-ecological systems.

Acknowledgements

We thank all involved participants who contributed valuable insight for our data collection and interpretations. This study was supported by the Swedish Environmental Protection Agency’s Wildlife Management Fund [grant number 802-0161-15], Future Forests a multi-disciplinary research programme supported by the Foundation for Strategic Environmental Research (MISTRA), the Swedish Forestry Industry, the Swedish University of Agricultural Sciences (SLU), Umeå University, and the Forestry Research Institute of Sweden.

Appendix A. Supplementary data

Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.envsci.2018.03.007.

References

Apollonio, M., Belkin, V.V., Borkowski, J., Borodin, O.I., Borowik, T., Cagnacci, F., Danilkin, A.A., Danilov, P.I., Faybich, A., Ferretti, F., Gaillard, J.M., Hayward, M., Heshtaut, P., Heurich, M., Hurynovich, A., Kashtalyan, A., Kerley, G.I.H., Kjellander, P., Kowalczyk, R., Kozorez, A., Matveytchuk, S., Milner, J.M., Mysterud, A., Ozoliņš, J., Panchenko, D.V., Peters, W., Podgórski, T., Pokorny, B., Rolandsen, C.M., Ruusila, V., Schmidt, K., Sipko, T.P., Veeroja, R., Velihurau, P., Yanuta, G., 2017. Challenges and science-based implications for modern management and conservation of European ungulate populations. Mammal Res. 62, 209–217.http://dx.doi.org/10.

1007/s13364-017-0321-5.

Armitage, D.R., Plummer, R., Berkes, F., Arthur, R.I., Charles, A.T., Davidson-Hunt, I.J.,

Diduck, A.P., Doubleday, N.C., Johnson, D.S., Marschke, M., McConney, P., Pinkerton, E.W., Wollenberg, E.K., 2009. Adaptive co-management for social–ecological complexity. Front. Ecol. Environ. 7, 95–102.http://dx.doi.org/10.1890/

070089.

Berkes, F., 2009. Evolution of co-management: Role of knowledge generation, bridging organizations and social learning. J. Environ. Manage. 90, 1692–1702.http://dx.doi.

org/10.1016/j.jenvman.2008.12.001.

Berkes, F., Colding, J., Folke, C., 2008. Navigating Social-ecological Systems: Building Resilience for Complexity and Change. Cambridge University Press, Cambridge.

Binder, C.R., Hinkel, J., Bots, P.W.G., Pahl-Wostl, C., 2013. Comparison of frameworks for analyzing social-ecological systems. Ecology Society 18.http://dx.doi.org/10.5751/

ES-05551-180426.

Bjärstig, T., Sandström, C., Lindqvist, S., Kvastegård, E., 2014. Partnerships implementing ecosystem-based moose management in Sweden. Int. J. Biodivers. Sci. Ecosyst. Serv.

Manage. 10, 228–239.http://dx.doi.org/10.1080/21513732.2014.936508.

Boyd, E., Folke, C., 2011. Adapting Institutions: Governance, Complexity and Social- ecological Resilience. Cambridge University Press, Cambridge.

Brock, W.A., Carpenter, S.R., 2007. Panaceas and diversiﬁcation of environmental policy.

Proc. Natl. Acad. Sci. 104, 15206–15211.http://dx.doi.org/10.1073/pnas.

0702096104.

Cox, M., 2011. Advancing the diagnostic analysis of environmental problems. Int. J.

Commons 5.

Cumming, G.S., Cumming, D.H.M., Redman, C.L., 2006. Scale mismatches in social- ecological systems: causes, consequences, and solutions. Ecol. Soc. 11.http://dx.doi.

org/10.5751/ES-01569-110114.

del Mar Delgado-Serrano, M., Ramos, P., 2015. Making Ostrom’s framework applicable to characterise social ecological systems at the local level. Int. J. Commons 9.http://dx.

doi.org/10.18352/ijc.567.

Dwyer, J., Hodge, I., 2016. Governance structures for social-ecological systems: Assessing institutional options against a social residual claimant. Environ. Sci. Policy 66, 1–10.

http://dx.doi.org/10.1016/j.envsci.2016.07.017.

Ebbesson, J., 2010. The rule of law in governance of complex socio-ecological changes.

Global Environ. Change 20, 414–422.http://dx.doi.org/10.1016/j.gloenvcha.2009.

10.009.

Edenius, L., Bergman, M., Ericsson, G., Danell, K., 2002. The role of moose as a dis- turbance factor in managed boreal forests. Silva Fennica 36, 57–67.http://dx.doi.

org/10.14214/sf.550.

Edenius, L., Ericsson, G., Kempe, G., Bergström, R., Danell, K., 2011. The eﬀects of changing land use and browsing on aspen abundance and regeneration: a 50-year perspective from Sweden. J. Appl. Ecol. 48, 301–309.http://dx.doi.org/10.1111/j.

1365-2664.2010.01923.x.

Ericsson, G., Edenius, L., Sundström, D., 2001. Factors aﬀecting browsing by moose (Alces Alces L.) on European aspen (Populus tremula L.) in a managed boreal landscape.

Ecoscience 8, 344–349.http://dx.doi.org/10.1080/11956860.2001.11682662.

Folke, C., Pritchard, J.L., Berkes, F., Colding, J., Svedin, U., 2007. The problem ofﬁt between ecosystems and institutions: Ten years later. Ecol. Soc. 12.http://dx.doi.

org/10.5751/ES-02064-120130.

Galaz, V., Olsson, P., Hahn, T., Folke, C., Svedin, U., 2008. The problem ofﬁt among biophysical systems, environmental and resource regimes, and broader governance systems: Insights and emerging challenges. In: Young, O.R., King, L.A., Schröder, H.

(Eds.), Institutions and Environmental Change - Principal Findings, Applications, and Research Frontiers. The MIT Press, Cambridge, MA pp. 147-182.

Hamann, M., Biggs, R., Reyers, B., 2015. Mapping social–ecological systems: Identifying

‘green-loop' and ‘red-loop' dynamics based on characteristic bundles of ecosystem service use. Global Environ. Change 34, 218–226.http://dx.doi.org/10.1016/j.

gloenvcha.2015.07.008.

Hinkel, J., Bots, P.W.G., Schlüter, M., 2014. Enhancing the Ostrom social-ecological system framework through formalization. Ecol. Soc. 19.http://dx.doi.org/10.5751/

ES-06475-190351.

Hinkel, J., Cox, M.E., Schlüter, M., Binder, C.R., Falk, T., 2015. A diagnostic procedure for applying the social-ecological systems framework in diverse cases. Ecol. Soc. 20.

http://dx.doi.org/10.5751/ES-07023-200132.

Jaren, V., Sinclair, A., Andersen, R., Danell, K., Schwartz, C., Peterson, R.O., Bowyer, R.T., Ericsson, G., 2003. Moose in modern integrated ecosystem management - how should the Malawi principles be adapted? Alces 39, 1–10.

Kardell, Ö., 2016. Swedish forestry, forest pasture grazing by livestock, and game browsing pressure since 1900. Environ. Hist. 22, 561–587.

Karimi, A., Brown, G., Hockings, M., 2015. Methods and participatory approaches for identifying social-ecological hotspots. Appl. Geogr. 63, 9–20.http://dx.doi.org/10.

1016/j.apgeog.2015.06.003.

Koontz, T.M., Gupta, D., Mudliar, P., Ranjan, P., 2015. Adaptive institutions in social- ecological systems governance: A synthesis framework. Environ. Sci. Policy.http://

dx.doi.org/10.1016/j.envsci.2015.01.003.

Leslie, H.M., Basurto, X., Nenadovic, M., Sievanen, L., Cavanaugh, K.C., Cota-Nieto, J.J., Erisman, B.E., Finkbeiner, E., Hinojosa-Arango, G., Moreno-Baez, M., Nagavarapu, S., Reddy, S.M., Sanchez-Rodriguez, A., Siegel, K., Ulibarria-Valenzuela, J.J., Weaver, A.H., Aburto-Oropeza, O., 2015. Operationalizing the social-ecological systems framework to assess sustainability. Proc. Natl. Acad. Sci. U. S. A. 112, 5979–5984.

http://dx.doi.org/10.1073/pnas.1414640112.

Levin, S., Xepapadeas, T., Crépin, A.-S., Norberg, J., de Zeeuw, A., Folke, C., Hughes, T., Arrow, K., Barrett, S., Daily, G., Ehrlich, P., Kautsky, N., Mäler, K.-G., Polasky, S., Troell, M., Vincent, J.R., Walker, B., 2012. Social-ecological systems as complex adaptive systems: modeling and policy implications. Environ. Dev. Econ. 18, 111–132.http://dx.doi.org/10.1017/S1355770X12000460.

Lindahl, K.B., Sténs, A., Sandström, C., Johansson, J., Lidskog, R., Ranius, T., Roberge, J.- M., 2015. The Swedish forestry model: More of everything? For. Policy Eco. 45–55.

(9)

http://dx.doi.org/10.1016/j.forpol.2015.10.012.

Liu, J., Dietz, T., Carpenter, S.R., Alberti, M., Folke, C., Moran, E., Pell, A.N., Deadman, P., Kratz, T., Lubchenco, J., 2007. Complexity of coupled human and natural systems.

Science 317, 1513–1516.

McGarigal, K., Cushman, S., Staﬀord, S.G., Staﬀord, S.G., 2000. Multivariate Statistics for Wildlife and Ecology Research. Springer New York, New York.

McGinnis, M.D., Ostrom, E., 2014. Social-ecological system framework: initial changes and continuing challenges. Ecol. Soc. 19.http://dx.doi.org/10.5751/ES-06387- 190230.

Olsson, P., Folke, C., Galaz, V., Hahn, T., Schultz, L., 2007. Enhancing the Fit through Adaptive Co-management: Creating and Maintaining Bridging Functions for Matching Scales in the Kristianstads Vattenrike Biosphere Reserve, Sweden. Ecol. Soc.

12.http://dx.doi.org/10.5751/ES-01976-120128.

Ostrom, E., 2007. A diagnostic approach for going beyond panaceas. Proc. Natl. Acad. Sci.

104, 15181–15187.http://dx.doi.org/10.1073/pnas.0702288104.

Ostrom, E., 2009. A General Framework for Analyzing Sustainability of Social-Ecological Systems. Science 325, 419–422.http://dx.doi.org/10.1126/science.1172133.

Ostrom, E., Janssen, M.A., Anderies, J.M., 2007. Going beyond panaceas. Proc. Natl.

Acad. Sci. 104, 15176–15178.http://dx.doi.org/10.1073/pnas.0701886104.

Pahl-Wostl, C., 2009. A conceptual framework for analysing adaptive capacity and multi- level learning processes in resource governance regimes. Global Environ. Change 19, 354–365.http://dx.doi.org/10.1016/j.gloenvcha.2009.06.001.

Plummer, R., Hashimoto, A., 2011. Adaptive co-management and the need for situated thinking in collaborative conservation. Hum. Dimens. Wildl. 16.http://dx.doi.org/

10.1080/10871209.2011.585434.

Rammel, C., Stagl, S., Wilﬁng, H., 2007. Managing complex adaptive systems — A co- evolutionary perspective on natural resource management. Ecol. Econ. 63, 9–21.

http://dx.doi.org/10.1016/j.ecolecon.2006.12.014.

Ratner, B., Meinzen-Dick, R., May, C., Haglund, E., 2013. Resource conﬂict, collective action, and resilience: an analytical framework. Int. J. Commons 7.http://dx.doi.

org/10.18352/ijc.276.

Sandström, C., Wennberg DiGasper, S., Öhman, K., 2013. Conﬂict resolution through ecosystem-based management : the case of Swedish moose management. Int. J.

Commons 7, 549–570.http://dx.doi.org/10.18352/ijc.349.

Schlüter, M., Hinkel, J., Bots, P.W.G., Arlinghaus, R., 2014. Application of the SES framework for model-based analysis of the dynamics of social-ecological systems. Ecol.

Soc. 19.http://dx.doi.org/10.5751/ES-05782-190136.

Thiel, A., Adamseged, M.E., Baake, C., 2015. Evaluating an instrument for institutional crafting: How Ostrom's social–ecological systems framework is applied. Environ. Sci.

Policy 53, 152–164.http://dx.doi.org/10.1016/j.envsci.2015.04.020.

Vogt, J.M., Epstein, G.B., Mincey, S.K., Fischer, B.C., McCord, P., 2015. Putting the "E" in SES: unpacking the ecology in the Ostrom social-ecological system framework. Ecol.

Soc. 20.http://dx.doi.org/10.5751/ES-07239-200155.

Young, O.R., 2011. Eﬀectiveness of international environmental regimes: Existing knowledge, cutting-edge themes, and research strategies. Proc. Natl. Acad. Sci. 108, 19853–19860.http://dx.doi.org/10.1073/pnas.1111690108.

Sabrina Dressel is PhD candidate in the Department of Wildlife, Fish & Environmental Studies at the Swedish University of Agricultural Sciences (SLU). Her research focuses on the adaptive capacity within Swedish moose management and the design of future wildlife governance. She has a BSc in International Forest Ecosystem Management (University for Sustainable Development, Germany) and a MSc (Distinction) in Wildlife Ecology and Wildlife Management (University of Natural Resources and Life Sciences, Austria). She is administrating and co-teaching a course on Human Dimensions of Fish and Wildlife Management (15 ECTS, advanced level) at SLU Umeå.

Göran Ericsson is Professor of Wildlife Ecology, and of Fish and Wildlife Tourism at the Swedish University of Agricultural Sciences (SLU). He focuses on plant-animal interactions, the use of the biological resource base and the human user groups, and interactions between the resource base and the users. As SLU’s strategic leader for “Animal Ecology”

he promotes and encourages basic research, their application for a wider societal user, and outreach. A key feature of their studies is multi-species habitat use and movement ecology using state-of-the-art technology and analytical approaches. He is one of the European forerunners of Human Dimensions of Wildlife.

Camilla Sandström is Professor in Political Science at Umeå University. Her research focuses on the governance and management of natural resources from a systems as well as institutional perspective. Through diagnostic approaches such as the mapping of policies and governance systems, her research contributes to the identiﬁcation of problems and prospects related to policy design, institutional arrangement and management methods.

She has contributed extensively to the development of Human Dimensions of Wildlife in Europe.

112