Elements of the IPBES framework

We addressed the ecological and social components that are the focus of the IPBES framework: biodiversity and ecosystems (nature), ecosystem goods and services (NCPs), human wellbeing (good quality of life), direct drivers, indirect drivers, and other institution and governance options (Díaz et al 2015). The categories of ecosystem goods and services we used in our analyses consisted of the NCPs adopted by IPBES (see Díaz et al 2018; for the lists of NCPs, see Supplemental material,  Table S1 (A1_mred-40-02-01_s01.pdf):  https://doi.org/10.1659/MRD-JOURNAL-D-19-00075.1.S1), and both the dimensions of human wellbeing and the direct and indirect drivers were taken from the Millennium Ecosystem Assessment (MEA 2005) and the IPBES assessments.

Data collection

Our analysis pertained to literature published between January 2015 and December 2018 that was available on Scopus ( https://www.scopus.com). For each of the 4 mountain systems, the literature was first filtered to include only publications pertaining to nature and ecosystems. The resulting subset was then independently filtered 6 additional times for publications pertaining to (1) the state of nature, (2) direct drivers of change, (3) indirect drivers of change, (4) ecosystem services, (5) human wellbeing, and (6) (institutional) responses (see Supplemental material,  Table S2 (A1_mred-40-02-01_s01.pdf) for the lists of search strings:  https://doi.org/10.1659/MRD-JOURNAL-D-19-00075.1.S1). While we welcome the concept of NCPs, the literature we reviewed did not widely reflect it. We therefore also included the term “ecosystem services” in our search and applied it throughout the present text. In addition to peer-reviewed articles, we included the 4 IPBES regional assessments published in 2018 ( https://www.ipbes.net/deliverables/2b-regional-assessments) and the fifth national reports to the Convention on Biological Diversity (CBD;  https://www.cbd.int/nr5/default.shtml) for the 26 countries sharing parts of the 4 mountain systems. Relevant information included in the IPBES and the CBD reports was searched using the find function and the keywords “mountain,” “montane,” and “alpine,” as well as the mountain systems’ names. To avoid introducing a bias in the level of detail reported for each of the mountain systems, we did not include the report of the Hindu Kush Himalayan Monitoring and Assessment Programme (Wester et al 2019).

The literature assessment was performed by 1 scientist and on abstracts only. Evaluation of a random set of abstracts by 2 additional scientists served to test for the repeatability and objectivity of the evaluation. Following the IPBES framework, each abstract was tagged for information on biodiversity (5 species groups) and ecosystems (6 mountain ecosystems/biomes), ecosystem services (3 groups of ecosystem services and/or 18 NCPs), drivers (5 direct drivers, 6 indirect driver categories, 8 categories of institutional and practical responses), human wellbeing (6 dimensions of human wellbeing), and 5 categories of interactions (see Supplemental material,  Table S3 (A1_mred-40-02-01_s01.pdf) for details:  https://doi.org/10.1659/MRD-JOURNAL-D-19-00075.1.S1). Additional tags were attributed to abstracts referring to any of the 17 Sustainable Development Goals (SDGs) or the 20 Aichi targets. In the cases of ecosystem services, human wellbeing, SDGs, and Aichi targets, we recorded whether the code attribution was explicit (ie a given SDG was explicitly mentioned) or inferred (for example, all abstracts that mentioned the conservation value of biodiversity or ecosystems were tagged with life on land (SDG 15), even if the SDG itself was not explicitly mentioned). To detect interactions between SDGs, we applied network analysis tools (Dalampira and Nastis 2019; Egelston et al 2019; Lusseau and Mancini 2019) to a weighted data matrix in which each row contained a paper and each column represented a different SDG. For interpretation, we used the number of SDG references in each paper (degree centrality), the importance of each SDG for the whole network (eigenvector centrality), and the SDG co-occurrences within the literature sample. The analysis was run using the R packages igraph (Csardi and Nepusz 2006) and gplots (Warnes et al 2019).

Comparative analysis of literature contents across mountain systems

The literature was most abundant for the Andes, followed by the Himalaya, the Alps, and the East African mountains (Figure 1). However, the Alps had the most publications per area (the area was calculated based on Körner et al 2011, 2017), with approximately 7.2 retained publications per 10,000 km², followed by the Himalaya (∼2.2; the literature on the Hindu Kush Himalaya [143 publications retained] consisted primarily of literature pertaining to the Himalaya sensu stricto [129 publications retained]; the value we present is based on these 129 publications, which pertain to an area surface of approximately 595,000 km²), the East African mountains (∼1.9), and the Andes (∼1.0). Abstracts primarily addressed direct drivers, ecosystems, and ecosystem services (with both explicit and inferred information; Table 1). Fewer abstracts specifically referred to species, institutions, governance, and indirect drivers, and even fewer to human wellbeing. Explicit mentions of the SDGs and Aichi targets were rare, but about half of the papers discussed issues relevant to these development and conservation goals. These observations generally apply to the 4 mountain systems. Differences between systems included proportionally more papers addressing sustainable development in the Andes and the East African mountains than in the other 2 systems, and particularly few references to notions of human wellbeing in the Alps.

TABLE 1

Percentages of abstracts addressing individual dimensions of the IPBES framework, SDGs, and Aichi targets.^a)

Biodiversity and ecosystems: The assessed literature presented large taxonomic gaps and did not offer enough systematic information on the state of and trends in biodiversity and ecosystems to perform a comparison. This was the case even though 3 of the 4 mountain systems overlap at least partly with 1 or more biodiversity hotspots (East African mountains: Eastern Afromontane biodiversity hotspot; Hindu Kush Himalaya: Himalaya biodiversity hotspot, Mountains of Southwest China biodiversity hotspot; Andes: Tropical Andes biodiversity hotspot, Tumbes–Chocó–Magdalena biodiversity hotspot, Chilean Winter Rainfall–Valdivian Forests biodiversity hotspot). An additional challenge inherent to the literature on species and ecosystems was the widespread use of different measures and indicators of biodiversity and ecosystems (eg population size, range, cover, condition, alpha and beta diversity, phylogenetic diversity).

Ecosystem services: When we made no distinction between ecosystems within mountain systems (Figure 2, bottom row), “habitat” (see Supplemental material,  Table S1 (A1_mred-40-02-01_s01.pdf) for the exact names of the ecosystem services:  https://doi.org/10.1659/MRD-JOURNAL-D-19-00075.1.S1) stood out as the most frequently mentioned ecosystem service in the 4 mountain systems. When ecosystems were analyzed individually (Figure 2, upper rows), “habitat” was the most frequently reported service in ecosystems above the treeline as well as in forests. It remained the most frequently mentioned service provided by grasslands and freshwater ecosystems in the Andes and the Alps. In grasslands of the East African mountains, “soil and hazards” was reported more frequently, and in those of the Himalaya, it was “food and medicine.” Ecosystem services associated with highly modified ecosystems (eg agricultural land) varied between systems, with “food and medicine” frequently referred to in the Andes in particular. Interestingly, abstracts pertaining to the East African mountains referred more systematically to a variety of ecosystem services.

FIGURE 2

Percentage of references to given ecosystem services in individual ecosystems. n: total number of publications retained for a given combination of mountain system and ecosystem. (See Supplemental material,  Table S1 (A1_mred-40-02-01_s01.pdf) for details on the ecosystem services:  https://doi.org/10.1659/MRD-JOURNAL-D-19-00075.1.S1.) (Mountain drawing modified from Mayoux 1996)

Direct drivers of change: When we made no distinction between ecosystems within mountain systems (Figure 3, bottom row), climate change was the most frequently reported direct driver of change in the Alps and the Himalaya. In the Andes, the number of abstracts mentioning climate change and land use change was nearly equal, whereas in the East African mountains, more abstracts referred to land use change. Unlike in other mountain systems, literature pertaining to the Himalaya included a comparatively high number of abstracts discussing issues of overexploitation, in addition to climate change and land use change. When ecosystems were analyzed individually (Figure 3, upper rows), patterns were more nuanced. In this case, climate change was more systematically discussed in abstracts pertaining to ecosystems above the treeline and freshwater ecosystems, and land use change in abstracts pertaining to forests, grasslands, and agricultural land. These observations were essentially true for the Andes, the East African mountains, and the Alps, whereas in the Himalaya, comparatively more references to climate change were made across all but the highly modified (agricultural land) ecosystems. References to invasive species and overexploitation were most frequent across all ecosystems in the literature pertaining to the Himalaya, whereas only abstracts pertaining to freshwater ecosystems mentioned pollution, albeit in all 4 systems.

FIGURE 3

Percentage of references to given drivers in individual ecosystems. n: total number of publications retained for a given combination of mountain system and ecosystem. (Mountain drawing modified from Mayoux 1996)

Institutions, governance, and other indirect drivers of change: Among the few references made to indirect drivers, economic drivers were reported most often, followed by demographic, cultural, and religious drivers. However, relative frequencies differed between mountain systems, with more frequent references to demographic drivers in the Himalaya and the East African mountains, and to economic drivers in the Andes and the Alps. For the Alps, references to demographic drivers were particularly few. The most frequently reported institutional responses pertained to “ecosystem and species management” and to “research and monitoring,” followed by “legal, regulatory and policy instruments” and “planning.” The occasional references to “economic and financial instruments” included specific measures such as payments for ecosystem services (especially for the Andes), subsidies, and the promotion of income-generating activities. Other specific measures frequently described included the establishment and management of protected areas (PAs) (and other area-based conservation measures).

Sustainable Development Goals: Based on a network analysis, research pertaining to the SDGs was primarily mentioned in the literature on the Andes and the Himalaya (Figure 4). Moreover, more abstracts simultaneously referred to several dimensions of sustainable development in the Andes than in any other system. Across systems, life on land (SDG 15) was the goal to which most abstracts referred (Figure 4, large yellow square in all subplots). In the Andes, the East African mountains, and the Himalaya, zero hunger (SDG 2) and climate action (SDG 13) ranked second and third, whereas in the Alps climate action (SDG 13) ranked second and sustainable cities and communities (SDG 11) third. Zero hunger (SDG 2), sustainable cities and communities (SDG 11), and climate action (SDG 13) were often discussed together with life on land (SDG 15) (Figure 4, dark red colors in the heat maps).

FIGURE 4

Networks of Sustainable Development Goal (SDG) co-occurrences based on the literature review. The networks illustrate to which SDG(s) a publication refers. Each publication node (black dot) is linked to the SDG(s) to which it refers in its own text. The number of links from an individual publication to a specific SDG can vary between 1 (smallest black node) and 7 (largest black node). The SDG node size (yellow square) corresponds to the SDG eigenvector centrality and the SDG label size to the betweenness centrality. Both values indicate the structural importance of each SDG within the literature sample. The heat maps illustrate the SDG co-occurrence throughout each region's set of abstracts. The number of SDG co-occurrences ranges from 1 to 9 (East African mountains), 10 (European Alps), 31 (Hindu Kush Himalaya), and 52 (Andes).

Literature-based assessment

A summary of the literature assessment along the dimensions of the IPBES framework is provided in Table 2, and the full text of the assessment, including the references, is provided as supplementary material (see Supplemental material,  Appendix S1 (A1_mred-40-02-01_s01.pdf):  https://doi.org/10.1659/MRD-JOURNAL-D-19-00075.1.S1).

TABLE 2

Summary of the literature assessment performed along the dimensions of the IPBES framework. (Table continued on next page.)

TABLE 2

Continued.

TABLE 2

Continued.

TABLE 2

Continued.

Biodiversity and ecosystems

The limited amount of information on the status of and trends in species is in line with previous observations that species information is generally very incomplete, both taxonomically and geographically (Payne et al 2017; Crouzat et al 2019).

Ecosystem services

The assembled literature consistently highlights the diverse palette of ecosystem services that mountains provide (Grêt-Regamey et al 2012; Crouzat et al 2019) to millions of people locally, in surrounding lowlands, and beyond (Schirpke, Tappeiner, et al 2019). Importantly, both the structured content analysis and the assessment illustrate the tight coupling between individual services or service bundles and ecosystem types (Figure 2), which means that the loss or degradation of individual ecosystem types can jeopardize the capacity of entire mountain landscapes to deliver vital services. The long-term provision of ecosystem services therefore calls for governance and policy frameworks applying to entire mountain systems along elevational gradients that extend all the way into the lowlands. This is of particular importance in view of the predicted shifts in ecosystems' and species' ranges under global change, which was reported across the reviewed literature. In line with a previous review of current research on mountain ecosystem services (Martín-López et al 2019), regulating ecosystem services, in particular services related to habitat and water, received particularly high attention in the assessed literature. Yet these results also became more nuanced at the scale of individual ecosystem types. Our literature assessment, in particular, points to the importance of trade-offs between individual ecosystem services and bundles of services (eg between material services, such as energy provision, and nonmaterial ecosystem services, such as aesthetic landscape perception), specifically in the Alps and the Himalaya, and to the importance of acknowledging and accounting for these trade-offs in management plans.

Drivers of change

Current discourses in nature conservation and sustainable development tend to attribute declines in mountain biodiversity, ecosystem conditions, and ecosystem extent primarily to climate change. In line with this narrative and with current funding priorities, we found that climate change indeed received considerable scientific attention in the literature we assessed (Figure 3). However, analyses at the level of individual ecosystem types revealed that climate is more systematically mentioned as a driver of change in literature pertaining to ecosystems above the treeline and land use change in literature pertaining to ecosystems below the treeline. Yet variation existed between mountain systems, with literature on the impacts of climate change also pertaining to ecosystems below the treeline in the Alps, for instance. These patterns emphasize the importance of accounting for the succession of ecosystem types along elevational gradients when studying global change drivers in mountain systems in the Anthropocene and when formulating policies and management approaches. The prevalence of land use change across the broad selection of publications we screened gives it a higher importance than the one it receives in the literature on mountain ecosystem services (Martín-López et al 2019). Our literature assessment (Table 2; and see Supplemental material,  Appendix S1 (A1_mred-40-02-01_s01.pdf):  https://doi.org/10.1659/MRD-JOURNAL-D-19-00075.1.S1) added further nuances, for example by attributing different forms of land use change and climate change to different mountain ranges and ecosystem types and by highlighting the importance of interactions between drivers of change. Specific interactions included those between climate change and land use change, as well as between invasive species and other drivers (Carboni et al 2018; Shrestha and Shrestha 2019).

Institutions, governance, and other indirect drivers of change

In line with recent literature reviews (Martín-López et al 2019), the number of direct and inferred references to indirect drivers of change was also low in our assessment. References to indirect drivers pertained primarily to economic and demographic factors, with differences between mountain systems. Literature information on governance and policy frameworks covered various instruments but was most specific about PAs and financial instruments. The importance given to PAs in the literature reflects the role of PA designation as a flagship contribution toward safeguarding nature in mountains, despite major shortcomings including gaps in coverage and mismatches with areas of high conservation value (Rodríguez-Rodríguez et al 2011; Elsen et al 2018). The references to payment for ecosystem services schemes, particularly in the literature pertaining to the Andes and to a lesser extent to the East African mountains and the Himalaya, is consistent with their increasing application in these regions (see Martín-López et al 2019 for references).

SDGs and Aichi targets

The relatively higher number of references to notions of sustainable development in the Andes might reflect the importance of the holistic concept of buen vivir (“living well”) in the political and scientific agendas of Bolivia and Ecuador, in particular (Vanhulst and Beling 2014), to which about half of the selected publications pertained. The importance of life on land (SDG 15) and climate action (SDG 13) in the assembled literature is in line with previous observations of an influence of life on land (SDG 15) on climate action (SDG 13) and vice versa (Ehrensperger et al 2019). It also supports previous calls for addressing environmental and climate issues together (Wymann Von Dach et al 2018). The importance of zero hunger (SDG 2) in the Himalaya, in turn, is exemplified by results from Nepal (Wymann Von Dach et al 2018). Differences in the importance of individual SDGs between mountain systems confirm the need for localized research and analyses at subnational and regional scale (Kulonen et al 2019), which can be facilitated by methods such as the SDG synergies approach (Barquet et al 2019). Context sensitivity is a main concern toward practical applications and the translation of findings on SDG interactions into concrete policy advice (Breuer et al 2019).

Methodology

The methodology we applied has some limitations. First, our literature assessment was based on a subset of articles published in English, of which we read only the abstracts. Accordingly, additional content pertaining to the various dimensions we assessed may have been included in the full articles. Second, with our focus on mountain biodiversity, the literature was first filtered to include only publications pertaining to nature and ecosystems. The output therefore reflects a topical prioritization of research on nature and people. Accordingly, comparatively many publications referred to or explicitly mentioned habitat ecosystem services and the SDG life on land (15), and the statistics on the numbers of publications assessed do not reflect the absolute number of publications addressing human wellbeing or governance and indirect drivers in mountains. A search without filters would have provided a more holistic view of ongoing research with a balanced emphasis for the social and societal components of the IPBES framework but would have yielded an unmanageable number of publications. Moreover, papers selected based on search terms for governance, indirect drivers, and dimensions of human wellbeing often do not establish an informative link with nature, biodiversity, and ecosystems.

Third, we might have missed a nonnegligible number of publications pertaining to the mountain systems of interest because of insufficient or inappropriate georeferencing. The inclusion of individual mountain ranges in the search string, which we did for the East African mountains with their relatively few individual mountain ranges (see Supplemental material,  Table S2 (A1_mred-40-02-01_s01.pdf) for the lists of search strings:  https://doi.org/10.1659/MRD-JOURNAL-D-19-00075.1.S1), would partly address this issue. However, given the number of mountain ranges in individual systems such as the Andes and the absence of a standardized hierarchy of mountain ranges within systems, this is impossible. Both better and more systematic georeferencing of papers based on meaningful standards and specific geographic search engines (eg JournalMap;  http://www.journalmap.org) could help to overcome geographic biases in literature searches (Karl et al 2013).

Fourth, we also performed our comparative analysis on information that was not always explicit but inferred during the coding of the abstracts. In particular, as the SDGs were not included in the search terms, information pertaining to them was mostly inferred. However, this coding was performed for all 4 systems by the same person and the results were coherent with recent publications (Martín-López et al 2019). Fifth, our choice of search terms and of sampling method means that our results give information about ongoing research and not strictly about the actual relationship between nature and people. However, it is likely that the research priorities reflect, at least partly, real relationships.

Finally, the application of a biodiversity lens and, even more so, the adoption of the IPBES framework quite certainly influenced the type of knowledge we assessed. The IPBES framework serves as a powerful tool to guide and organize data acquisition with a focus on biodiversity, facilitate comparisons, deliver global recommendations, and enable the uptake of mountain-specific knowledge on biodiversity and ecosystems into global narratives. However, the adoption of standardized categories (eg of drivers or ecosystem services) in a general integrative framework (Díaz et al 2015) entails a certain loss of information. This information includes nuanced, detailed, and management-as well as policy-relevant information at pertinent scales, which is particularly needed in heterogeneous and complex mountain social–ecological systems. Accordingly, taking into account further conceptualizations of social–ecological systems, such as the conceptual model proposed by Klein et al (2019) or the general framework developed by Ostrom (2009), might grant access to more actionable forms of knowledge. This might particularly be the case for the model of Klein et al (2019). This model is specific to mountains and concentrates on some of their distinct characteristics, on particular paradoxes emerging from nonlinear interactions among these characteristics, and on the distinction between sustained and episodic drivers that affect mountain social–ecological systems at different temporal and spatial scales.

Recommendation 1: improving biodiversity information

Recommendation 2: generating spatially explicit knowledge on ecosystem services

Recommendation 3: integrating knowledge and action along elevational gradients

Recommendation 4: generating knowledge on interacting effects of global change drivers

Recommendation 5: generating knowledge relevant for transformative action toward sustainable mountain development

Recommendation 6: addressing knowledge gaps by using comprehensive concepts and codesigned approaches