Challenges and Solutions for Forest Biodiversity Conservation in Sweden: Assessment of Policy, Implementation Outputs, and Consequences

Abstract

Swedish policies aim at conserving biological production, biodiversity, cultural heritage and recreational assets. This requires compositionally and structurally functional networks of representative habitats, the processes that maintain them, and resilient ecosystems. The term green infrastructure (GI) captures this. We review (1) policy concerning forest biodiversity conservation from the 1990s; (2) the implementation outputs, including the formulation of short-term and evidence-based long-term goals for protected areas, education, and the development of hierarchical spatial planning; (3) the consequences in terms of formally protected and voluntarily set-aside forest stands, as well as conservation management and habitat restoration. We assess the successes and failures regarding policy, outputs and consequences, discuss challenges to be addressed, and suggest solutions. Policies capture evidence-based knowledge about biodiversity, and evidence-based conservation planning as an output. However, the desired consequences are not met on the ground. Thus, the amount of formally protected and voluntary set-aside forests are presently too low, and have limited quality and poor functional connectivity. GI functionality is even declining because of forestry intensification, and insufficient conservation. Challenges include limited collaborative learning among forest and conservation planners, poor funding to conserve forest habitats with sufficient size, quality and connectivity, and national politics that ignores evidence-based knowledge. As solutions, we highlight the need for diversification of forest management systems with a landscape perspective that matches forest owner objectives and regional social-ecological contexts. This requires integrative approaches to knowledge production, learning and spatial planning.

1. Introduction

To protect, manage and restore ecosystem services (ES) for human well-being and welfare, e.g., [1,2], the continued alteration, fragmentation and loss of natural forest and cultural woodland biotopes and urban green space must be tackled. Simultaneously, increased production of provisioning ES on forest and agricultural land is desired, and more space is used for grey infrastructures, e.g., [3]. At the same time, global climate change [4], knowledge resistance [5] and an increasingly ideologically polarized world affecting migration and competition for natural resources [6,7] have led to uncertainties. This stresses the urgency to improve the adaptive capacity and secure resilient social-ecological ecosystems. Coping with these complexities is a challenge to the sustainable development (SD) process toward sustainability—a development that implies that finite resources and the environment are not consumed or degraded in an irrevocable manner, to the detriment of future generations, e.g., [8].

In Sweden, a long history of forest management for sustained yield wood production [9,10,11] has moved managed forest landscapes far away from their natural and historical range of variability in terms of forest disturbance processes and functions, e.g., [12,13,14,15], habitat structures, e.g., [16,17] and species composition, e.g., [18,19,20]. In urban landscapes, green spaces shrink as housing, transport, communication and energy and different kinds of grey infrastructure demand more space [21], which is a development that poses threats to human health [22].

Concerns about species’ extinction emerged in Sweden more than a century before the term biodiversity appeared [23]. Already in 1877, Säve [24] called for actions to halt the loss of species. While the Swedish Parliament passed an act for the establishment of National Parks to protect nature for the benefit of science and tourism in 1909, modern forest conservation in terms of setting aside protected areas with habitat for species emerged only in the mid to late 20th century [25]. The State Forests (Domänstyrelsen) began set-aside of forest areas (Domänreservat) in 1913, and stipulated nature considerations in managed forests in 1924 [26,27]. Public reactions against intensive forest management, such as to establish forest plantations on cultural woodlands [28], large clear-cuts [29], use of herbicides to remove the deciduous trees in young forests in the 1970s [30], and loss of old-growth forest [31], triggered this development.

The Nature Conservation Act of 1964 included the creation of nature reserves. The Swedish Environmental Protection Agency was created in 1967, focusing on environmental protection including nature conservation and outdoor recreation [32]. The increase in protected areas from the 1980s was based on the first nation-wide old-growth forests inventory conducted between 1978 and 1981 [33].

In 1993, the Swedish parliament deregulated the national forest policy, and established an environmental goal on par with the previous, long-standing goal of high wood production. The establishment of a series of national environmental objectives adopted by the Swedish parliament in 1999 triggered increased efforts to protect forest areas. The “Living Forests” objective with the interim target to increase the amount of formally protected forest by 400,000 ha, and the area voluntarily set-aside forests by 500,000 ha in productive forests below the mountain region before the end of 2010 was a result of this increased effort [34,35].

To communicate the need for improved biodiversity conservation and the provisioning of ecosystem services supporting human well-being, a range of policies targeting the sustainable use of natural resources have emerged since the 1990s at EU and Swedish policy levels, e.g., [36,37,38,39]. The synonymous concepts functional habitat network and green infrastructure (GI), used in both research [40] and policy [37], captures the urgency to sustain adaptive and resilient ecosystems with sufficient functional connectivity of representative land overs in landscapes and regions. These can then provide ecosystem services, and support adaptation to climate change and socio-economic drivers.

However, the evidence-based policy vision forming the foundation for GI work is in stark contrast to the present poor functionality of habitat networks in Sweden [20,41,42,43]. Captured in the current Swedish forest policy [44], the Government’s environmental quality objectives and the strategy for biodiversity and ecosystem services [44], the Government [45] commissioned a range of authorities to produce guidelines and plans for implementation of regional action GI plans at the level of county administrations.

While biodiversity conservation and GI development have been clearly pronounced in Sweden through international and national laws and policies, e.g., [25,38,46], the subsequent implementation process in terms of creating protected areas needs to be assessed regarding its effectiveness to satisfy agreed policy. In a review of formally and informally protected forest area development 1991–2010, Angelstam et al. [41] showed that establishment of functional GIs involves a long chain from policy creation via the outcomes of implementation processes including outputs in terms of policy instruments, and finally consequences on the ground in ecosystems and social systems.

However, in addition to the establishment, management and restoration of networks of functional habitats including protected areas, GI development also includes the management of the matrix in the surrounding landscape. Thus, it is also necessary to understand the consequences of actions by multiple actors and stakeholders at multiple levels, which aim at improving the managed forests surrounding protected areas both rural and urban settings [47]. Such research is inherently interdisciplinary, e.g., [48], and requires collaboration among actors and societal sectors involved with management and governance of GIs in forest, rural and urban landscapes [41,49,50,51,52].

The aim of this study is to endeavor an assessment of progress with forest biodiversity conservation by focusing on the contributions to conservation, management and restoration of representative habitat networks during a 30-year phase from the early 1990s of the current iteration of Swedish forest policy cycles. This period is characterized by the ambition to balance wood production and environmental objectives. We review the extent to which evidence-based knowledge from conservation biology and landscape ecology is applied in policy, leads to outputs, and has consequences on the ground [53]. First, concerning policy development we focus on the environmental objective slogan “Living forests”. Second, we cover the outputs in terms of evidence-based performance targets for conservation, education and hierarchical planning. Third, we concentrate on the consequences for GI functionality on the ground as a means of supporting biodiversity conservation and human well-being. This forms the base for discussing the extent to which policy objectives are met, and the prospects of maintaining forest biodiversity by zoning encompassing strict protection, nature restoration and forestry, in landscapes and regions. Taking stock of transnational effects on forest ecosystem services related to global climate change and an increasingly polarized world, we also discuss the importance of understanding both ecological and social sub-systems of GI development. Finally, stressing the urgency of adaptation and securing resilient ecosystems supporting multifunctional forest landscapes, we argue in favor of a system transition that encourages new modes of transdisciplinary integrative knowledge production and collaborative learning.

2. Methodology

2.1. Sweden as a Case Study

Sweden hosts regions with considerable forest loss due to clearing for the development of agriculture over millennia, regions with high forest management intensity, and the last remnants of the European Union’s intact forest landscapes [47]. When the international frontier of wood mining reached Sweden, the process of transforming once naturally dynamic forests, e.g., [16], into an effective wood production system, commenced. Currently, most forests are now managed by clear-felling systems with short rotations compared to naturally dynamic forests. The frontier of forest landscape transformation continues to reduce the remaining remnants of near-natural forests in Sweden [54,55,56]. Sweden has 23.6 million ha of productive forest (annual wood production > 1 m³ha⁻¹) and 4.6 million ha unproductive forest land [57]. There are five different forest ecoregions that range from broad-leaved deciduous nemoral forests and hemi-boreal forests in the south, to south and north boreal as well as sub-alpine forests in the north. These five forest ecoregions are linked to both latitudinal and altitudinal factors affecting the vegetation growing period, forest site production capacity and species distributions. The northern borders of the nemoral and hemi-boreal forest ecoregions broadly parallel the northern contiguous distribution of beech (Fagus sylvatica) and oak (Quercus robur), respectively. Together with the two boreal ecoregions further north, which are dominated by Norway spruce (Picea abies) and Scots pine (Pinus sylvestris), these four ecoregions are widely used for intensive wood production. In contrast, the sub-alpine ecoregion, confined to the Scandinavian mountain range’s eastern edge, hosts the lowest proportion of productive forest among all ecoregions and is dominated by Norway spruce and mountain birch (Betula pubescens ssp. czerepanovii). When it comes to land ownership (Figure 1), Swedish forests are mostly owned by non-industrial private forest owners (49%), private forest industry (23%), and the rest by the National Property Board, the state forest company Sveaskog Co., public bodies such as municipalities and regions, the church and forest commons (28%) (Figure 1). The human population density is high in the south and very low in the northwest (Figure 1).

2.2. Framework for Evaluation of Policy, Implementation and Consequences

Evaluating governance and management for biodiversity conservation is a crucial step for understanding the progress towards satisfying agreed policy goals in the real world. This heading’s topic matches the essence of the concept policy cycle, and how development of complex processes can be assessed (Figure 2). This requires studying the policy creation process, the implementation outputs, and the consequences on the ground [53]. Evaluation of the policy creation process involves assessment of what is good or democratic governance [8,58,59], including elements such as more and improved information management and learning, a legitimate process, and the normative aims of transparency and participation. According to Rauschmayer et al. [53], the outcomes of policy creation processes have two parts. The first part is about implementation outputs in terms of policy norms and rules to be applied by governors at multiple levels, and pronouncements of norms [60] in terms of strategic performance targets. Examples include short-term and long-term goals for the required amount of protected areas, e.g., [61], retention of fine-scale nature consideration in the managed landscape surrounding protected areas [62], as well as collaborative tactical planning and operational management approaches to enhance functionality of green infrastructures [50]. The second part is about the consequences on the ground of actual operational implementation by managers of strategic plans. This includes the progress towards a sufficient and functional network of representative forest and woodland habitats as a GI that conserves biodiversity [37], and provides human well-being [63].

This study spans a period of ca. 30 years, from the early 1990s when the current forest policy was formulated until 2023, and relies on a wide range of informants, and sources published in a diversity of contexts over a long time. This period matches the current forest policy cycle with its explicit focus on maintaining viable populations on naturally occurring species. Methodologically, we concentrate on summarizing a suite of thematically and temporally narrow studies by applying a holistic policy cycle approach (see

Table 1. The term policy cycle, e.g., [64,65,66] is a simplified description of the long-term dynamic of iterated policy creation in a particular field. This can be linked to Rauschmayer et al.’s [53] proposed series of three systematic steps to understand (1) the policy creation, (2) and implementation outputs that lead to (3) consequences on the ground in both ecological and social systems. Finally, assessment of the extent to which policy requirements are met on the ground in social-ecological systems can be made. The methodology, results and discussion in this paper about formally protected and voluntary set-aside forests and trees contributing to green infrastructure for biodiversity conservation track steps 1–3 in the extended policy cycle shown in Figure 2.

Analytic Steps		Phases in Policy Cycle (Extended Version)	Type of Science	Includes the Following Sub-Steps	Key Sources for This Review
1. Policy creation		Policy process	Social	Problem perception, agenda setting, decision-making	Bush [46]
2. Implementation		Policy implementation outputs	Social	Policy instruments and norms	Angelstam et al. [41]
3. Consequences	3.1.a. Ecosystem; Stand level	Protected area development	Natural	Outcome consequences in terms of protected area	Angelstam et al. [20,41]
	3.1.b. Ecosystem; Landscape level	Habitat network functionality/Green Infrastructure	Natural	Outcome consequences in terms of green infrastructure functionality	Angelstam et al. [20,41,54,67], Jonsson et al. [47], Svensson et al. [68]
	3.2. Social system	Operational planning processes	Social	Consequences in terms of spatial planning processes	Angelstam et al. [20,41,54,67], Eriksson and Hammer [50]
Assessment		Is policy leading to desired states and trends?	Integrative	Holistic evaluation of protected areas and matrix, and planning processes	(Comparison of step 3 with step 1 in the discussion)

). Our study can thus be characterized as participatory due to the senior author having been engaged by ministries and agencies in several assessments, which have only been published in Swedish. The second author published a comprehensive analysis of relevant policy, and the third author was engaged in several analyses of the consequences on the ground.

2.3. Policy Creation Process

This review is based on public records, reports and interviews with staff at the Swedish Forest Agency and the Swedish Environmental Protection Agency made 2010–2012 (see also [41]), and also actors in forest companies and County Administrative Boards 2019–2022. Records examined included the archives of the forestry policy review committee established to develop what became the Forestry Act of 1993, and archives of the review of the Nature Conservancy Act, as well as the histories of the earlier versions of the Forestry Act and the environmental legislation enacted in 1988 and 1991. Review of legislative records encompassed relevant ministry publications, Swedish government official reports, referral submissions from affected organizations, and government propositions presented to parliament; in the case of the 1993 Forestry Act, this also included an examination of the legislative debate in committee and in parliament. Items relating to developments in forestry policy following passage of the revised Forestry Act include materials from the Swedish Forest Agency, the Swedish Environmental Protection Agency, the Swedish Society for Nature Conservation, and applicable news reports. For methodological details, see Bush [46], Angelstam et al. [9,20,41,42]. To assess the match with evidence-based knowledge and policy, we refer to fundamental conservation biology principles for the conservation of naturally occurring species such as habitat loss thresholds [40,69], representation [70] and habitat quality [71].

2.4. Implementation Outputs

The formulation of a new forest policy triggered a long sequence of activities to translate policy to practice via strategic, tactical and operational outputs supporting tangible consequences on the ground in both social and ecological systems. We summarize the policy implementation process concerning formally protected and voluntary set-aside areas as presented in Angelstam et al. [20,41,42]. The process of implementing Swedish biodiversity conservation policy by creating protected forest areas was divided into four phases: (1) interpretation of policy content and norms for implementation in planning and practice, and the subsequent hierarchical conservation planning process in terms of (2) use of evidence-based knowledge about forest and woodland ecology and conservation biology as a base for formulation of long-term strategic quantitative targets regarding the necessary amount of protected forest areas in Sweden, (3) education and public awareness of stakeholders, (4) development of systematic conservation planning to establish functional networks of protected areas in relation to the government’s provision of financial resources.

2.5. Consequences on the Ground

2.5.1. Ecological System: Protected Area Development

We compiled data about the amount of formally protected and voluntarily set-aside areas [20,41,42,61,72,73,74]. Data are presented both for the period 1991–1997 before the interim target for formally protected and voluntarily set areas was formulated, and for the period of implementation (1998–2021).

2.5.2. Ecological System: Habitat Network Functionality

The functionality of formally protected and voluntarily set-aside areas as building blocks of functional GIs can be assessed by spatial modelling [20,70,75,76]. Functionality depends on the extent to which forest habitat patches have sufficient quality and size, and their spatial configuration allows for persistence of local populations in the short and long term (e.g., [77]). With good knowledge about the interconnectedness and functional links among species, habitats and processes for forests, e.g., [78,79], rapid assessment using estimator-surrogate data such as habitat types sensu [80] is possible, e.g., [81]. This requires (1) wall-to-wall digital spatial data of the habitats of interest, (2) knowledge about focal species’ [82] habitat requirements, which is based on the idea that conservation of specialized and area-demanding species can contribute to the protection of many other less demanding co-occurring species, e.g., [83], and (3) suitable spatial modelling algorithms [70,84].

Functional GIs are formed by both natural and anthropogenic disturbance regimes [37,85,86,87]. Natural disturbance regimes in Sweden can be divided into three broad types of natural forest dynamics, e.g., [79,88]. First, gap dynamics where regeneration of shade-tolerant trees (e.g., Norway spruce in boreal forest, and broad-leaved tree species in the hemiboreal and nemoral ecoregions) take place in small patches (i.e., gaps) created when one or a few trees disappear from the canopy because of mortality. Kuuluvainen and Aakala [89] sub-divided this category into patch dynamics driven at fine scales (<200 m²) and intermediate scales (>200 m²). Second, succession related to large-scale disturbance caused by high intensity fire, wind throw or insect outbreaks, often favoring deciduous trees in early and mid-successions. Third, cohort dynamics with partial loss of shade-intolerant trees on dry sites (e.g., Scots pine in the boreal, and oaks in the hemiboreal and nemoral ecoregions) caused by low-intensity fires. In addition, especially in southern Sweden, there are high conservation value cultural woodlands with a mosaic of forest, wooded grasslands, large trees and agricultural land [90,91,92] as a result of anthropogenic disturbance regimes [85].

Results of analyses of GI functionality can be presented in three steps, e.g., [67,75]. The first reflects the amount of habitat (e.g., combination of biotopes as for example different land cover types) for the focal species. The second step concerns the selection of all resulting habitat patches, which meet the area requirements of focal species individuals. The third step is to identify tracts with concentrations of suitable habitat that satisfy species-specific critical thresholds for the occurrence of a local population. Angelstam et al. [76] provide an overview of variables and parameter values for focal bird species listed in, for example, EU-level policies linked to biodiversity, and thus useful for assessment of GI functionality for different habitat types. For Sweden, several types of wall-to-wall land cover data have been used for assessments by practitioners, and in research [41]. The first is the k-NN dataset produced by the Swedish University of Agricultural Sciences [93]. It was derived using a combination of remote sensing of satellite images and data from the Swedish National Forest Inventory. The second is the Land Cover Data (SMD) from the National Land Survey. The SMD emits from the EU CORINE land cover program [94], and has been updated as a National Landcover Data. The third is the mapping of High Conservation Value Forests (HCVFs) [20]. A fourth approach is to expand these data using artificial intelligence; see [95].

2.5.3. Does Sweden Satisfy Its Forest Biodiversity Targets

To conserve forest biodiversity in Sweden, set-asides are made in many different ways at different spatial scales. First, trees, groups and strips of trees are left from harvesting within stands (so-called retention forestry, e.g., [62,96]). Second, some stands with high conservation values (e.g., woodland key biotopes) are voluntarily set aside, for example, in the context of forest certification schemes [97,98]. Finally, clusters of stands or entire landscapes are managed for the benefit of different species [99]. Key challenges are to measure, aggregate and assess these efforts in a landscape or an ecoregion so that it is possible to communicate the consequences of the conservation efforts at different spatial scales, i.e., tree, stand and landscape, and to different stakeholders [99]. Ideally, in addition, correction factors describing the efficiency and longevity of conservation considerations at each spatial scale in each main forest ecosystem and ecoregion should be made. Next, the total proportion of functional habitat could be compared with performance targets based on the habitat thresholds for focal species with different degrees of specialization at each spatial scale.

We summarize the approach applied by Angelstam et al. [20] to use CBD’s Aichi target #11 quantitative and qualitative criteria reflecting conservation science as a relevant normative model [100] for assessing contributions from formal protection, voluntary set-asides, unproductive forest and other attempts to establish effective area-based conservation measures to GI functionality. We compiled data about formally protected areas, voluntarily set-asides, nature consideration areas and unproductive land officially presented as potential assets to meet CBD’s Aichi target #11. To evaluate the effectiveness of these conservation instruments, we also compared these conservation instruments with respect to their size, duration, decision-making, control and method for monitoring (for details, see [20]). We also summarized the conclusions of the Swedish Forest Agency’s in-depth recent report evaluating the environmental objective “Living forests” [43].

2.5.4. Social System: Planning Processes

The operational spatial planning process to implement forest biodiversity conservation policy on the ground can be studied by qualitative methods. The examples given in this study include analyses of the content and visions of policies, and planners’ understanding, capacity, and willingness to act according to policies [101,102] using interviews [103,104]. This includes several steps. An interview manual is first developed, based on a normative model for the implementation derived from environmental and forest policies [49,52]. The normative model can be described as a translation of the policy content to an ideal approach for implementation. To identify interviewees, a bottom-up approach is used, meaning that the study included informants at the operational level of the conservation planning process [102]. The informants are then asked about their understanding, capacity, and willingness to act related to biodiversity conservation planning and collaboration among stakeholders. During interviews, the interviewees should be given full freedom to express themselves. The interviews are then transcribed and analyzed with qualitative methods with the aim that the results should be thoroughly supported by and grounded in empirical data [103,105]. A framework for data analysis would use the following steps: (1) a thorough reading and initial analysis of the data with the aim to build a general and comprehensive picture of the data; (2) an evaluation of the validity of the data including cross-evaluations with the results from the natural science analysis; (3) structuring and writing of a rough version of the results, which resulted in an unstructured text; (4) structuring the text; (5) a return to the data for comparisons, confirmation and rewriting; (6) comparison and confirmation with other scientific studies of similar subjects; (7) discussions within the group of authors. The writing and analysis process went through much iteration, back and forth between these seven points. The results are thus repeatedly scrutinized by iterative comparison with data, other research and discussions [105,106]. This approach provides empirical data to assess the level of compliance between planners’ planning processes and the normative model derived from policies [101].

3. Results

3.1. Policy Creation Process

Sweden’s current forest and environmental policies are the result of several major forces. The first is the growth of modern environmentalism and its subsequent focus on the conservation of remnants of boreal forests with high conservation value in northern Sweden, as well as cultural woodlands in southern Sweden. Nature conservation as a social movement in Sweden dates to the late 19th century but expanded greatly in reach and influence after World War II [30,107,108,109]. The Swedish Society for Nature Conservation (Swe: Svenska Naturskyddsföreningen) and the Swedish Environmental Protection Agency (Swe: Statens Naturvårdsverk) created in 1967 started to focus on the growing scientific knowledge regarding relationships between organisms and their environments, and the changes that contemporary forestry practices caused to ecosystems [110,111,112,113,114]. The conservation of habitats for endangered or threatened plant and animal species through formal area protection became a dominant theme, achieving official recognition in Swedish environmental legislation and policy. This expanded the limited nature conservation provisions added to the Forestry Act in 1974 by explicitly including the conservation of biological diversity [115,116]. Notably, this occurred at a time when environmental ideas were at the forefront of concerns among Swedish citizens [117,118]. A subsequent review of the Nature Conservation Act led to the creation of a special type of legal conservation for small biotopes [119,120]. Questions surrounding the interaction between this new form of conservation and Forestry Act regulations [121], as well as the environmental legislation, subsequently became catalysts for a full legislative review of the Forestry Act initiated in 1990, which was directed to establish a precise environmental goal for Swedish forestry [122].

The second major force was the revisions to the Forestry Act in 1948 and 1979 which increased national regulations of the activities of Swedish forest owners. The 1948 revisions were a major expansion of government influence over private forestry, designed to foster larger, more valuable harvest volumes, to maintain employment and ensure a steady supply of raw material to the forest industry [123]. The 1979 revisions added an explicit legal requirement on forest management to maintain a high and valuable wood yield, and introduced a host of new regulations including an extensive system of silvicultural subsidies and a nationwide inventory of forestland owned by individuals, all supported by a special forestland tax. That law also established rotation forestry as essentially the only permissible type of forest management, to maximize timber and pulp production and thereby support national economic goals [115,124]. The strong focus on wood production under the 1948 and 1979 revisions of Swedish forest legislation had significant negative environmental consequences. For example, only about half of the forest structures required to be protected remained on areas subject to final felling under the 1974 general nature conservation provision of the Forestry Act, which was secondary to the production goal [125]. Ultimately, the environmental effects of industrial-scale silvicultural system under the Forestry Act in this period led the Swedish Society for Nature Conservation to conclude that the forest industry was impoverishing the natural environment in the pursuit of purely economic benefits [126], and forest owners to protest against what they saw as administrative micro-management of forestry [123].

The third force was the dramatic political and economic change that occurred while the forest policy review was underway. In the middle of the committee’s work, Sweden experienced its deepest economic crisis since the Great Depression. High unemployment and a massive bank bailout caused the ouster of Social Democratic-led government in the autumn of 1991 [127,128]. The winning conservative coalition reconstituted the review committee, and directed the committee to focus on deregulation and by phasing out most of the silvicultural subsidies and the forestland tax [129]. Yet, formulation of a new environmental goal was still required. Thus, the committee had to find a way to safeguard forest productivity and biodiversity while controlling costs. Committee debate initially focused on an estimate that conservation of roughly 15 percent of productive forestland below the mountain regions would be necessary, in the absence of changes in common silvicultural practices [130]. The fiscal and practical problems with the type of approach forced the committee to consider improvements to production forestry in general that could simultaneously improve biodiversity and lower the potential cost [131,132]. Ultimately, and despite questions about the forest industry’s ability and willingness to adopt general changes, e.g., [133], as well as the possibility of success to achieve the goals set out in the committee’s directives, e.g., [134], the committee chose this approach [135].

Simultaneously, a gradual development of nature conservation policy regarding the managed forest landscape took place, e.g., [25]. In 1979, a section (§21) was added to the 1948 forestry act with the aim to implement stand-scale nature considerations in operational forest management in general. From the late 1980s, forest conservation was influenced by national and international environmental organizations, e.g., [136], the emergence of the sustainable development and sustainability policy principles, and different international agreements and conventions about forests and biodiversity, e.g., [137,138]. Regarding voluntary forest protection, the introduction of forest certification was crucial (i.e., FSC and PEFC; [139]).

After more than two decades of gradually increased societal interest in nature protection, the conservation of biodiversity, i.e., the composition, structure and function of ecosystems [140], became one of the nationally agreed forest policy objectives in Sweden [44,46,141]. From 1993, conservation and production were equal objectives of forest management in Sweden, e.g., [46]. In addition to this national policy development, Sweden has adopted several Pan-European [142,143] and EU policies and directives, such as the EU Birds, Habitat and Water Framework Directives [137,138,144], all of which include different legal obligations related to biodiversity conservation in forests.

The proposed government bill from 1990 [145], reflected a strong Swedish and Fennoscandian species-centered tradition, stating that naturally occurring species should be conserved by maintaining viable populations. This was continued with a policy addition aiming to secure the productive capacity of all forest land and to increase the protection for threatened species and different types of habitats [44]. In accordance with the principle of representation of conservation areas by ecoregions [146,147], the conservation discussion was divided in 1991 into two parts: productive mountain forests (Fjällnära skog in Swedish) and productive forest below them [148]. Moreover, it was stressed that natural functions and processes in forest ecosystems should be maintained [34]. Forest biodiversity conservation was also included into the environmental quality objectives, established by Parliament. The quality objective “Living Forests”, and its four interim targets (of which one focused on protected areas), signifies biodiversity as being important [34,35].

Following EU strategies and CBD [149], the Swedish Government’s strategy for biodiversity and ecosystem services [150] formulated both qualitative and quantitative targets for GI development as a tool to support the sustained delivery of ecosystem services supporting human well-being [63], and welfare built on nature-based outdoor recreation and tourism [151]. This strategy was the foundation for the government’s commission to a range of authorities to produce guidelines and plans for implementation of green infrastructure regional action plans at the level of county administrations [45].

To conclude, national policies concerning forest biodiversity and ecosystem services have remained intact until present time [43]. The content has been reinforced by the post-Paris agreement regulations concerning climate [152,153,154], the revised national forestry accounting plan for Sweden for 2012–2025 including a revised proposed forest reference level, EU-level strategies about biodiversity and forests [154,155], and proposed regulations on nature restoration [155].

3.2. Implementation Outputs

3.2.1. Interpretation of Policy

With clear guidelines at international, Pan-European, EU and national levels—and even within forest companies—that formulate society’s desire to conserve biodiversity in forest landscapes, formulation of tangible outcomes is indeed possible [73,74]. Thus, as Angelstam et al. [41] concluded: “the Swedish policy pronouncements evidently capture the definitions of biodiversity and conservation well. Science-based biodiversity conservation thus gradually emerged.” and “The environmental objective of the Swedish forest and environmental policy pronouncements can be interpreted as having three key words and phrases concerning biodiversity conservation. These are “all”, “naturally occurring species” and “viable populations”.

The word “all” refers to the interpretation that not only generalist species should be maintained, but also species adapted to natural and cultural disturbance regimes [41], which often have high demands on both habitat quality and area configuration [82]. This represented a so-called zero-vision for biodiversity loss compatible with EU-policy and the global 2010-target formulated in 2002 [149]. In reality, it is impossible to follow and manage all species, as for instance Sweden hosts more than 25,000 species connected to forest ecosystems. In Sweden, the national book of red listed species [156] has had a strong influence on practical conservation measures, including the development of 210 specific action plans for more than 500 species and species groups, many of which have their primary occurrence in forests. Additionally, the focal and umbrella species concepts [82,83,157] were accepted at the policy level. Hence, evidence-based knowledge of such species could be used to formulate quantitative conservation targets [73]. Empirical studies confirm that this is a useful approach [158,159], and have been validated by studies of how endangered species respond to changes in the amount of habitat [81,160]. However, there are still many knowledge gaps on the requirements of species with different life history traits and their thresholds levels for habitats, which is indicated in the EU forest strategy [154].

The term “naturally occurring species” links to the notion of representativeness, and that the Swedish forest and environmental policy does not require conservation of species introduced by humans. This means that patches and networks of protected areas and other set-asides should represent the biological variation in each ecoregion [161,162]. Hosting several types of natural forests [163] and cultural woodland regions [164], Sweden has a wide range of habitats with trees, each of which containing different species assemblages. When designing GIs for biodiversity conservation, and thus in the formulation of conservation targets, all forest and woodland systems need to be represented. There is also a temporal dimension: a long history of land use has transformed the landscape and thereby alienating landscapes from the range of natural variability, e.g., [78,165,166]. Species have adapted through natural selection to different natural processes in ecosystems (e.g., fire and flooding), and in the pre-industrial landscape to different traditional ways of managing forests and trees (e.g., grazing and mowing, pollarding). If processes are significantly altered, habitat quantity and quality will be reduced, as well as species’ population sizes and genetic diversity, potentially leading to extirpation. Hence, there is a need for detailed knowledge of various ecosystems’ ecology and historical development in different parts of the country’s ecoregions.

The term “viable populations” refers both to population ecology and population genetics [167,168,169]. One approach for establishing how much habitat is needed in the long term for the persistence of naturally occurring species is to follow the umbrella species hypothesis. This requires estimating how much forest habitat the most demanding species need in the long term, for each representative disturbance regime and development stage in each natural geographic region. Because many properties in a forest environment are dynamic, one must consider the entire landscape dynamics over time. This means that some forest habitats can exist only in certain areas in perpetuity, while others will move throughout the landscape over time depending on forest stand age and use. It is thus a spatial planning and management issue to ensure the functionality of habitat networks. Finally, while the policies on biodiversity are reasonably explicit as to the level of ambition, the spatial scale for conservation is not. Species whose individuals are small are likely to require less area than species with large body size, and operating at higher trophic levels. Additionally, should all policy targets be accomplished on all land, in every municipality, county, or in the country as a whole? This complexity allows actors with different interests and power to interpret policies differently.

3.2.2. Use of Evidence-Based Knowledge

Forest and Woodland Ecology

A foundation for formulation of strategic goals for how much area ought to be devoted to conservation requires a thorough understanding of the composition, structure and function of forest ecosystems in time and space. This includes forest landscape history, e.g., [170,171,172], the emergence of the natural disturbance regime concept in forest policy and management, e.g., [86,173,174,175], and the insight that cultural woodlands are also important habitats for forest species, e.g., [176]. These research novelties were indeed incorporated in the discussion about biodiversity conservation as the policy implementation process evolved [61]. Contrary to what may be suggested by overly simplistic indicators of biodiversity, such as forest cover [177], there are many different forest ecosystems involving a rich diversity of species, habitats and processes at different spatial scales. Thus, forest-living species representing a wide variety of adaptations must be used to study the relation between the presence and viability of populations under different levels of human-induced changes to forest ecosystems [178].

How Much Habitat Is Enough?

Swedish forest and environmental policies presently focus on the maintenance of naturally occurring viable populations. A population’s persistence in a forest landscape or region depends on how much habitat there is, whether individuals or propagules can move between different patches of suitable habitat, and how the habitat networks are maintained over time [167,168,179]. Additionally, the role of the matrix surrounding habitat patches aimed at focusing on conservation needs to be understood. While the term biotope refers to an environmentally uniform landcover, a habitat is defined by the properties that define the requirement of a species or a population [180]. Thus, a habitat often consists of several biotopes, such as for feeding, cover and breeding. Therefore, there is a need to identify and assess the quality of biotopes that form habitats. In addition, factors other than biotopes mapped as land cover types are parts of the habitat of a species, such as predators and competitors, as well as micro- and macroclimate [15,181]. The combination of decreasing amounts of habitat, which decreases the number of individuals that can be supported, and increased fragmentation, which makes it harder for individuals to move about in the landscape, are the most common reasons why species disappear locally and regionally, and finally completely.

There is both theoretical and empirical evidence for the existence of thresholds for extirpation of a population as the amount of available habitat is reduced [182,183,184,185]. The threshold refers to the fact that the risk for population extinction shifts from low to high within a limited range of further loss of habitat. The fact that there are limits to how much of different forest habitats may disappear without threatening the viability of populations of all naturally occurring species forms the basis for the formulation of long-term goals for how much of different forest habitats are needed, e.g., [19]. There is a clear parallel to the concept of critical load, which addresses the question of how much deposition of, for example, nitrogen and sulfur ecosystems can tolerate [186]. Studies attempting to answer the question, how much of different forest habitats are necessary for species persistence, are of two different types.

The first type of studies addresses the proportion of a local landscape that must consist of biotopes that form suitable habitat. An example is the statement that a species would need at least 30% of old Norway spruce forest in a local landscape to be present as a local population, no matter how much old spruce forest once existed in the natural landscape [187,188]. Systematic studies of how habitat loss affects species with different requirements can be used to formulate performance targets as to how much habitat species require [158]. Angelstam et al. [189] proposed the following steps: (1) stratify the forests into broad cover types as a function of their natural, or anthropogenic, disturbance regimes; (2) describe the historical spread of different management impacts in the respective ecoregions that moved the system away from forest naturalness or cultural landscape authenticity, e.g., [85,190]; (3) identify appropriate response variables (e.g., focal species, functional groups or ecosystem processes) that are affected by habitat loss and fragmentation; (4) for each forest type identified in step 1, combine steps 2 and 3 to look for the presence of non-linear responses and identify intervals of risk and uncertainty; (5) identify the “currencies” (i.e., species, habitats, and processes) which are both relevant and possible to communicate to stakeholders.

Empirical research shows that there is large variation in terms of what different species require of habitats depending on the life-history traits, and scale and ambition of the conservation work, e.g., [69,76,178]. For several specialized forest species, the presence of thresholds has been documented for the necessary amount of habitat at the landscape level [76,159,188]. For 17 species (birds, mammals and insects), the proportion of habitat needed was 10–50% with a mean of 19% [76]. Svancara et al. [69] reviewed evidence-based knowledge and norms agreed in policy processes about the area proportion needed for conservation. On average, the proportion of area recommended based on evidence-based studies in terms of conservation assessments (31%) and threshold analyses (42%) was almost three times as high as those recommended in policy-driven processes (13%). This is consistent with previous findings that 10–30% of a species’ habitat is needed to maintain viable populations in a landscape [182,183,185].

The second type of studies providing knowledge relevant for evaluation of how species are affected by different amounts of resources involve comparative studies along forest history gradients [9,191,192,193,194]. Such studies focus on how much of various resources are needed compared to the range of variation in naturally dynamic reference landscapes regarding dead wood, e.g., [195,196,197,198,199], the proportion of deciduous trees in stands [200,201], and naturally dynamic forest old growth forest stands found in the managed landscape [17]. Research on how much of different characteristic habitat properties are needed in managed landscapes to maintain species dependent on natural forest properties indicates that at least 10–40% of the natural amount habitat needs to be maintained [76,196,202]. As this level is much higher than what is left in today’s Swedish forest landscapes [195,196,203,204,205], specialized species are threatened [156,206].

Summarizing, to answer the key question “How much habitat is enough?”, evidence-based knowledge is needed about (1) the composition, structure and function of the pre-industrial characteristics of forest landscapes and cultural woodland, e.g., [61,207,208], and (2) how much loss of that can be accepted. Hanski’s [40] rule of thumb “a third of a third” is appropriate and is now being matched by evidence-based policy targets allocating 10% strict protection plus 20% aimed at conservation management and nature restoration, e.g., [154,209].

3.2.3. Education and Public Awareness of Stakeholders

To communicate the emerging evidence-based knowledge about forest and woodland ecology, cultural woodlands and conservation biology, Sweden’s National Board of Forestry arranged several educational programs where nature conservation and sustainable forest management were important parts, i.e., Richer Forests, Cultural Heritage in the Forest, and Greener Forests [210,211,212]. Additionally, green forest management plans with specific focus on maintaining habitat for species appeared [213]. Several efforts have aimed at developing ways to contribute to a landscape perspective for biodiversity conservation (Landscape Ecological Core Areas (Swe: Landskapsekologiska kärnområden (LEKO)) [214]; inspired by the Finnish METSO programme, the KOMET programme [215] aims at providing complementary methods for nature protection (Swe: kompletterande metoder för skydd av natur) [216,217]; Regional Landscape Strategies (Swe: Regionala Landskapsstrategier) [218,219]. Since 2002, the Swedish Forest Agency has worked with first a national, and then also regional and local forest sector councils as an attempt to develop an interface towards and establish collaboration with the main forest sector stakeholders. This effort has had varied success, and in many places the local level has been omitted. Experiences from the national level show that stakeholder collaboration is not an easy task [220]. The main tools used for implementation of forest and environmental policies in Sweden are counseling and education. This informational approach to policy implementation has been shown to influence the behavior among forest owners in the short term but, might not affect underlying values and preferences [221]. In contrast, in a study where 25 forest and conservation planners were interviewed in central Sweden [41], some forest planners described the acceptance of the new forest policy in the early 1990s as a long process. Knowledge and gender are linked to the attitude toward conservation. For example, Uliczka et al. [222] showed that self-estimated knowledge about conservation and knowledge about forest species were all related to a positive attitude towards conservation. Thus, education can affect conservation consequences.

3.2.4. Systematic Conservation Planning

The Emergence of Hierarchical Planning

The introduction of the woodland key habitat concept [223,224] and a corresponding nation-wide mapping of biotopes with high conservation value, and substantially increased resources for protection of forest areas with high natural values for conservation purposes during the 1990s, created a potential foundation for spatial planning of voluntary set-aside areas [20]. However, conservation of viable populations requires sufficient amounts of suitable habitat configured to form functional networks [20,54,225,226]. In parallel, the first ideas about landscape ecological planning regarding forest conservation emerged [67,227]. As a result, the policy implementation process to conserve biological diversity gradually became hierarchical with strategic, tactical and operational planning in several steps similar to forest management planning [228].

Strategic Planning: Regional Gap Analysis

The purpose of a regional gap analysis is to estimate how much of different habitats remain in different regions compared to the pre-industrial amount and distribution [161,229,230,231]. Focusing on the role of protected areas for forest biodiversity conservation, Zackrisson et al. [130] and Liljelund et al. [232] pioneered attempts to formulate area targets for forest protection in Sweden. Nilsson and Götmark [146] conducted analyses of representation of protected areas for different types of land cover, and found that productive sites were underrepresented. SOU [73,74], summarized by Angelstam and Andersson [61] and Angelstam et al. [41], took the gap analyses concept a step further by also estimating the extent to which there were gaps in the amount of habitat to maintain viable populations of naturally occurring species in each of Sweden’s main ecoregions. A short ABC for a quantitative gap analysis [41,61,233] includes several steps (

Table 2. Summary of variables associated with quantitative regional gap analyses concerning the proportion of a forest habitat or attribute that needs to be conserved (including protection, management and restoration) to maintain viable populations of naturally occurring species in an ecoregion; see [41,61,233].

Variable	Description
A	The amount of a particular forest environment which species have adapted to in the region a
B	Today’s amount
B/A	Representation
C	Performance target or norm based on knowledge about the proportion out of the area of a particular natural forest environment required for retaining a viable population
A × C	Long- term target for the amount of a particular forest environment
B–(A × C)	Gap (if the value is negative)

a—in naturally dynamic boreal forest landscapes, e.g., [16], or traditional cultural landscape, e.g., [166].

):

(1) To estimate the pre-industrial area of the different representative forest habitats in a particular region (A). (2) To compare (A) with estimates of the current quantities of the same forest habitats (B), makes it is possible to estimate how representative different habitats are today (i.e., B/A or representation). (3) With knowledge about what proportion of a particular naturally occurring forest environment that is required for the most demanding species, i.e., focal or umbrella species, to maintain a viable population (C), one can estimate how much of different representative forest types need to be maintained to secure viable populations of all species. The quantitative gap analysis is thus based on the difference between B and A × C. A negative value indicates a gap in habitat area, and hence the need of restoration and re-creation of habitats [234,235]. However, the presence of habitat may still not lead to re-colonization of species. Species with poor dispersal ability may thus remain only as relicts of past landscapes doomed to extirpation, because they are unable to colonize isolated areas [236,237].

The existence of non-linear responses of species to habitat is central for the opportunity to formulate evidence-based norms for the conservation of biological diversity, e.g., [182,183,184,185,238]. By incorporating contemporary knowledge about forest ecology, forest history and conservation biology, SOU [73] concluded that in the long term (~50 years) 8–16% of forest landscapes, depending on ecoregion, should consist of functional GIs [61,74]; see

Table 3. Summary of results of the quantitative gap analysis concerning productive forests below the mountain forest region in Sweden [74]. Using a general threshold value of 20% as a target for the necessary proportion of remaining habitat in the long term (I), the following steps were taken: individual assessment of 12 natural forest and 2 cultural woodland types according to their expected occurrence in the different ecoregions and (II) assessment of which of these forest types managed landscapes can deliver. The remainder (III) became the long-term target for set-aside of forests to maintain viable populations of naturally occurring species. This long-term target is satisfied by summing up (IV) the already protected area in 1997, taking into account (V) the nature values created by nature consideration and landscape planning in regular forest management, setting aside (VI) forests and woodlands with high nature values that were not protected, (VII) including the area of wooded grasslands of the cultural landscape, and finally (VIII) restore habitat by nature conservation management.

Item	Description	Average Proportion and Regional Variation (in Brackets) of Productive Forests below the Mountain Forest Region in % of 218,800 km2
I	Threshold rule of thumb based on empirical studies of species’ requirements (C in %; see Table 2)	≈20
II	Forest environments without needs for forest protection (%) (PG *)	10 (4–12)
III	Long-term goal (%) with sub-components IV–XIII below	10 (8–16)
IV	Formally protected area 1997 (%)	0.8 (0.4–1.6)
V	Reduction of the need for forest protection due to functional nature considerations at the stand level (%) (PF/K *)	0.9 (0.3–1.7)
VI	Short-term goals defined by existing unprotected forests with high conservation value (%) (NS and NO *)	3.2 (1.9–3.5)
VII	Wooded grasslands in cultural landscape (%)	0.8 (0–2.2)
VIII	Restoration needs (%) (PF/K *)	≈4 (3–11)

* the codes PG refers to wood production with general nature considerations, PF to production with reinforced natural consideration or K (combined goals), NS to nature conservation management.

. Subsequently, a short-term interim target was formulated by the government, stating that by the end of 2010 the amount of formally protected and voluntarily set-aside forests should increase by 400,000 and 500,000 ha, respectively [34,35]. These 900,000 ha correspond to 4.1%-units of productive forests at the national level.

It must also be noted that for many species, habitat is indeed maintained in managed forest landscapes even with conventional sustained-yield forest management systems. For example, area-demanding species such as Moose (Alces alces), brown bear (Ursos arctos) and capercaillie (Tetrao urogallus) [15,181,188,239], which have been extirpated in many parts of Europe, may thrive in managed forests. Thus, the estimated need to set aside forests to conserve viable populations was lower than the 20% rule of thumb [73,74], and varied among forest regions due to differences in the composition of forest environments and their dynamics in relation to how they are managed. Since clear-felling with tree retention is the norm in Sweden [13,54], and forests with internal dynamics (such as broadleaf or wet spruce forests), and cohort dynamic forests (i.e., multi-layered oak and Scots pine forests), and cultural woodlands, are more common in southern than in northern Sweden, the estimated need for forest protection was higher in southern (16%) than in northern Sweden (9–12%) (Table 3). Note that these estimates rested on the assumption that environmental considerations in managed forests reached the expected targets for tree retention in stands, that the network of protected areas was fully functional, i.e., composed of biotope patches with sufficient quality and connectivity, and that the general considerations are coordinated with all formally protected areas and voluntary set-asides.

Much has happened since the emergence of the contemporary forest policy in 1993 and the gap analysis from 1997. The evaluation of forest policy [240], a revised forest policy [141] “En skogspolitik i takt med tiden”), the addition of a 16th environmental quality objective, A Rich Diversity of Plant and Animal Life, and the new challenges of climate change and increased globalization are a few examples. To date, three audits of the 1997 regional gap analysis have been made. First, on 11 February 2004 the National Board of Forestry in Kristianstad organized a hearing on “Gap Analysis of Nemoral Forest”. Second, the scientific background to quantify goals for maintenance of viable populations was the subject to a report [241], and a conference at the Royal Academy of Forestry and Agriculture on 21 March 2006. The conclusions from these revisions of the scientific background were that the regional gap analysis approach was a robust strategic planning tool, that new knowledge about species requirements showed that they were rather higher than lower compared to the Environmental Advisory Council estimates from 1997, and that there was a need to monitor and evaluate the results of investment in biodiversity conservation continuously. Finally, a review of the interim target Living Forests [242] did not change the conclusion about the required amount of protected areas, but pointed out that landscape ecological planning and collaboration among land managers need to be improved for formally protected and voluntarily set-aside areas to form functional green infrastructures for forest biodiversity conservation [42].

According to the Swedish Forest Agency’s [43] in-depth evaluation of reach Living Forests, they will not be reached. The five most important problems to solve are:

• Decline and lack of important habitats in the forest landscape, and several types of habitats are becoming increasingly fragmented.
• Unfavorable status or negative development for many forest-dwelling species. Many threatened and sensitive species are declining, and populations are becoming increasingly fragmented.
• Several of the forest ecosystem services have insufficient status.
• Cultural heritage remains are destroyed in the forest landscape due to forestry measures.
• Negative impact on watercourses of the forest landscape.

Tactical Spatial Planning

The next step was to optimize functionality of forest habitat networks at the county level, e.g., [243] (see methods section for an approach to assess functionality of forest habitat networks). To aid this planning process, a national compilation of high conservation value forests [20,244] and analysis of the location of core areas for forest protection [245] were conducted by each county in Sweden. While the regional analysis in the previous strategic planning step only distinguished four broad forest regions, the tactical analysis was spatially explicit to match the resolution of individual protected areas across Sweden. The spatial planning strategy pronounced how protected area candidates should be selected for formal protection. The guiding principle for selection was the conservation value of an area, including structure and species composition of the forest itself, as well as its connectivity in the local landscape context in terms of distance to other high value forests. Additional criteria for formal protection were recreation and cultural heritage. Finally, the extent to which the protection was practical was considered. The need for dialogue with forest landowners was also stressed as an important component. Subsequently, the County Administrative boards and the Swedish Forest Agency formulated regional county-level strategies, including spatial analyses.

Forest companies have embarked on spatial planning by developing landscape plans [227]. For example, in 2003 the state forest company Sveaskog Co. developed the Ekopark concept [99]. This involved an approach for identifying forest landscapes of different types that should be devoted to the maintenance of viable populations of species based on sufficient habitat qualities across three spatial scales, i.e., tree, stand and landscape [99]. Today, there are 37 Ekoparks covering 2.5% of Sveaskog Co.’s holding, and with an average size of 6500 ha [246]. They have a much higher proportion of old forest (52% > 100 years old) compared to 15% outside the Ekoparks.

3.3. Consequences on the Ground

3.3.1. Ecological System: Protected Area Development

Productive Lowland Forests

Angelstam et al. [41,42] summarized the development of formally protected and voluntarily set-aside forests for the period 1991–2010. The first review of formally protected areas showed that about 0.5% of the productive forests below the mountain forest region was formally protected in 1991 [72]. By 1997, a total of 0.8% (174,000 ha) of the productive forest was formally protected [74]. By the end of 2008, the formal forest protection figures had reached 1.1% (244,500 ha). This corresponded to 61% of the interim target for formal protection to be reached by the end of 2010. As described in Angelstam et al. [41], the government promised to transfer up to 100,000 ha productive forest land from Sveaskog Co. to the state to speed up the process of reaching the interim target in time. This forest was to be used as a pool for forest land replacement when creating protected areas on privately owned land [247]. The most recent data add up to 11% of productive forests, and are equally distributed between formally protected and voluntarily set-aside forests.

Regarding voluntarily protected areas, these are less precise than the formally protected areas. A survey of woodland key habitats began in the early 1990s [223], and voluntary set-aside of forest commenced in the early 1990s. In 1998, the total area of voluntarily protected forests with conservation values was estimated at 230,000 ha below the mountain forest region [248,249], and in 2008 the Swedish Forest Agency [6,250] reported that this number had increased to about 936,000 ha. However, [7,250] estimated that about 75% of the voluntary set-asides had significant nature conservation values. The interim target of 500,000 ha voluntarily set-aside forest formulated after the regional gap analyses made in 1997 was thus probably reached by the end of 2010. The increase in formal protection and voluntary set-aside for the period 1909–2021 is summarized in Figure 3, and includes transfers from voluntary set-asides to formal area protection around 2015.

The Mountain Forest—The EU’s Last Intact Forest Landscapes

The sub-alpine mountain forest region’s forests and woodlands along the Scandinavian mountain range covers ca. 3.5 million ha, of which 1.5 million ha counts as productive [68,148]. According to Naturvårdsverket [72], 265,000 ha was protected as Domänreservat in 1991 (i.e., state forest company protected areas). Additionally, there were 325,000 ha nature reserves and national parks, thus amounting to a total of 590,000 ha (38%) with formal protection. According to SOU [74] and Naturvårdsverket [252], a total of about 660,000 ha (~43%) of the mountain forests was formally protected in 1997 [253]. Currently, 57% of the mountain forest region’s productive forest is formally protected for conservation purposes ([251], Figure 4). Regarding voluntarily set-aside productive forest in the mountain forest region, Skogsstyrelsen [6,250] reported 197,000 (13%). The corresponding figures below the mountain forests are 3–5% and 6%, respectively.

The last large intact forest landscapes along the Scandinavian Mountain range in Sweden offer unique opportunities for conservation of biodiversity, viable populations and ecological integrity and resilience in the European Union ([47], Figure 4). Additionally, with a European perspective, the forests along the Scandinavian Mountains and the Ural Mountains, which run north–south, offer better conditions for species to survive the stress of climate change [254] than forest species in the Carpathian Mountains, the Alps and the Pyrenees, which run east–west. However, these last large intact forest landscapes are at a crossroad between intensified wood production aimed at bio-economy, and rural development based on multi-functional and resilient forest landscapes for future-oriented forest value chains [68].

Indeed, policy regulations have been successful in limiting forest harvesting since the beginning of the 1990s. However, like other unique natural forest remnants such as in the Bialowieza Forest in Poland [255], the Swedish mountain forests remain as a battleground. Key issues regard intensification of forest use and logging of forests that have never been subject to clear-felling systems, vs. reindeer husbandry, conservation of biodiversity and wilderness as foundations for rural development based on value chains other than the forest industry’s [68].

3.3.2. Ecological System: Habitat Network Functionality

Spatial Differences

The historical and presently ongoing fragmentation and loss of natural forests and cultural woodlands implies that not all habitat patches satisfy criteria in terms of size, quality and connectivity, and will thus not form functional GIs [236,256,257,258]. Assessments of habitat network functionality over entire counties and regions confirm this [20,54]. In collaboration with the County Administrative Boards of Dalarna and Gävleborg, Angelstam et al. [67] conducted analyses of functional connectivity for focal species representing different forest and woodland habitat types, and found that regional gap analysis overestimated the area of functional habitat area. Using the same approach, covering nine counties in south-central Sweden, Angelstam et al. [41] applied spatial modelling to assess the functionality of three different forest habitats (old pine, old spruce, old deciduous) and one type of cultural woodland (forest-farmland edge). The analysis showed that on average 15% of all presently existing land with these land covers formed functional habitat networks (Figure 5). However, there were significant regional differences among the four forest habitats in the different boreal ecoregions depending to the history of forest use. This study did not assess the landscape level amounts needed to reach different conservation ambition levels, but only the extent to which presently existing biotopes formed functional habitat networks in landscapes. Later the same kind of analyses were conducted for all of Sweden. While functional connectivity was very good in the mountain forest regions, the situation deteriorated from northern to southern ecoregions. This clearly illustrates the reduced functionality of GI due to scattered and small remaining high conservation value areas (Figure 5).

Temporal Trends

Few retrospective studies have compared the temporal consequences of loss vs. protection of high conservation value forests. Focusing on the two large counties Dalarna and Jämtland representing the characteristic expansion of the timber frontier in northern Sweden, Angelstam and Manton [54] showed that formal forest protection grew rapidly in the two counties from 1968 to 2020, and reached only 4% of productive forests. In contrast, from 2000 to 2019, habitat network functionality for old Scots pine declined by 15–41%, and for old Norway spruce by 15–88%.

3.3.3. Does Sweden Reach Forest-Related Environmental Quality Objectives?

The first evaluation of the implementation of the 900,000 ha interim target area for forest protection until 2010 [259] concluded that it would be difficult to reach this target by the end of 2010. Hence, Miljömålsrådet [260] stressed the need for intensified activities to reach this area target. Subsequently, Statskontoret [261] proposed that the government-owned Sveaskog Co. should offer compensation areas for productive forestland with identified conservation values on land belonging to industrial forest owners. Political pressure to speed up the area protection process prior to the Swedish parliament elections in autumn 2010 forced some county administrative boards to focus on establishing protected areas by purchasing the land designated for land exchange with the Sveaskog state forest company to reach the interim area target. As pointed out by Angelstam et al. [41], this exemplifies “how economic and political circumstances may overthrow a well elaborated planning process”. As a result, because Swedish state forests are biased towards less productive forest types such as dominated by Scots pine, representativeness and functionality of habitats are reduced compared to if the tactical planning approach had been pursued.

The size, duration, decision-making, control and method for monitoring of the formal and voluntary conservation instruments, as well as unproductive forests, mean that they differ in conservation effectiveness (

Table 4. Basic information about four groups of conservation instruments in Sweden [20].

	(i.i) Formal; According to the Environmental Code		(i.ii) Formal; According to the Land Code	(ii) Voluntary Set-Asides	(iii) Nature Considerations (§ 30, Forestry Law)	(iv) Unproductive (Wood Production <1 m3ha−1yr−1) (§ 13a, Forestry Law)
Aim	National park, nature reserve: Conserve and develop nature of high value for plants, animals and people	Biotope protection: Conserve terrestrial or aquatic habitat for threatened species	Conservation agreement: Conserve and develop qualities for biodiversity	A complement to formal protection	Consideration to biodiversity conservation in managed forest	Wood harvest not recommended
Establishment	1909 and 1964, respectively	1998	1993	1995	1979	1979
Target size	Usually >20 ha	Usually <20 ha	Variable	>0.5 ha	<0.5 ha	>0.1 ha
Duration	Permanent	Permanent	Variable	Unknown	Unknown	Permanent
Decision by	Parliament, Government, County, Municipality	Forest Agency, Municipality	Agreement between the State or Municipality and owner	Landowner	Parliament, Government, Forest Agency	Parliament, Government
Control	County	Forest Agency, Municipality	State	Forest certification	Forest Agency	Forest Agency
Monitoring	Georeferenced GIS polygons	Georeferenced GIS polygons	Georeferenced GIS polygons	GIS data and questionnaires	Random field sampling	National Forest Inventory

). Angelstam et al. [20] showed that there was a clear decline in the patch size and duration of different conservation instruments from formally protected areas (>20 ha and permanent) via voluntarily set-asides to nature consideration areas (<ca. 0.5 ha and unknown).

In Sweden, forests not used for wood production currently cover 26% of all forest land (28 million ha). Excluding current nature consideration areas, the figure is 24%. To assess the extent to which HCVF patches actually contribute to GI functionality, Angelstam et al. [20] presented spatial analyses adjusted for the lower biodiversity value of unproductive forest, which suggested a 50% reduction to 12%. Of this, 6%-units of forest land was formally protected, 3% voluntary set-aside, and 3% unproductive. In contrast, in the sub-alpine forest ecoregions 72% of the total forest area potentially contributed to functional GI, of which 54%-units contributed to Aichi target #11, and of which 44% was formally protected. For the four other forest regions, in which the focus is on production of industry raw material, the corresponding numbers were 14–23%, 3–8% and 1–3%, respectively, of all forests.

Angelstam et al. [20] concluded that there are two key aspects of the distribution of the four types of set-asides listed in Table 4 as components of forest GI in Sweden. First, there is a large difference in the GI functionality of the sub-alpine forests being dominated by unproductive forests compared with the other four forest ecoregions in which the focus is on high sustained yield forestry. Second, there is a considerable difference between the total area of different set-aside types and the estimated area of functional GI (Figure 6) and the lower biodiversity conservation value of unproductive forest.

3.3.4. Social System: Operational Planning Processes

In Sweden, forest areas withdrawn from wood and biomass production can be divided into areas formally protected by law (national parks, nature reserves, biotope protection areas and conservation agreement), and voluntarily set-aside areas (Table 4). To conserve, manage and restore functional network of forest habitats in a district, county, or other geographical area requires collaboration between stakeholders, including landowners, government agencies and others representing different interests and uses. Statskontoret [261] presented results from interviews with county administration staff, who reported that collaboration related to protected area designation was only rarely a problem. This is corroborated by the fact that only 2% of the proposed protected areas led to disagreements and court cases. However, the counties felt limited by the staff available, access to forest land to compensate forest owners’ loss of productive land, taxation rules and, above all, funding to compensate landowners. To facilitate the implementation of the 16th environmental quality objective on biodiversity (A Rich Diversity of Plant and Animal Life), landscape planning of protected areas was encouraged in terms of a pilot project commissioned by the Government in 2005. The objective was to develop regionally adapted landscape strategies, i.e., working arrangements and planning processes for conservation and sustainable use of natural resources using a holistic and cross-cutting perspective [262]. A total of seven pilot areas were included, ranging from mountain to regular managed forests, as well as in urban and rural areas. The case studies also represented different phases in the development of collaboration. As a result, a handbook was produced [263], and Jonegård [219] summarized the Swedish Forest Agency’s experiences.

Few studies have systematically assessed the extent to which planning processes succeed with spatial conservation planning across different forest management units and forest owners’ holdings, and different spatial scales. The Swedish model for biodiversity conservation is built on a shared responsibility among landowners, the forest industry and the government, and the principle of each sector’s responsibility for the environment (see [145]). However, available knowledge on which to base future conservation decisions is not as comprehensive as the information used for decisions related to timber production. In a study of large forest companies in Sweden, Eriksson and Hammer [50] noted gaps in terms of absence of information about habitat connectivity at the landscape and smaller scales, and the effectiveness of protected areas for conservation. Similarly, as indicated by results from interviews made with 25 forest and conservation planners in central Sweden, Angelstam et al. [41] could not trace this shared responsibility at the landscape or regional level. The collaboration focused on the object or at a stand level, i.e., with the aim of identifying red-listed species, or certain biotopes. There was no general collaboration among forest owners, or with government agencies and forest owners, aiming to create functional forest habitat networks that cross ownership borders. Thus, the regional conservation strategies were never brought down to the ground, which means that they are not used by forest planners to assist a landscape perspective in their planning. In fact, regional administrations claimed that their responsibility is only for the protected areas, and not the whole territory. There were neither efforts to involve the public in collaborative learning processes, nor to develop socially robust solutions for conservation or to develop a common knowledge base among different stakeholder groups. Similarly, studies in Poland [49] and Lithuania [51] report no or low levels of collaboration among sectors and levels of governance. Reviewing other cases of natural resource planning processes, Sandström et al. [264] concluded that the government is still more important than governance (see also [265]).

4. Discussion

4.1. Assessment of Trends in Policy, Implementation Outputs and Consequences

4.1.1. Overall Patterns since the 1990s

Overall, policy concerning environmental dimensions of forests shows remarkable stability over time (

Table 5. Overview of the long-term temporal development of policy, instruments for implementation, and outcomes supporting the Swedish environmental objective in social-ecological systems on the ground.

Analytic Steps	Sub-Categories				Items				Overall Comments
		1990		2000	2010		2020
Policy	International		SFM		CBD Aichi	Paris climate		CBD 2022	Stable evidence-based policy
			policy
	EU					GI	climate +	nature	Increasing role of EU under the
						policy	biodiv.	restorat.	“The Green Deal”
	National		Forest policy with			GI			Stable evidence-based policy
			environmental objective			policy
	Voluntary		FSC						Stable negotiated policy
Instruments	Carrot			Funding Protected Areas					Gradual increase of PAs,
									depending on politics
	Stick								NA
	Sermon		Gap analysis		Gap analysis				Sustained science-policy
				Conservation planning		GI plans	GI plans		interface
Outcomes	Protected areas	Formal protection							Gradual increase of PAs,
			Voluntary set-asides						depending on politics
	GI						Net loss		Forestry intensification
									over-rides effects of PAs
	Collaboration								Abundant, but often not
									evidence-based
	Planning								Functional on public land and
					Landscape planning				some industry; otherwise not
		1990		2000	2010		2020

). However, while public policy clearly is built on evidence-based knowledge about conservation biology and landscape ecology at international, EU and national levels, voluntary policy such as forest certification is not [266,267,268].

Policy instruments can be divided into the three categories “carrot, stick and sermon” [269]. The dominant type of tools for supporting the development of habitat networks as functional GIs has been carrot in terms of the state purchasing land to create formally protected areas, and sermon in terms of education campaigns and evidence-based analyses of high conservation value forests and their spatial configuration.

Outcomes in terms of increased number and area of formally protected areas took off in the early 1990s, and have continued to increase at rates that mirror the amount of state funding made available. However, rapid loss of remnants of high conservation value forests outweighs the increase of high-quality protected areas [20,54]. Platforms for stakeholder collaboration are widespread at national and regional levels of governance [270,271], but are generally neither linked to concrete landscapes on the ground, nor based on principles of evidence-based systematic conservation planning.

4.1.2. Policy about Forest Biodiversity Conservation Remains Intact

In Sweden, forest biodiversity conservation is dealt with in both forest and environmental policies. The opening paragraph of the current Forestry Act from 1979 [272] and its revisions establish production and environment as equally important objectives [141,273,274,275]. The system of regulations and subsidies under the previous law were abolished, and replaced with minimal regulations and high expectations for voluntary conservation by all actors within the forestry sector to achieve the goals of the new law [276]. This has been described as “freedom under responsibility” [115,277]. In exchange for greater freedom of forest management, forestland owners and the forest industry are expected to share a collective responsibility to voluntarily ensure that both production and environment objectives are met. This is sometimes referred to as the Swedish model for conservation, and the Swedish forestry model, as described by the Royal Swedish Academy of Agriculture and Forestry [278]. The aim was to foster forest management systems that are adapted to different site conditions able to both conserve and enhance biodiversity, and maintain and develop the productive capacity of forestland, reflecting recent scientific knowledge about the importance of species and ecosystem variation [44].

Major elements of the previous law that the latest Forestry Act retained include the reporting requirement for final harvests, regeneration regulations, and a prohibition on the conversion of existing deciduous forests comprised of species of significant ecological value [273]. The main role of the Swedish Forest Agency has shifted from a focus on regulatory oversight and administration of silvicultural subsidies to emphasis on national inventories of ecologically valuable habitats for formal and voluntary protection under the new policy, as well as policy implementation by providing information, education and advice to landowners and others in the forestry sector [210,211,221,223,279,280,281].

Additionally, the legal protection for small habitat patches, now encompassed within the revised Environmental Code of 1998 [282], remains in effect by the Swedish Forest Agency. The formal policies are also now supplemented by voluntary forest certification systems that have increased their area coverage significantly since the passage of the current Forestry Act [283,284,285], and with a combined cover a total of 19.2 million ha of forestland [286,287].

With its extensive reliance on voluntary contributions for biodiversity conservation and limited regulation, the new forest policy has been described as depending on “voluntary action as control”. This implies a system built on self-responsibility and relying on non-compulsory but nevertheless standardized sector-wide measures such as guidelines, evaluations, and certification systems to achieve the environmental objective. This involves unresolved tensions between the production and environmental objectives [288]. Despite these concerns, in the latest of the periodic policy reviews, the Swedish government reaffirmed the major elements of the current forestry policy, including the emphasis on voluntary action to achieve nature conservation goals [141]. However, recent reports from the Swedish Forest Agency show that a significant share of final felling areas do not meet the minimum requirements of the Forestry Act [289,290], as well as reports of harvests in old-growth and other high-conservation forests [291]. This has fueled criticism of the forest policy as being too heavily reliant on voluntary contributions to reach the goals of the Swedish environmental and forest policy, e.g., [292,293]. In recognition of these problems, the Swedish Forest Agency has recently indicated that it plans to issue more legal orders prohibiting final harvests entirely, or requiring more detailed and extensive conservation than affected landowners had planned. The intent is to force landowners to oppose the agency in court in order to clarify the legal praxis surrounding the agency’s interpretation of the conservation regulations established under the Forestry Act [290,294]. This would then provide guidance about how high a level of nature conservation that the Forestry Act actually requires landowners to accept, which was a central issue at the time the policy was created [295].

At the same time as the most recent forest policy revision has retained the balanced production and environmental goals [141], it states that raw material production should increase using more intensive methods [296]. Poudel et al. (2012) estimated that intensive forestry may increase forest production by up to 26% and annual harvest by up to 19%. The campaign “skogsriket” (English: “the forest kingdom”) initiated by the ministry of rural development aimed at producing more raw materials in the forest industry and increasing the net value of forest product exports [275]. However, in the context of implementing the current forest and environmental policies, Strengbom et al. [297] concluded that intensively managed forests will “only harbor species that are common and widespread in conventionally managed stands and that species of conservation interest will be lacking, due to the low heterogeneity and light intensity of even-aged monocultures with dense canopies, short rotation times and low availability of coarse woody debris”. In particular, the effects of management strategies for increased biomass production on soil resources, specialized species and water quality at landscape scales are inadequately understood [297,298]. A decade later, Felton et al. [14] found that these issues remained unsolved. In a similar study, Angelstam et al. [9] concluded that a strong forest management cropping system tradition can be a burden for reaching sustainable forest management objectives.

In addition, global factors are affecting Swedish forests and forestry. Beland Lindahl and Westholm [299] found that four areas stand out as particularly important: changing energy systems, emerging international climate policies, changing governance systems, and shifting global land use systems. Both domestic challenges of biodiversity conservation and rural development, and global challenges, will continue to be important for future Swedish forests and forestry. Hence, they concluded that the forest sector “must be disembedded and approached as an open system in interplay with other systems”. This calls for integrated approaches to natural resource governance, planning and management, e.g., [9].

Summarizing, in parallel with the long delivery time for creation and restoration of ecological dimensions such as old trees, stands with several tree generations and decayed dead wood, a long-term perspective on the different phases of forest use as in successive long-term coarse policy cycles is appropriate (e.g., “societal contracts” sensu [9,300]). This study focuses on what can be characterized as a third phase in the development of Swedish policy cycles (

Table 6. Current, past and emerging policy cycles in Sweden covering the time span for developing high conservation value forest ecosystems (i.e., >200–300 years [16]). This study focuses on Phase 3.0 and the emerging 4.0 (see Table 5).

Phase in the Evolution of Forest Policy	Approximate Time Period	Short Description
Phase 1.0	Medieval to industrial revolution	Livelihoods based on multiple use of landscapes in traditional village systems
Phase 2.0	ca. 1830–1970s	Even-aged sustained yield forest management for industrial raw material in three steps:
Phase 2.1	Mid-1800s	- Sustained yield of wood for charcoal emerged regionally in mining and metallurgy
Phase 2.2	1850s to 1903	- North Sweden is reached by successive frontiers of “wood mining”, which triggered development of forest policy
Phase 2.3	1903 and 1947 forest laws to 1970s	- State subsidies and advice to increase wood production and industrial value was taken in several steps, and led to the 1947 policy focusing on forests as effective cropping systems
Phase 3.0	1970s to 1990 and the proposed forest law, and to the 2020s	Emerging focus on nature conservation in the mid-1970s, which shaped the 1993 forest policy by introducing production and environmental objectives under the slogan “Living Forests”
Phase 4.0	2020s–	Increased EU and international influence concerning climate, energy and nature restoration (EU’s Green Deal and Biodiversity Strategy as well as regulations of emissions and removals from the land use, land use change and forestry (LULUCF), Deforestation and RED3)

). However, international [209] and EU policies about climate [152], biodiversity [153] and forests [154], as well as proposed nature restoration regulation [155] hint to a fourth phase in the evolution of policy about forests. Emerging new policy components include the importance of satisfying evidence-based conservation targets, halting the harvesting of primary and old-growth forests, adapting to and mitigating climate change, and coping with conflicts and competition in an increasingly multi-polar world [301].

4.1.3. Implementation Outputs

The need to increase the amount of protected areas in Sweden in the late 1990s was a consequence of Swedish and international guidelines and targets, which mirrored evidence-based knowledge about forest ecology and conservation biology. Following a hierarchical spatial planning approach that included strategic quantitative gap analysis for each ecoregion in Sweden [73,74], Angelstam et al. [41] concluded that “there was a straight chain of decisions from the short-term interim target for protected areas decided by the parliament, a government decision, strategies by governmental agencies, and to the regional administrations’ tactical planning to mitigate habitat fragmentation through spatial planning, as well as operational planning for designation, management and restoration of formally protected forests”. Hence, society has, so far, accepted evidence-based knowledge as a basis for biodiversity conservation (see [302]). However, to promote efficient conservation policy implementation consequences on the ground, it is important that these three planning levels are interconnected (e.g., [303]).

The process of implementing biodiversity conservation policies in terms of establishing protected areas in Sweden made use of contemporary knowledge about conservation biology, forest ecology and landscape ecology. This is in accordance with Barbour’s et al. [304] information ladder, which includes three levels. Starting with data, analyses and peer-review publication, information transfer takes place by facilitators, and this satisfies users’ need for “sound bytes” and narratives that interpret policy contents in an uncomplicated manner suitable for media and politics. Quantitative knowledge about species’ requirements was indeed used in strategic spatial planning, and forest and landscape ecology as well as approaches to collaboration advocated in tactical planning.

Angelstam et al. [41] concluded that the difference between the long-term policy goal for protected areas based on the quantitative gap analysis regarding forests below the mountain forest region (on average 10% across all ecoregions) on the one hand, and what was protected in 1997 (approximately 0.8%) was planned to be reduced by about 5%-units. This corresponds to the short-term interim target of 900,000 ha for forest protection formulated for the period 1998–2010 [34,35]. In addition, there was a long-term restoration target of an additional 4%, thus about 10% in total. However, to reach the 20% long-term conservation target, it was assumed that improved voluntary conservation through application of increased proportions of other forest management systems than rotation forestry based on the clear-felling system, and higher levels of retention of natural forest structures, would be applied [86,174].

As discussed by Angelstam et al. [41], at the end of 2010 “the short-term target (400,000 ha) for formal protection below the mountain region was reached to 80%, and the voluntary set-aside target (500,000 ha) was estimated to be reached, but with poorly known quality”. To fill the gap for formal protection, a pool of Sveaskog Co. land (100,000 ha) was made available. To conclude, while the political will was there to reach the interim target, and the support provided by the Sveaskog Co. was very important, the 900,000 ha area target was not fully reached.

Moreover, there are at least three additional challenges that need to be overcome to satisfy the policy target in terms of maintenance of viable populations of naturally occurring species in the long term. (1) To fill the gap between present amounts of habitat and what is needed to satisfy policies, different forms of nature conservation management, restoration and re-creation are needed (see [305,306]). (2) To ensure habitat quality of protected and set-aside forest areas (e.g., late successional stages and gap dynamics), and renewal of transient habitats (e.g., early successional stages such as burned forest and deciduous successions after disturbance), dynamic reserves may be needed [307]. (3) To assess the functionality of areas of different forest environments as representative habitat networks at the landscape and regional levels.

To conclude, a certain percentage of a region that is formally protected or voluntarily set aside, or low-productive forest not subject to harvesting, does not mean that functional GIs are in place in terms of providing sufficient amounts of habitat networks for viable populations of representative forest types. First, the quality of different categories of forest management potentially contributing to GIs differs considerably [201]. Additionally, the results emerging from spatial modelling to assess the functionality of different networks show that the functionality of small voluntary set-asides is generally unfavorable compared to formally protected areas [98,257,266]. Thus, reported levels of habitat network functionality may be overestimates because spatial modelling is based on remote sensing data with limited thematic resolution in terms of the ability to identify high conservation value forests [75]. Similarly, studies of biodiversity conservation planning show that there is very limited collaboration across forest ownership borders with the aim to improve habitat connectivity in landscapes [20,49,50,303].

Angelstam et al. [20,41] thus concluded “the existing areas of high conservation value forests in Sweden are presently too small and too fragmented in relation to the current forest and environmental policy ambitions”. Biodiversity conservation thus requires a combination of maintaining existing conservation values, conservation management, and restoration of forest habitats in protected areas of different kinds, as well as in the surrounding matrix. Formal forest protection represents the main investment in biodiversity conservation.

4.1.4. Consequences in Ecological and Social Systems

Protected Areas and Levels of Ambition for Biodiversity Conservation

Conserving biological diversity spans a range of levels of ambition. These range from (1) presence of species in the short term, (2) maintaining viable populations of all naturally occurring species in the long term (i.e., a current environmental objective) to (3) ecological integrity and (4) social-ecological resilience [69,189] (Figure 7). Swedish forest and environmental policy regarding biodiversity conservation clearly goes beyond the first ambition level, and focuses on the second level of ambition. In this respect, one can categorize species into five groups that vary in specialization from being generalists to highly specialized [42]. First, species which can withstand virtually whatever we do with the forest; second, those that are not threatened today, but that depend on general consideration for their long-term survival; third, specialized species which require protected or especially managed habitats; fourth, species that are doomed under current conditions, but that can be conserved with active restoration measures, and fifth, those which are already doomed to extirpation no matter what is done (i.e., the so-called extinction debt [169].

Even if Swedish forest and environmental policy goal can be interpreted as to conserve all native species in viable populations (the second level of ambition), evidence suggests that only the first level of ambition (presence of species for some time) may be possible to reach with the present levels of formal protection and voluntary set-aside of forest habitat. The reason is the Swedish landscape history context, representing a long history of strong focus on sustained yield forestry [9,165], and thus GIs with limited functionality. To achieve higher levels of biodiversity conservation ambition, remaining high conservation value forests need to be conserved and if necessary managed as parts of functional networks of different forest habitats, and large-scale efforts for active restoration of forest habitats commenced [9]. In contrast to the almost total domination of even-aged rotations in Nordic forestry with gradually lowered final felling ages, uneven-aged and mature even-aged forests (>80 years old) as well as protected areas are important to maintain biodiversity in boreal forests [308]. Their comprehensive meta-analysis thus highlighted that remnants of high conservation value forests need to be conserved to ensure the future of forest dependent species in Fennoscandia and European Russia. Thus, the most effective approach is to maintain mosaics of different forest types and development stages within landscapes.

One problem for effective conservation is that many forestry actors are not aware of, or are unwilling to accept, the need to sustain a range of different forest types and development stages, and to consider evidence-based conservation targets and the spatial configuration of conservation areas across forest ownerships [50,309]. Four methods that will contribute to achieve the policy objectives for the conservation of biological diversity are (1) forest management systems that mimic natural or cultural disturbance regimes [86], (2) conservation, management and restoration of habitats [306,310], (3) landscape ecological planning [166,227], and (4) if necessary to re-establish extirpated populations. Performance targets for the second level of ambition, to conserve native species in viable populations, imply that 10–30% of the forest should have biodiversity conservation as the main target [69,196,234,311], and the precise proportion will depend on how well the managed matrix satisfies species’ requirements. Empirical assessments of the role of the managed landscape for GI functionality are thus crucial.

The third level of ambition is ecological integrity [312,313]. Large predator populations’ control of the effects of large herbivores’ browsing is one example of this [15,181]. Migratory fish and their relationship to water flow regimes and dynamics of other species in entire catchments is another [314,315]. Another example is the interaction between forest fires and species that depend on them [316,317]. The highest level of ambition represents resilience, which means to maintain a system’s ability to recover from large-scale disturbances [318]. This issue was highlighted by Sweden at the International Meeting on Sustainable Development in Johannesburg in 2002 [319]. More frequent severe storms, long-term climate change and the associated risk of expanding fungal deceases and pest insects are other examples of threats to resilience. Additionally, international shifts in economics and desired products such as bioenergy may result in increasing pressure on the forest ecosystems. The resilience concept also has a deeper dimension linked to how societies are organized, how natural resources can be used sustainably, and how humans can live on or respond to and restore previous functionality after a large-scale disturbance. Adaption to climate change is another example. To emphasize the interconnectedness between ecosystems and people, resilience of social and ecological systems [320] or coupled human and natural systems [321] is stressed. Development of forums for inter-sectoral collaboration among actors from different sectors and at different levels plays a key role [322]. Possibly with the exception of the Swedish mountain forests, given the current intensive forest management regime, this highest level of ambition can only be satisfied in remote parts of the boreal biome on the European continent [47,323]. For this level of ambition, target levels in terms of set-aside necessary proportions are higher, likely around 40–60% [69]. This matches the current high proportion allocated to nature conservation in the Swedish mountain forests.

To conclude, during the past three decades the focus in Sweden has been to conserve species in the short term through the provision of small patches of formally protected or voluntarily set-aside forest areas. The long-term goal to conserve naturally occurring species in viable populations involves a higher level of ambition. EU-level policies pronounce even higher levels of ambition such as ecological integrity and resilience [37,154,324]. Increasing ambition levels of biodiversity conservation requires increased proportion of functional habitat networks in landscapes and regions [20,41,69]. Wilhere [302] criticized that researchers provide policy recommendations and called evidence-based conservation a myth. However, we agree with Rompré’s et al. [238] conclusion that management approaches that combine thresholds to maintain managed landscapes within the limits of natural variability are a necessary avenue.

Habitat Network Functionality for Nature and People

Well-designed networks containing sufficient amounts of protected areas with sufficient quality, size and connectivity are important building blocks for the development of GIs for species and ecosystem functions, as well as ecosystem services and benefits of nature to people. This reflects that there are both biocentric and anthropogenic perspectives.

The understanding of ecological sustainability has developed from a biocentric view towards a more anthropocentric view on ecosystems, e.g., [325]. Biodiversity in the sense of composition, structure and function of ecosystems [140], as well as ecosystem services and benefits of nature [1,2,326] mirror this. A good recent example that aims at operationalizing this dual perspective is the emergence of the concept of GI and its role for humans [37]. Human health and well-being are dependent on biodiversity and ecosystem services provided by ecosystems’ species, habitats, and processes in landscapes [327,328,329,330]. This was emphasized strongly by the Millennium Ecosystem Assessment [1]. Policy decisions, implementation processes and operational conservation, management and restoration of GI will thus have positive consequences for human well-being [331]. This includes a wide range of aspects from infectious diseases to impacts on quality of life, stress relief and often with complex functional relationships [330,332]. Although modern medicine is constantly making progress in fighting diseases and ill health, with few exceptions, about 60% of all causes of bad health, disease and premature death cannot be sought in simple relationships, such as exposure to pathogenic bacteria or genetic factors [333,334]. A considerable amount of current health hazards are lifestyle related, such as an increasingly sedentary life, physical inactivity and chronic psychological stress [335]. The ultimate reason behind such issues is a mismatch between the physical and social environment in which the human species evolved, and the dramatically changed environmental conditions in which modern humans live [336]. Recognition of the need to restore GI not only for wild species, but also humans, is necessary to reduce many of such problems [63].

Evidence from lab and field confirms positive health effects of contact with natural environments. Effects have been observed at cellular, individual and population levels [330,337]. Such effects can be utilized in health promotion. They seem most profound on diseases and disease pathways, which are responsible for a large proportion of the burden of poor health in 21st century Europe related to physical inactivity and psychological stress, poor mental health, cardiovascular and respiratory disease. This is significant because of the scale of health problems Europe faces. According to a systematic review of data from community studies in European Union (EU) countries, 27% of the adult population had experienced at least one mental disorder in the past year; an estimated 83 million people are affected. The economic cost of such problems in the EU is conservatively estimated to be 3–4% of the gross national product. The situation with physical disease is as bad. Cardiovascular disease causes over 4.3 million deaths in Europe per year, nearly half of all deaths in Europe (48%) [337]. To handle this requires a holistic approach that includes interventions related to nutrition, lifestyle, living environment as complements to pharmaceutical treatments [338,339]. Indeed, policy documents from governments, health service providers and land managers highlight the potential for natural environments to play a role in reducing the burden of poor health and narrowing health inequality. However, more effective landscape planning, management and access are needed to maximize potential benefits, and this requires a solid understanding of how natural environments, health and well-being are, and could be connected [63,340]. Research show that several different qualities are of significant importance—above all species richness, spatial extent of natural environments and silence [341]. To address this, Baldwin et al. [325] stressed the need to integrate traditional academic disciplines such as systematic conservation planning, and environmental design and planning, into biophilic design.

Integration of Planning Processes

The long history of forestry to supply the forest industry with raw materials in the Nordic countries, which have similar approaches to forest management, is one of the main reasons that a large number of forest species are red-listed [206]. The Swedish model for forest biodiversity conservation is characterized by small protected areas and general considerations in the surrounding landscape [20]. While the latter appears to have positive effects on some bird species [256,342,343], empirical studies indicate that while small protected areas of unproductive forest and retention trees do contribute to the conservation of biodiversity, they are insufficient to satisfy the environmental objective “Living forests” [20,172,201,257].

Currently, there is an increased interest in intensified forest management [9,14,54], cost-effective conservation, and development of attractive landscapes for tourism, recreation and human well-being [63,300]. To design functional GIs for biodiversity, ecosystem services and human health, all these driving forces need to be handled through increased integration of spatial planning processes for management of land and water. Additionally, more diversified suites of forest management systems are discussed [13,86,344,345]. As pointed out by Sandström et al. [300] and Axelsson et al. [344], this requires integrated bottom-up approaches, but also transparent information about the state of landscapes and regions [20,309,346].

There is indeed technical opportunity to develop input to communication, learning and spatial planning based on knowledge about species, habitats and processes by using geographical information systems to analyze data and produce maps [346], system analysis through group modelling [342,347], as well as decision-support systems [348,349]. However, because different sectors generally work in isolation from each other this is not enough. Additionally, improved collaboration among stakeholders to assure acceptable and socially robust solutions is needed. Moreover, responsible businesses and government agencies need to overcome scale mismatches and break down national and regional strategic plans to advise and counsel at tactical landscape and operational planning levels. However, in Sweden no organization has responsibility for spatial or territorial planning across sectors at the level of entire landscapes and regions. While landowners with large contiguous management units indeed have this opportunity, the most common situation is that many landowners and landowner categories are in the same local landscape or region. Three main public sector actors of relevance for GI planning are the Swedish Forest Agency, with a responsibility for forest biodiversity and the environmental objective Living forests, the county administrations, with a responsibility for protected areas, and municipalities. Swedish municipalities have a monopoly in spatial comprehensive planning, while counties and national-level government authorities produce strategic plans and ensure that municipal planning follows applicable national and EU policies [350,351]. There is an ongoing process to expand the responsibility of municipalities to cover all sustainability dimensions [352]. Additionally, knowledge-based collaborative learning forums or platforms for government functions, landowners, and other stakeholders representing different sectors and different societal levels can be encouraged [353]. The proposed way to address this issue is often called “landscape approach” [354,355,356]. With an international perspective, Biosphere Reserve [357,358] and Model Forest [353,359,360] are examples of such fora or platform concepts.

4.2. Challenges

4.2.1. Adding Conservation Efforts and Risks for Creative Book-Keeping

As a base for discussing what a certain percentage of protected areas and green tree retention actually means for biodiversity conservation, Shaffer and Stein [361] and Tear et al. [311] used the terms representation, redundancy and resilience. Representation means capturing all ecological elements or target of interest (e.g., a population, species, biotope, landscape type or ecoregion) [362]. Redundancy (i.e., to protect more than is required for a specific ambition level) is necessary to reduce the risk of losing representative examples of these targets [363]. Resilience, often referred to as the “quality” or “health” of an ecological element, is the ability of the element to persist through severe hardships [364].

The investment in biodiversity conservation by reaching the short-term interim target regarding protected areas formulated in 1997 [73,74] in terms of creating additionally 900,000 ha of formally protected areas and voluntary set-aside by the end of 2010 was reasonably successful from a numerical point of view [41]. However, spatial analyses presented in the same study indicated that requirements in terms of representation, redundancy and resilience were not satisfied. Additionally, the role of varying levels of green tree retention in final felling areas made in the matrix surrounding protected areas [96], intended to provide habitat and improve the permeability of the matrix surrounding protected areas, needs to be understood. This is also the case for areas not managed for wood production, such as forest and wooded land with low biological productivity (Swe: impediment) [20].

Regarding long-term targets for protected areas, a wide range of percentages are currently circulating among different stakeholders regarding the area proportion of forestland in Sweden that is and should be devoted to conservation of biodiversity, including species, habitats and processes. For example, Anon. [365] argued that a quarter of Sweden’s forests are not used for wood and biomass production and that “Sweden satisfies the Nagoya agreement”. This quote refers to CBD’s strategic plan for biodiversity 2011–2020 and the Aichi target number 11, which states that 17% of lands and waters shall be protected [149]. The Aichi target of 17% protected areas refers to the result of negotiations at the CBD COP meeting held in Nagoya 2010 about whether 15–25% should be protected (P-O. Ståhl, pers. comm.). Additionally, target 11 states that areas “are conserved through effectively and equitably managed, ecologically representative and well-connected systems of protected areas and other effective area-based conservation measures, and integrated into the wider landscape and seascape” [366]. The Aichi targets consider both pattern and process, and address both quantitative and qualitative criteria [234]. More recently, the EU Forest strategy [154] has nominated a 10% strict protection target and an additional 20% protection target with management, thus totaling a 30% protection target. In 2022, the Kunming-Montreal Global Biodiversity Framework [209] replaced the Aichi targets, and set the same level of ambition, namely to “ensure at least 30 per cent of areas of degraded terrestrial, inland water, and coastal and marine ecosystems are under effective restoration” (Target 2) to “ensure at least 30 per cent of terrestrial, inland water, and of coastal and marine areas are effectively conserved and managed” (Target 2).

However, there is a very large difference in the proportion of forests not used for wood production between the generally unproductive subalpine mountain forest with a total of 83% of formally protected, voluntary set-aside, non-productive forests and retention set-asides, respectively, and 17–20% in the four other ecoregions. However, we stress that those numbers do not include assessments of the extent to which these areas form functional GIs [20]. For example, formally protected forest areas are generally neither representative [146] nor with sufficient functional connectivity [41,98,257]. Additionally, voluntarily set-aside forests are subject to higher losses due to lower levels of spatial planning [41,346] and do not always host species dependent on natural forest components (compare [172] and [367]). Finally, non-productive forests host only 2% of the red-listed species [368,369,370]. Additionally, the role of edge effects [257] and degradation of natural disturbances, and cumulative effects, needs to be understood and taken into account. These limitations of conservation instruments other than formally protected areas are illustrated by Kyaschenko et al. [201]. While slight increases in structural components indicating habitat quality were observed, this has not been reflected in documented improvements for red-listed forest species because increases in the availability of forest structural components are simply insufficient.

This clearly suggests that functionality for biodiversity conservation of the ca. 25% forests and woodlands not managed for wood and biomass production is severely over-estimated. Considering estimates for efficiency rates for the four categories of forest not managed for wood and biomass production, respectively, then the effective area should be considerably lower (11% for the whole Sweden, 47% in the mountain forest region and 5–6% in the four ecoregions mainly used for wood production; see Figure 6). While this argumentation merely points at the need for assessing the consequences on the ground for biodiversity conservation, the estimates show that with 7.5% formally protected area, Sweden does not reach international agreed conservation targets. Additionally, if the different ecoregions are assessed separately, then 93% of Sweden forming the productive forests reaches up to only 2–3% units of formal protection secured in the long term (Figure 6).

Additionally, considering the managed forest landscape it is fair to include areas with low productivity, and retention forestry. These methods can potentially be viewed as tools to both create habitat and to make the matrix around formally protected areas more permeable for dispersal of individuals of different species. The coarse estimate of total functionality of (1) formally protected areas, (2) voluntary set-asides, (3) areas not used for forestry and (4) retention forestry can then be viewed as estimates of the sum of Aichi targets 11 and 7. However, even then Sweden does not reach agreed targets.

Moreover, at the same time as there are positive effects of operational management supporting biodiversity conservation at multiple spatial scales, there are also negative effects in terms of continued gradual loss and transformation of forest stands never subject to clear-felling [20,47]. These host compositional, structural and functional aspects that are more favorable to biodiversity conservation than stands that originated from clear-felling and intensive forest management for wood and biomass [140,196,371]. Additionally, the effects of intensified forestry to improve wood and bioenergy yields need to be understood [96]. Taken together, this argumentation about set-asides of stands and landscapes for conservation, and matrix management by retention forestry, emphasizes the need to understand the cumulative effect on biodiversity of two groups of drivers. These are establishing protected forests areas with or without conservation management, active habitat restoration and what can be achieved by increased nature conservation in the managed matrix on the one hand, and what happens to the last remnants of high conservation value forests on the other [54,372]. This is a major unresolved challenge.

Summarizing, while forest biodiversity policies are evidence-based in Sweden, and relevant hierarchical planning processes have been developed for formal area protection, there are gaps when it comes to landscape planning processes that integrate formally protected and voluntarily set-aside areas across forest ownership categories, and with too limited funding to secure high conservation forest remnants. Sweden is thus far from a functional landscape planning process. Researchers and engineers can develop technical solutions such as systems for analysis and reports to support decision-making. Often, however, these are not socially robust [373], which means that they are not accepted, understood or practically useful for the involved parties.

4.2.2. The Perspective of Industry or Individual Forest Owners?

In Sweden, individual forest owners as a group have freedom and great opportunities to choose forest management system. One can focus on applying cropping systems to produce industrial raw materials, or apply a diversity of forest management practices that focus on delivering many different ecosystem services. Appreciation of a wider portfolio of ecosystem services and nature’s benefits may lead to management aimed at mixed coniferous and deciduous forests, longer harvesting rotations and voluntary set-asides. The increased focus on adaptation to climate change has indeed increased the application of such adaptations [374,375]. Employing alternatives to even-aged rotation forestry that rely on natural regeneration reduce forest owner’s costs, which in turn yield increased net monetary income because expenses decrease.

For individual forest owners, income from forestry motivates <10% of them to own forest land [376]. Instead, increasing real-estate values as well as social, cultural and ecological values are important [9,377]. At the same time, Hafmar [378] showed that a large proportion (53% in Jämtland county) of individual private forest owners are interested in using alternatives to the even-aged clear-felling system. In contrast to this, the industry’s timber buyer market stresses the benefits of the clear-felling system with arguments that it maximizes the timber flow to the industry [379], which erodes the trust of forest owners [380]. The low application (ca. 3% in 2021) of “clear-cut free” forestry in Sweden has explanations that go far beyond the lack of knowledge and ecological limitations. Culture, forestry education, industrial investments, coalition networks and timber markets are important factors [9,381]. This points to a need for more comprehensive advice to forest owners on how different value chains can be developed most effectively [379]. However, an increased diversity of forestry methods means that smaller volumes of industrial raw material can be delivered.

4.2.3. Knowledge Production and Learning for Biodiversity Conservation

This review argues that to assure the functionality of GIs for biodiversity conservation and resilient ecosystems, as pronounced in policies at multiple levels, it is necessary to develop multi-sectoral and multi-level social learning processes [382] and knowledge-based collaboration [166,383]. A well-developed collaboration can be referred to as a partnership [384]. All involved stakeholders need to share the responsibility and feel that they are important parts of the problem-solving process. Obviously, no single stakeholder has all the knowledge, skills and resources needed to solve the challenges of biodiversity conservation in Sweden. The alternative is to learn stepwise through ongoing evaluations [385] and by active adaptive management [386,387]. Continuous evaluations are needed to improve policy processes, the outputs and the consequences on the ground [53].

At the same time, there is a need to understand the different levels of ambitions for ecological sustainability objectives as expressed in policies, and evidence-based knowledge about what this requires in the terms of composition, structure and function of ecosystems. Additionally, there is a need to create widespread awareness among stakeholders and the public about the contribution of different efforts in terms of conservation, management and restoration of species, habitats and ecosystem processes at multiple spatial scales. This involves social learning, i.e., how local actors learn about their place and the state and development trends of biodiversity with the aim to create an interest and to actively become a part of the development and to steer it [388,389]. This is in line with the World Forestry Congress’ [390] recommendation that environmental monitoring and assessment should include stakeholders to improve their understanding, learning and awareness.

This necessary focus on both social and ecological systems, and their interactions, contrasts in many respects with current management and governance structures, which focus on social or ecological systems in isolation from each other. A transition away from this requires the development of neutral fora and platforms for partnership development, collaboration, and informed learning about biodiversity in combination with the use of decision-support systems, e.g., [300], and appropriate policy instruments. The development of multi-sectoral learning processes may benefit also other sustainability dimensions than the ecological. We conclude that there is an urgent need to (1) increase collaboration among academic and non-academic stakeholders to facilitate learning, collaboration and sharing of knowledge and experience, e.g., [383], and (2) develop evidence-based knowledge, e.g., [178,189,391] and approaches for integrated spatial planning of GI at scales from local to trans-national, which are adapted to local and regional contexts.

Ultimately, the increased range of desired goods, services and values from landscapes requires transformation of the Swedish forestry model characterized by general considerations, e.g., [9,278] into a zoning approach [68,392], such as TRIAD including conservation, multiple-use and production [267,393]. However, depending on landowners’ preferences and available policy instruments, the opportunities for this vary considerably among local landscapes in different parts of Sweden [394]. The implementation of active adaptive management is challenged by the fragmented pattern of land ownership, limited collaboration among different forest and conservation planners [41], and the traditional sectoral management of natural resources and territorial development. This has been considered one of the reasons why the implementation of a participatory process in environmental governance is still rather limited in Sweden [264]. However, despite the inherent governance and management complexity of evidence-based participatory processes, they are a prerequisite for the sustainable development process, and ecological sustainability in the long term [320,382,395]. Informed stakeholder participation has the potential to create socially robust changes in attitudes, values and behavior because the process operates at a level where more basic human social and behavioral aspects can be reached and influenced [396]. In particular, there is a need to develop trust, equity and empowerment among involved stakeholders [397], with the aim to create space for experiential and adaptive learning where different stakeholders’ perspectives and experiences meet, and new ideas develop to allow innovations, solutions or new ways to handle complex natural resource management situations [382]. Nowotny [398] called this “the hybrid space”. This approach will in turn increase compliance and legitimacy [399,400]. Participation in collaborative learning processes with the aim to establish functional GI is a challenge since some forest sector stakeholders may see this as a threat to industrial forestry and employment in the forest sector [401]. However, the same study concluded that so far conservation has had only a limited impact on employment in the forest sector when compared to the impacts of internal processes of rationalization and mechanization. This suggests that funding for the protection and management of protected areas needs to be widened to also support concepts and initiatives aiming at cross-sectoral multi-level place-based evidence-based collaboration with the aim to create and maintain functional habitat structures in the landscape. Biosphere reserve [357,402], Model Forest [360] and Long-Term Socio-Ecological Research (LTSER) [403,404] are appropriate examples which also allow for exchange of international experiences.

Research that supports solutions to these challenges must integrate human and natural sciences, and academic and non-academic actors, e.g., [387,395,405]. Place and area-based production of new knowledge and ongoing evaluations of policy processes, outputs and consequences [53,385], and collaborative learning processes [382,406] are two important tools to carry out integrative problem-based research [405]. Thus, scholars and practitioners agree about the need to move away from the paradigm of “best-practices to be taught” based on disciplinary research, to transdisciplinary knowledge production, which produces knowledge that is socially robust and useful on the ground [385,407]. The importance of understanding ecological, societal and behavioral processes in the governance and management of GI clearly emphasizes the need for transdisciplinary knowledge production [395,407,408].

Eight actions requested in the short term by Skogsstyrelsen [43] to improve the implementation of the Swedish Environmental Objectives are:

• Intensify the development of digital high-quality geographical data about high natural and cultural forest values.
• The Government ensures that there are sufficient financial means to compensate landowners for the creation of formal forest protection, and to provide forest management recommendations.
• Clarify and elucidate today’s contradictory political signals about how forests with high natural values should be managed; for example, is final felling of high conservation value forests allowed?
• The government ensures increased resources for relevant authorities to carry out more supervisory activities with the aim of achieving better legal compliance.
• The government develops a portfolio of measures to develop and promote clearcut-free forest management methods.
• The Forest Agency and the Environmental Protection Agency to propose financial instruments and measures aimed at making visible and incorporating conservation forests and forests with high natural values into the market economy in the same way as forests aimed at wood production.
• Specify how Sweden is to achieve the national environmental quality goals as well as international commitments for biological diversity.
• Systematically monitor biological diversity in entire forest landscapes.

Efforts to cope with climate and forest landscape change must include and integrate both ecological and social systems at multiple spatial scales, i.e., what geographers call landscapes. A development from “Business-As-Usual” forestry focusing on wood production, to proactively plan use and conservation and coping with climate change and climate adaptation, is complex, e.g., [234,374,375,409]. This requires collaboration between different stakeholders and learning based on evidence and systems analysis [9]. The concept of landscape approach has therefore been developed as a method; see [410]. Systems analysis is one such way to achieve this, e.g., [310,342,347]. A robust model for this approach is presented in

Table 7. Overview of the need for development towards system analysis [347] and landscape approach [354,356] for climate change and climate adaptation that includes entire landscapes as linked ecological and social systems.

Aspects of a Landscape		At Present	In the Short Term (Decades)	In the Long Term (Centuries)
Ecological system	Forest	Even-aged clear-felling system dominates, 50–70 year rotations, monocultures	Apply forest management systems that store more carbon, longer rotations, more deciduous trees	Multiple methods, ”triad” approach through zoning of functions, resilient forest ecosystems
	Farmland	Focus on high production of few crops	Increase the use of permacultures, agroforestry and grasslands	Focus on maintaining and improving soils
	Wetland	Landscapes with many ditches have reduced the capacity of retaining and storing water	Removing ditches, re-wetting to improve carbon storage; effects on biodiversity and trends for greenhouse gases	Re-create lost wetlands that can retain and store water
Social system	Ideology	View landscapes as predictable cropping systems	Transformation to a focus on handling uncertainties and risks	Ethics and moral focusing on the future, precautionary principle, reduced consumption
	Scales for planning and management	Forest stands, fields, individual landowners; fragmentation and polarization of actors and stakeholders	Improve collaborative learning with focus on functional habitat networks and ecological functions	Management for multiple ecosystem services in entire landscapes
	Governance	One dominating sector for forestry and agriculture, respectively; sectors as silos	A diversity of value chains based on both material and immaterial values; risk analyses	Integrated governance and planning of landscapes and regions, trade-offs among ecosystem services

.

4.2.4. Wicked Problems, Disciplinary Silos and Knowledge Resistance

Across multiple natural resource sectors, increased demands of goods, services and values on the one hand, and limited supplies of those on the other, have led to conflicts and controversies among actors and stakeholders. Salmon recovery, fracking and forestry are three examples. Increasingly, these can be characterized as wicked problems [411]. According to Rittel and Webber [412], such problems share three key characteristics by being unstructured, crosscutting and relentless. Unstructured refers to the complexity and uncertainty, and little consensus on neither problems nor solutions. Crosscutting refers to a diversity of problems that cut across sectors and levels of governance. Relentless refers to solutions that are unlikely and affect other sectors as well. Current politics around forests and their biodiversity, and climate, in Sweden are a good example [413,414].

The emergence of a fourth forest policy cycle (Table 6) highlights the contrast between proactive EU and international regulations and policy to cope with biodiversity conservation, climate change and consequences of war in Europe on the one hand, and narrow national politics advocated by the traditional forest industry sector ignoring evidence-based knowledge about biodiversity on the other. The recent debate about the role of forests and forestry to cope with climate change, and of protected areas for conservation and restoration of biodiversity is highlighted by recent communications to the European Commission. A “Scientist letter sent to European Commission, regarding the need for climate smart forest management” with >500 signatures [415] argued for less forest conservation and more wood production. In response, a scientific group, also with >500 representatives, responded with a letter to the European Commission on the need to reduce forest logging for the sake of mitigating climate change and safeguarding biodiversity [416]. This underlines the need for defining the system borders for knowledge-based deliberations (Table 7), as well as deep levers that can resolve wicked problems [9,14]. Unfortunately, however, while Sweden has a long tradition of stakeholder dialog at different levels of governance, we argue that informed collaborative dialog based on evidence-based knowledge about states and trends of different dimensions of sustainable forest management is limited. Such collaboration cannot be restricted to the narrow “forest sector”, but should secure cross-cutting participation of actors and stakeholders engaged also in other value chains.

4.3. Solutions—Different Philosophies for Forest Biodiversity Conservation

4.3.1. Integration—Segregation—Triad

In Europe, there are increasing expectations that forests and forest landscapes should be multifunctional and provide many different ecosystem services. However, to achieve desirable levels of the various expectations is scale-dependent. To avoid trade-offs at small spatial extents (individual forest stands), one can manage conflicting goals on larger spatial extents (forest management unit or estate, and the entire landscape) by deliberately doing different things in different areas. A long series of articles discuss this for both forestry and agriculture [20,417,418,419,420], often based on three variants [421,422] (Figure 8).

Integration (“Land Sharing”)

Conservation considerations for various forestry measures during a rotation period is an attempt to integrate species conservation and timber production [96,423,424,425] (Figure 8). This is considered particularly important in forest landscapes with a large proportion of privately owned forests, which is the case for many parts of Europe [421,426]. If the amount of formally protected and voluntarily set-aside areas is not sufficient, and deficits in forests’ structural diversity are found in the managed landscape, then nature restoration is needed [427]. Two crucial questions are how much of different habitat structures are conserved at different scales, and for how long they survive over time [428]. Current recommendations to save 5 to 10 trees per ha at final felling [424] is below the level based on evidence-based knowledge [196,267]. To maintain 90% of the unique species richness in a naturally disturbed area, 75% of its surface should be left unharvested [425]. This illustrates that the target of 5–10 saved trees (with 500–700 stems/ha this corresponds to 1–2%) mentioned above is very low. Threshold values for dead wood for intact diversity of different taxa related to dead wood vary from 20–30 m³/ha in boreal forests to 45–50 m³/ha in temperate forests [425]. The current amount of dead wood in Sweden averages about 8 m³/ha, but does not include the diversity of decomposition stages [201,429]. For Slovenia, Nagel et al. [430] concluded that integrated management practiced on a large scale is insufficient to maintain viable populations of species dependent on naturally dynamic and old-growth forests. The same conclusion has been found in Finland [431,432] and Sweden [433]. Kuuluvainen et al. [268] expressed this kind of mismatch as “The development of retention practices in Finland indicates that the aim has not been to use ecological understanding to attain specific ecological sustainability goals, but rather to define the lowest level of retention that still allows access to the market”.

Segregation (“Land Sparing”)

This term refers to a spatial separation of high and efficient production of industrial raw materials on the one hand [337], and formally protected and voluntarily set-aside areas as components of GIs on the other (Figure 8). This requires forest management systems that achieve a regional balance between different goals. In Sweden, this has so far been solved by forests with high natural values being bought with state funds, or set aside voluntarily. Angelstam et al. [20] showed, however, that the extent of habitat networks and the functionality of representative GIs do not reach the conservation targets formulated in Swedish [45] and international policy [149]. Different species have different requirements. Those that disappear at too low levels of structures such as dead wood and old growth forest, or that need large undisturbed areas, are heavily dependent on segregative methods, but other less demanding species can be conserved through integration [419]. Combining both approaches is therefore necessary [430].

TRIAD

Combining several different management methods by zoning in landscapes is also called TRIAD, and has long been proposed as a system for sustainable forest landscape management [411,420,434]. According to this concept, protected areas and intensive forestry systems make up part of the landscape, while the rest is occupied by integrative, close-to-nature or ecological forest management systems (Figure 8). The latter forms a matrix around protected areas that can provide linkages among patches of habitat for forest species, and as a buffer to intensively managed forest stands. The Swedish state forest company Sveaskog’s division of forest stands with different objectives into Ekoparks and other landscapes with an increasing focus on production, is an example [99,246]. Within complex mosaics of forest ownership polygons, however, this can be difficult. Nevertheless, Pohjanmies et al. [435] showed that planning over small forest areas (200 ha) can contribute effectively to trade-offs among different ecosystem services. Thus, landscape planning can be feasible even in small-scale forestry if it is combined with learning [377] and tools for financial compensation [436], which of course must be adapted to the interests and abilities of different forest owners.

4.3.2. Adding Efforts of Forest Owner Categories

After a long history of providing rural livelihoods by maintaining locally multifunctional landscapes in Sweden, when most people lived in the countryside (Phase 1.0 in Table 6), forestry became focused on producing industrial raw materials (Phase 2.0 in Table 6). Currently, however, after Phase 3.0 dealt with in this study, in a new appearing Phase 4.0 forest policy cycle encompassing biodiversity conservation, effects of climate change and demands of multifunctional landscapes, are emerging [9,437]; see Table 6.

Sweden, like the rest of Europe, has many different forest owner categories. These have different desires and opportunities to produce different portfolios of ecosystem services, and represent different driving forces for and against cooperation in and about landscapes [438]. With a landscape or regional perspective, consequences on the ground of applying policy instruments should, we argue, be seen as joint efforts based on different forest owners’ own abilities and interests. A combination of several different forest management methods is an effective way to support biodiversity conservation, and a development towards resilient and multifunctional forest landscapes [439,440]. To illustrate this, the values that forests provide can be simplified into three different themes (i.e., biomass in various forms, multifunctionality, and habitat for species), and forest ownership into three groups (i.e., private industry, individual, and finally state and other public owners). This is visualized in Figure 9.

Biomass is in demand in a wide range of different forms, and increasingly in an imagined future bioeconomy. In Sweden, about 97% of the available growth of wood is harvested [20,441], and increase in monetary value takes place in various types of industry focusing on export. Increased fellings therefore requires faster growth, or reduced efforts toward conservation and nature restoration. Shorter rotation times after harvest can contribute to reduced risks of storm damage [442] and forests with multiple tree species deliver more of more ecosystem services [443,444,445]. Forestry practices that lead to a greater proportion of high-quality timber with a higher price, the opportunity to store more carbon [409] and in the future receiving payment for this [446], and the creation of a higher property value, are therefore favorable. Local value-added products based on several different forest values can provide income and contribute to the local economy and social capital [47]. Other currencies for valuation than monetary are also needed. Multifunctional forests are represented by those found in and around urban areas, and are essential for people’s well-being and health [447,448]. Additionally, forest companies today have good income from wind farms, land exploitation in urban areas and leases for hunting. This means a much wider use of the forests and several new value chains [47]. Habitat for species includes many different types of forest and woodland habitats. However, forests where the focus is on efficient timber production have difficulty delivering quality habitats, unless voluntarily set-aside old-growth forests, nature reserves, and forests with conservation management are available. Traditionally managed cultural landscapes with-wooded grasslands have been severely reduced, and it is crucial to maintain existing remnants [174,256].

Private forest industry companies focus on effective production of industrial raw materials based on a culture where forestry is seen as a cropping system [9,381]. Today, basic nature considerations are applied, but efforts aimed at functional GIs, ecological integrity and resilience are insufficient [449,450]. In contrast, individual forest owners are a heterogeneous group, which with a broader profile of advisory services than the one offered today, could further increase the breadth of ecosystem services delivered. This group also has a central role for biodiversity linked to traditional cultural landscape trees and wooded grasslands, which host much of the forest biodiversity in southern Sweden. The state and other public owners, in various forms, own most of the reported remaining natural and old-growth forests. Within the EU, Sweden, Bulgaria, Finland and Romania are the countries with the largest areas of high-conservation forests [451].

We argue in favor of encouraging that the portfolios of benefits (biomass, multiple use and habitat) should be differentiated among different forest owner categories (e.g., industry, individuals, state) (Figure 9). Finally, we discuss the opportunities for land sharing, land sparing and TRIAD forestry at the landscape and regional level to satisfy this diversified approach.

The sole use of land sharing is insufficient to conserve biodiversity. Tree retention [452] and small patches set aside within production forests do contribute to biodiversity conservation [453]. However, the retention approach cannot substitute for larger protected areas [201,308]. This is contrary to hopes of achieving higher levels of voluntary conservation ambition in earlier estimates of protected area needs [208,454].

Multiple use forest management based on a diversity of approaches is becoming of increasing interest among individual forest owners, especially because only about 20% of Swedish forest owners declare substantial income from forestry [455]. Instead, real-estate values, the feeling of owning forests, hunting and recreation are motivations for owning forest, e.g., [9,377].

The triad approach appears as the most effective in supporting multifunctional forest landscapes. Reviewing publications from Fennoscandia and European Russia, Savilaakso et al. [308] showed that compared to the current shorter (60 to 80 years) rotations of even-aged forest management systems, uneven-aged and mature even-aged forests (>80 years old) are important to maintain biodiversity in boreal forests. Importantly, their results also show that set-aside areas of natural forest remnants are needed to ensure conservation of forest dependent species. They concluded, as we do, that biodiversity conservation is best achieved by ensuring a mosaic of different forest management approaches within landscapes.

5. Conclusions

Developing planning and forest management practices that can deliver and maintain multifunctional forest landscapes, and adapting them to local and regional contexts, is complex and complicated. Simplistic statements about the areas or proportions of a certain forest type that can be harvested in time and space, protected or restored, or which forestry method is “best”, are insufficient. Instead, several complementary strategies and measures need to be combined with a landscape perspective that involves both social and ecological systems at multiple spatial scales and levels of governance [71,372,456]:

(1)

Create and maintain fora and platforms able to adapt planning and forest management to the desired ecosystem services and benefits of nature, and to local and regional conditions [457,458].

(2)

Maintain Sweden’s, and thus Europe’s, last intact forest landscapes [47,68].

(3)

Aim at maintaining representative functional habitat networks as GI by setting aside sufficient amounts of areas with sufficient quality, size and connectivity [459,460].

(4)

If necessary, also restore and re-create habitat structures at different spatial scales, and regulate processes such as grazing and browsing pressure [461], allow natural disturbances [86], and regulate predation on ground-nesting birds and large herbivores [15,342].

(5)

Support spatial planning and monitoring; combine multiple data sources to describe and measure natural forest and cultural woodland values. Note that forests that are less valuable from a wood production point of view due to low timber volumes (“green lies”) can indeed have a high degree of naturalness [462], and deliver many non-wood benefits [417,463].

(6)

Cope with wicked goal conflicts. For example, climate mitigation solutions that rely on forest bioenergy can be in conflict with carbon sequestration and storage in forests, and with climate adaption and the conservation of biological diversity [464]. Actions to manage climate change and conserve biodiversity must be integrated [465].

(7)

Although evidence-based dialogue processes for learning can be used to evaluate outcomes in real life, the decisions made depend on the worldviews, cultures, professional identities and power of different actors and stakeholders, as well as political and legal realities [255].

(8)

Encourage informed dialogues; use evidence-based qualitative and quantitative goals at multiple levels to support learning about the maintenance, management and restoration of multiple ecosystem services [20,149].

(9)

View forests as complex adaptive systems [456], strive to maintain variety at all scales and at different levels, and accept uncertainties and unpredictable events in both ecological and social systems [301,466].

(10)

Altogether, this places great demands on transforming forestry to match the objectives of different forest ownership categories, and forestry training so that forest management methods can contribute to a diversity of ecosystem services [14,154,467].