ECOLOGICAL Role

The ecosystem‐level consequences of biodiversity loss

As mentioned in Duffy et al. 2007, "Global biodiversity is increasingly threatened by human domination of natural ecosystems and concomitant impacts that accelerate rates of population and species extinction, and homogenization through invasion (Vitousek et al. 1997; Sala et al. 2000). These changes raise fundamental questions, such as: What are the community and ecosystem‐level consequences of biodiversity loss? Will extinctions alter basic ecosystem processes, including those that produce food, purify air and water, and decompose harmful wastes? To address such questions, the relationship between biodiversity and ecosystem functioning has emerged during the last decade as a vigorous new research area linking community and ecosystem ecology (see general syntheses in Loreau et al. 2001, 2002; Hooper et al. 2005)."

Link between biodiversity and ecosystem function

As developed by Strong et al. (2015), "Biodiversity and ecosystem function (BEF) relationships, if present, can take many forms. These arise from linear positive or negative relationships (proportional gain or loss) or exponential (high and low redundancy models) (Naeem and Wright, 2003).

• Positive, linear BEF relationships suggest that additional units of biodiversity (this can be taxonomic units such as species richness or functional diversity) have an equal and additive contribution to an ecosystem function (Naeem and Wright, 2003). This would be indicative of situations where complementarity (transgressive over-yielding) was occurring, i.e. individual species perform better in diverse communities when compared to monoculture due to facilitation and niche partitioning in shared resource use.

• When several biodiversity units are capable of providing the same function, and therefore the same change in ecosystem function, logarithmic (or redundancy) relationships are likely. The species range beyond the asymptote are often considered to be redundant (Naeem and Wright, 2003).

• Complementarity provides what might be considered the truest form of BEF relationships. However, in situations where particular species have a disproportionate functional role (keystone species), they can also generate positive BEF exponentiel relationships and are termed identity effects (this form of non-transgressive over-yielding can also be called sampling or selection effects).

• Where identity effects are prevalent, stepped or riveted relationships might be apparent.

Functional diversity measures (such as Biological Traits Analysis that uses a series of life history, morphological and behavioural characteristics of species present in assemblages to indicate aspects of their ecological functioning e.g. Bremner, 2008), rather than taxonomic methods, are suggested as way of partially compensating for both redundancy and identity effects."

The role of keystone species

As explained in https://taniaroseesteban.wordpress.com/science/, "In different ecosystems, each specie plays a role within a community and can influence the food web and ecosystem functioning. However, the relative impact of each species can vary in terms of importance. Such species that have disproportionate effects on ecosystems relative to their population are known as keystone species. According to network theory, keystones are intimately linked via ecological networks of highly connected and complex webs. These include species at different trophic levels.

Apex predators exert top-down effect on these levels, known as trophic cascades; whereby these strongly connected species indirectly influence community structure and ecosystem function. The robustness of food webs to species removals varies, depending on the species and ecosystem type, where certain removals have greater impacts on ecosystem functioning and structure. Many apex predators are classed as keystone species since the loss of predator species can have impacts that cascade down a food chain to plants, altering basic ecosystem processes (trophic cascades). Predators directly impact upon herbivore numbers as well as indirectly through risk effects. This then influences the relative abundance of producers- hence a cascading effect." This avenue of research on the role of apex predators as keystone species is especially important in light of growing evidence that a variety of human impacts cause preferential extinction of top predators (Dobson et al. 2006) and that top‐down control extends farther, on average, through food webs than do bottom‐up effects of resource supply (Borer et al. 2006)."

As explained in Duffy et al. 2007, "One classic example is the kelp – sea urchin – sea otter food chain in the northeast Pacific. Hunting of sea otters by fur traders in the late 19th century caused a population explosion of their sea urchin prey, and consequent overgrazing of kelp forests (Estes & Palmisano 1974). Loss of kelp led to local extirpation of numerous other species that depend on kelp for habitat, as well as increased coastal erosion and storm damage since kelp was a primary buffer from wave action. Similar cascading effects of predator removal have since been documented in a wide variety of ecosystems (Pace et al. 1999; Borer et al. 2005)."

Another famous example is the Yellowstone wolf story. Wolves were hunted to extinction in the Greater Yellowstone Ecosystem in the early part of the 20^th century. Decades and decades went by without this top-predator. As a result, the northern range elk population (the larger of two elk herds in the park) eventually sky-rocketed. All these elk needed a lot of food, which is especially limited in the winter, and so the vegetation was overgrazed and subsequently declined.

The 1995/1996 reintroduction of gray wolves (Canis lupus) into Yellowstone National Park after a 70 year absence has allowed for studies of tri-trophic cascades involving wolves, elk (Cervus elaphus), and plant species such as aspen (Populus tremuloides), cottonwoods (Populus spp.), and willows (Salix spp.). After wolf reintroduction, elk populations decreased, releasing browsing pressure on woody plants, while beaver (Caster canadensis) numbers increased, possibly due to the increase in available woody plants and herbaceous forage resulting from less competition with elk. This has ultimately consequences on stream morphology, coyotes and bears, scavengers, birds, small mammals and insects, etc.

The role of ECOSYSTEM ENGINEERS

Ecosystem engineers are organisms that directly or indirectly modulate the availability of resources (other than themselves) to other species by causing state changes in biotic or abiotic materials. In so doing they modify, maintain and/or create habitats” (Jones et al. 1994).

• Allogenic engineers (e.g. woodpeckers, beavers) change the environment by transforming living or non-living materials from one physical state to another, via mechanical or other means.

• Autogenic engineers (e.g., trees, corals) modify the environment by modifying their own physical structures (their living and dead tissues).

Beaver are familiar examples of organisms acting as allogenic ecosystem engineers. By cutting trees and using them to construct dams they alter hydrology, creating wetlands that may persist for centuries. "These activities retain sediments and organic matter in the channel, modify nutrient cycling and decomposition dynamics, modify the structure and dynamics of the riparian zone, influence the character of water and materials transported downstream, and ultimately influence plant and animal community composition and diversity" (Naiman et al. 1988).

Another example of allogenic ecosystem engineer is Prairie dogs which are considered as allogenic ecosystem engineers due to the fact that the species has the ability to perform substantial modifications by burrowing and turning soil. They are able to influence soils and vegetation of the landscape while providing underground corridors for arthropods, avians, other small mammals, and reptiles. This has a positive effect on species richness and diversity of their habitats which results in the prairie dogs being labelled as keystone species.

Trees are a good example of autogenic ecosystem engineer, because as they grow, their trunks and branches create habitats for other living things; these may include squirrels, birds or insects among others. In the tropics, lianas connect trees, which allow many animals to travel exclusively through the forest canopy.

Another example of autogenic ecosystem engineers in marine environments would be scleractinian corals as they create the framework for the habitat most coral-reef organisms depend on.

The role of mobile-link species (active subsidies)

Animals can also affect ecosystem processes by moving nutrients, energy and their consumption of other organisms within and among ecosystems (spatial subsidies; Polis, Anderson, & Holt, 1997). Indeed, many species are highly mobile across space and time and can use different habitats for their different activities over the day and year. As a result of their movements between different habitats across landscapes, they can connect provide links between communities, habitats and ecosystems that otherwise remain separate, and influences ecosystem dynamics, inducing unique biotic and abiotic flows (such as seed, parasite and nutrient flows), at varying distances across the landscape. Those species are called ‘mobile‐link species’ (Lundberg & Moberg 2003; Jeltsch et al. 2013), mobile agents (Kremen et al. 2007), dispersal vectors or active subsidies (Earl & Zollner 2017).

As developped in Earl & Zollner (2017), "Active subsidies can have large effects on trophic dynamics, animal behaviour and nutrient status in recipient ecosystems (Marcarelli, Baxter, Mineau, & Hall, 2011). Further, they are related to a number of active areas of ecological research, such as alternative stable states (Nakazawa, 2011), apparent competition (Murakami & Nakano, 2002), source- sink dynamics (Loreau et al., 2013), maintenance of biodiversity (Nakano & Murakami, 2001), and the relationship between biodiversity and ecosystem function (Allen, Vaughn, Kelly, Cooper, & Engel, 2012). Beyond their ecological importance, they can also support ecosystem services (Kremen et al., 2007), affect the conservation of species, and transport contaminants (e.g. Walters et al., 2009) and pathogens. This concept provides a functional perspective, i.e. the movement of individuals is investigated with respect to its effects on other processes that impact biodiversity".

"Active subsidies involve the movement of materials (i.e. nutrients and energy) by an animal vector from a donor to recipient ecosystem. The concepts of donor and recipient ecosystems fit within more generalized concepts of sources and sinks within ecology and the earth sciences (Loreau et al., 2013). Animals are characterized by traits (i.e. animal characteristics including behaviours, recalcitrance of tissues, digestive efficiency and prey choice) that affect how they deposit materials and subsequent ecosystem effects. The movement process itself is also modified by the animal vector’s traits (e.g. movement strategy) interacting with landscape and ecosystem characteristics. Further, subsidy effects in the recipient ecosystem depend on community structure (Nakazawa, 2011), and environmental drivers can affect aspects of these processes. Thus, active subsidy research is at the interface of ecosystem, community, behavioural and landscape ecology, which presents a number of challenges. Massol et al. (2011) previously identified challenges for research on meta- ecosystems in general and proposed an integration of food web meta-community and landscape ecosystem approaches."

"When active subsidies are nutrients, energy, or prey, they can alter population dynamics and behaviour of consumer organisms in recipient ecosystems (e.g. Kato, Iwata, Nakano, & Kishi, 2003) and cause bottom- up effects. The nutrients and energy can be the animal’s own body as prey or carrion, waste products, carcasses of other animals, reproductive materials (eggs, embryos), or other carried organisms, such as parasites or seeds that become prey in the recipient ecosystem. These diverse materials differ in quantity, quality and stoichiometry, which determine their effects on the recipient system (Marcarelli et al., 2011). One well- known example is salmon moving from marine environments to streams where they breed and die. This transfer of nutrients and energy increases stream water nutrients and the density of benthic organisms (Janetski, Chaloner, Tiegs, & Lamberti, 2009). Many consumers of subsidies track emergence of prey subsidies behaviourally by shifting foraging areas (Nakano & Murakami, 2001), resulting in altered growth patterns (Sabo & Power, 2002). Further, the presence of subsidies can alter the status of a habitat as either a source or a sink (Loreau et al., 2013). When active subsidies are consumers, they can cause top-down effects through the transfer of their metabolic demands to the recipient ecosystem. This increased consumption of prey or plants can alter the stability and structure of communities (Polis, Holt, Menge, & Winemiller, 1996). For example, dragonflies metamorphosing from ponds can negatively affect populations of pollinators leading to insufficient pollination and lowered terrestrial plant seed set (Knight, McCoy, Chase, McCoy, & Holt, 2005). Subsidies can also result in complex interactions, such as pparent competition between prey subsidies and in situ prey (Murakami & Nakano, 2002)."

Based on what animals primarily transport and translocate between areas, these links have been categorized as resource, process, and genetic links (Lundberg and Moberg 2003). As explained by (Jeltsch et al. 2013):

• "Resource linkers transport energy, organic and inorganic material, e.g. nutrients or minerals (e.g., seabirds concentrating nutrients via guano deposits, salmon transporting nutrients and energy upstream and possibly redistributed by scavengers and predators).

• Genetic linkers transport mainly ‘genetic material’ into a community, e.g. by transporting genes within seeds, propagules, microbiota or other organisms (e.g., flying foxes on oceanic islands that disperse seeds and pollen of various endemic plant species, large herbivores that disperse seeds and improve their germination through endozoochory in Savannah ecosystems)

• Process linkers engage in some activities that provide new or intensify existing ecological processes (e.g., grazing of big mammals or herbivorous birds, which affects nutrient cycling, biomass production, disturbance regimes and consequently plant species composition, or predation)."

"Mobile links provide a multitude of different functions, such as pollination (Allen-Wardell and others 1998; Buchmann and Nabhan 1996), seed dispersal (Hutchins and others 1996), the translocation of nutrients (Meyer and Schultz 1985; Polis and others 1997), and grazing (Carpenter 1986; Walker 1993). Therefore, they often have pivotal effects on ecosystem processes, especially following disturbance, through their input of, for example, seeds, pollen, mycorrhiza, as well as essential processes such as grazing. Furthermore, they are essential components of ecological memory (Bengtsson and others forthcoming; Nyström and Folke 2001). Organisms that have mobile link functions can have substantial effects on ecosystem functioning and structure (Mills and others 1993). In this sense, they often provide functions analogously to keystone species (Paine 1969). However, the classic keystone species concept has focused on the experimental manipulation of less mobile top-level carnivores and their top–down effects. (...)

Thus far, these biotic vectors have not been considered to the same extent in terms of their role in the functioning and dynamics of ecosystems, but the situation is changing." Indeed, we have started to realize that mobile links are crucial for maintaining ecosystem function, memory, and resilience (Gilbert 1980; Nystrm and Folke 2001; Lundberg and Moberg 2003).

Humans, through anthropogenic environmental changes (e.g., habitat loss and fragmentation), can impact behaviours and especially spatial ecology of mobile link species, and as such, their ecological role, with important consequences on ecosystem functioning.

As stated by Lundberg & Moberg (2003), "There is a growing appreciation among ecologists that flows of matter and organisms can link seemingly isolated systems and exert a substantial influence on local patterns and dynamics (Hilderbrand and others 1999; Huxel and McCann 1998; Polis and others 1997). (...) The rate, timing, duration, frequency, and spatial extent of a mobile link function could all be affected (Dukes and Mooney 1999; Harrington and others 1999; Hughes 2000), leading to profound changes in local ecosystems (Post and others 1998). For example, the lack of pollinators and seed dispersers might lead to recovery failures and phase shifts after disturbances (Buchmann and Nabhan 1996; Cox and others 1991). In addition, so that an existing organism-mediated link could contribute to the spread of aggressive alien species (Larosa and others 1985; Simberloff and von Holle 1999; Woodward and others 1990), genetically modified organisms (Scheffler and Dale 1994), disease (Epstein 1999), pathogens (Olsen and others 1995), or pollutants (Ewald and others 1998)."

In particular, humans affect the dispersal process in a variety of ways, the result of which can be characterised as human-mediated dispersal (HMD), which presents two main forms (Bullock et al. 2018):

• Human-vectored dispersal, which occurs when humans transport organisms directly.

• Human-altered dispersal, which encompasses the indirect effects of humans on dispersal by altering landscape structure, dispersal vectors, and animal behaviour.

As Bullock et al. (2018) explained:

"This HMD modifies long-distance dispersal, changes dispersal paths, and overall benefits certain species or genotypes while disadvantaging others. It is leading to radical changes in the structure and functioning of spatial networks, which are likely to intensify as human activities increase in scope and extent. "

"Particular species or genotypes benefit from increased dispersal ability under HMD, including new linkages among areas of suitable habitat; conversely, others suffer from loss of dispersal opportunities and linkages, as well as increased costs. In total, HMD is expected to rewire spatial networks through the reconfiguration of links among nodes, particularly by changing the distances over which individuals disperse and the creation of highly connected nodes (hubs)."

"Habitat fragmentation involves habitat loss leading to the division of large areas of continuous habitat into smaller patches, isolated from each other by less suitable matrix habitat. It is well documented that this fundamental change to landscapes often has negative effects on dispersal, due to avoidance of the matrix and/or increased dispersal costs across less hospitable environments. However, effects vary and, in some cases, dispersal can be unaffected or even boosted following fragmentation, such as where animal vectors move longer distances to access isolated habitat patches. Within fragmented landscapes, adding green infrastructure can enhance dispersal, while red infrastructure can further restrict movement. Green infrastructure includes anthropogenic linear features, such as canals, unpaved roads, and firebreaks { soft linear development (SLD)}, which often act as dispersal corridors, for example by allowing the movement of frugivorous animals. SLD might be able to conserve many of the structural features of the original habitat (e.g., extant or planted vegetation along canals and roads). Red infrastructure includes high-impact developments, such as highways, which can act as physical barriers that impede movement, or as behavioural barriers when animals are reluctant to cross them because they are perceived as a risk. Indeed, the extent of animal movements is typically much reduced in areas with a large human footprint, such as those with a high building density. The level of neophobia (i.e., a fearful response to novelty) of a particular species or individual is an important characteristic that determines how anthropogenic infra- structure affects its movement. For example, while roads act as barriers to some species, they are preferred dispersal routes for others because movement is easier than within dense, intact vegetation. "

"Harvesting can reduce animal emigration by relaxing density dependence, but conversely can induce emigration through risk avoidance (landscape of fear ). At a broader scale, localised population control or harvesting can induce immigration from non- hunted areas, potentially creating source – sink dynamics. Finally, some species suffer dispersal limitation due to the loss of their natural animal vectors through human activities, such as hunting, cascading effects of biological invasions, or anthropogenic habitat destruction."

"HMD-driven changes in dispersal patterns within ecological systems can have a wide array of impacts, which might be negative or positive depending on the characteristics of the human activities. While these effects are likely to be multifarious, we can make some predictions at various levels of organisation. For example, we describe above the limitation of dispersal through habitat fragmentation and loss of dispersal vectors. This has been shown to constrain metapopulation processes, such as rescue effects, and to lead to population bottlenecks. We expect that species dispersed by extinction-prone animals, such as large- bodied vertebrates, will be particularly sensitive to these effects. Dispersal collapse might even occur long before vector extinction, particularly when remnant vectors cannot provide sufficient dispersal. Importantly, human activities rarely happen in isolation, and so we expect that, in some cases, these detrimental effects might be compensated for to a certain degree by positive effects of HMD. For example, HVD might increase dispersal rates and distances, which have been shown to benefit certain species by: connecting isolated habitat patches and so increasing metapopulation size; promoting range expansions; and creating dispersal hubs that can maintain metapopulations. However, because increased dispersal in metapopulations might also exacerbate dispersal-driven population synchrony and the likelihood of extinction, we anticipate that effects of HMD on metapopulations will vary according to the relative contributions of human-altered and -vectored dispersal. A further consideration is how HVD changes the locations at which individuals settle. Natural dispersal, even when passive, is often directed such that individuals settle at locations that support survival and establishment, presumably as a result of dispersal evolution. We predict that HVD will disrupt this process if the new settlement locations are less favourable.

The combined effects of HMD on the individual species in spatial networks will drive local community and ecosystem properties and, ultimately, those of metacommunities and meta- ecosystems. First, in contrast to typical views that human impacts cause foodwebs to become more connected and less modular, we anticipate complex changes in local communities because both mutualisms and antagonisms are shaped by both lower and higher-order interactions. These interactions are shaped dynamically by the relative abundances of the different species as determined by human-imposed changes in habitat quality and also increased immigration under HMD. Second, HMD-driven changes in dispersal patterns might eventually change the ecological functions within communities in a nonrandom fashion, for example through network-wide domination by certain species with an homogenising impact on the network structure and local functioning. Here, the correlation between those traits of species promoting the likelihood of HVD, or the filtering effect of HAD, and those shaping interspecific interactions and ecological functions (e.g., body size) will condition the structural and functional consequences of HMD. Finally, we predict that HMD will change metaecosystem processes. Again, depending on the human activities, we can predict different consequences. While colonising species might export nutrients by moving among or switching foraging areas, constrained dispersal for large-seeded plants due to the loss of large- bodied animal vectors leads to metacommunities dominated by small-seeded plants with consequences for regional carbon storage."

Ecosystem services are functions provided by nature that improve and sustain human wellbeing (Daily 1997). Some ecosystem services, such as pollination, pest control and seed dispersal, are produced at a local scale by mobile organisms foraging within or between habitats (Gilbert 1980; Lundberg & Moberg 2003; Sekercioglu 2006). Kremen et al. (2007) called these services, generated by mobile link species, mobile agent‐based ecosystem services (MABES). As explained by Kremen et al. (2007), "Although these mobile organisms deliver services locally, their individual behaviour, population biology and community dynamics are often affected by the spatial distribution of resources at a larger, landscape scale. Managing for mobile organisms and the services they provide therefore requires considering not only the local scale where services are delivered, but also a landscape scale that reflects both the spatial distribution of resources and the foraging and dispersal movements of the organisms themselves."

As developped in Earl & Zollner (2017), "Active subsidies alter the production of ecosystem services. Animals transporting subsidies can provide provisioning services, such as migratory game species (e.g. geese), and support provisioning services by transporting nutrients, energy or pollination to crop, game or timber species (Kremen et al., 2007). Further, nutrient cycling, a key aspect of active subsidies, is itself a supporting service necessary for maintaining ecosystem function (Millennium Ecosystem Assessment, 2005). For example, salmon input marine- derived nitrogen to streams and riparian forests (moved by bears; Quinn, Carlson, Gende, & Rich, 2009), which increases the growth rate of trees (Drake, Naiman, & Helfield, 2002). Salmon also provide a vari- ety of ecosystem services as an important game fish and providing food for other game species and species that attract ecotourism (Darimont et al., 2010). Spatial subsidies can substantially affect species and ecosystem conservation. Endangered species can be important nutrient vec- tors, and as these species have declined, many subsidies have been greatly depleted (Doughty, Roman, et al., 2016). Spatial subsidies can also be important resources for endangered species. For example, en- dangered Orcinus orca (killer whale) populations in Washington, USA, depend heavily on migratory salmon (Foster et al., 2012), and Myotis sodalis (Indiana bats) depend on emergence of aquatic insects (Kurta & Whitaker, 1998). Further, subsidies can impact species and ecosystem conservation through anthropogenic enhancement of ecosystem connections. Active subsidies can be a substantial component of edge effects (Ries, Fletcher, Battin, & Sisk, 2004). Habitat fragmentation increases the ratio of edge to interior habitat, facilitating an increase in subsidies. For example, spillover of predators and herbivores from agriculture can damage native plants and reduce native insect popu- lations (Rand, Tylianakis, & Tscharntke, 2006), and pathogens can also be transferred to new areas (e.g. Nobert, Merrill, Pybus, Bollinger, & Hwang, 2016).

Active subsidies are additionally capable of degrading ecosystems. This degradation can occur through an anthropogenic change in one area that increases the subsidy vector population and results in transferring excess nutrients or consumption. For example, expansion of agriculture in the United States has increased the food supply for migratory geese in their winter ranges, which increased geese transport of nutrients to wetlands causing eutrophication (Kitchell et al., 1999), as well as raising geese populations. The larger populations led to greater transfer of geese consumption of plants in the arctic summer range, resulting in complete removal of plants in some salt-marshes (Jefferies, Henry, & Abraham, 2004). This example shows that changes in populations of subsidy vectors can lead to negative changes on multiple spatial scales. Animal vectors can transport contaminants, particularly by species that biomagnify heavy metals and persistent organic pollutants. Aquatic insects move polychlorinated biphenyls and heavy metals from aquatic to terrestrial systems, resulting in elevated contaminant levels in terrestrial spiders and birds (Walters et al., 2009). Anadromous fish and seabirds biomagnify marine contaminants and deposit them in freshwater and terrestrial ecosystems (Krümmel et al., 2003). Contaminant transfer by active subsidies can cause toxicity in recipient organisms (Blais et al., 2007) and subsequently control pred- ator populations (Paetzold, Smith, Warren, & Maltby, 2011). Blais et al. (2007) suggested that contaminant biotransport was likely to have larger impacts when animals congregate during subsidy deposition or involve animals with very large biomass (e.g. whales)."

To conclude, “We have only a basic understanding of how behavioral responses in one species, through its effects on others, might alter ecosystem processes” (Wong & Candolin 2014) as well as ecosystem services. Importantly, the impact of mobile animal species, such as large vertebrates, on the functioning of ecosystems and ecosystem services must be understood on a broad scale (landscape or meta-ecosystem scales, or even regional scales) and explicitly take into account their movements and behaviours (habitat use, activity rhythms, feeding, dispersion, migration).

THE SOCIO-ECOLOGICAL SYSTEM FRAMEWORK

As explained in Bretagnolle et al. (2019):

"In the Anthropocene (Lewis and Maslin 2015), humankind’s global footprint in terrestrial ecosystems gradually increased from 5% to more than 50% in just 3 centuries (Ellis et al. 2010). Already, human impacts on ecosystems worldwide have resulted in a dramatic decline in biodiversity (Pimm et al. 2014), with measurable consequences for ecosystem services (ESs; Balvanera et al. 2014). Ecosystems will be even more intensively used in the future because the human population is still growing rapidly (Carpenter et al. 2009). Altogether, increased human pressure on ecosystems, global change, finite resources, and economic instability urge decision makers to frame new paradigms for sustainable development to achieve human well-being for all (Ellis 2015). Locally relevant indicators of the system’s state were developed to prompt public action (e.g., Dearing et al. 2014), but the analysis of the relationship between social and biophysical conditions at broader scales, e.g., the landscape scale, as a tool to foster changes in management from a system dynamics perspective is still lacking.

Environmental problems result from social, technical, economic, and ecological variables that not only form complex systems on their own, but also can interact to create wicked problems with intricate causes and consequences. Solving them calls for a new research posture, shifting from monodisciplinary approaches to transdisciplinarity (Jahn et al. 2012). The latter allows accounting for various and diverging viewpoints and involves explicit stakeholder knowledge, as well as cooperation between science and society (Spangenberg et al. 2015, Church 2018).

The social-ecological system (SES) framework is particularly adapted to implement such an approach based on interdisciplinary and transdisciplinary research that links social and ecological systems as an integrated science-policy research agenda. Indeed, it makes explicit the coupling interfaces between social and ecological templates to use leverage tools and promote action for active social-ecological system (SES) stewardship (Chapin et al. 2010).

Most natural ecosystems have been colonized and exploited by humans, becoming SESs. SESs combine interdependent social and ecological dynamics that involve multiple interactions and feedbacks between the human and ecological components (Collins et al. 2011), are adaptive (Folke et al. 2005, Levin et al. 2013), and loop into co-occurring complex (Holling 2001) and cross-scale (Levin 1998, Cash et al. 2006) dynamics. Addressing solely the social dimension of resource management without ecosystem dynamics or focusing only on the biophysical processes as a basis for decision making for sustainability both lead to narrow conclusions that may result in unexpected outcomes and even the collapse of SESs, e.g., the Aral Sea. The system therefore needs to be considered as a whole because of the tight couplings among components and across scales (Redman et al. 2004).

Going beyond Collins et al.’s (2011) conceptual framework, we suggest that SES key elements can be coupled into two process-based interacting interfaces, each comprising three core items: the (1) “ecosystem services interface” with functions, goods, and benefits/values; and the (2) “adaptive management interface” with collective action and colearning, multiple resource uses, and practices. Both interfaces are set within a given landscape (Fig. 1). We consider these six core items as leverages influencing the dynamics of the SES, though they differ in scale and nature. The two interfaces and their core coupling elements share characteristics despite having their own variables, methods, analytic tools, vocabulary, and semantics (Abson et al. 2014, Rissman and Gillon 2017). Having many meanings, their use conveys concepts with dialectically vague frontiers. As such, they can be seen as boundary objects that can promote opportunities for transdisciplinarity (Schröter et al. 2014)."

Another model particularly suitable for studying such complex issues involving interactions between social and ecological systems is the the Press Pulse Dynamics Framework (PPD). The PPD framework links the social domain (characterised by socio-economic activities) with the biophysical one (characterised by ecosystem structure and functions) through pulse-press dynamics and ecosystem services. The dynamics within the biophysical domain are driven by “pulse” events or by “press” events that are sustained and chronic. Over time, presses, pulses and press-pulse interactions alter the relationship between the biotic structure and the ecosystem functioning, which in turn affect essential services humans obtain from ecosystems.

As explained in Mirtl et al. (2018):

"The Press Pulse Dynamics Framework conceptual model is dynamic (iterative and including feedbacks), holistic (including both the social systems and biophysical systems including the critical zone), and considers multiple spatial and temporal scales. It can be used to focus on long-term ecosystem, biodiversity, critical zone and social-ecological research agendas through the identification of, and connections among, six strategic research questions (H):

• How do long-term press disturbances and short-term pulse disturbances interact to alter ecosystem structure and function (H1)?

• How can ecosystem structures be both a cause and consequence of ecological fluxes of energy and matter (H2)?

• How do altered ecosystem dynamics affect ecosystem services (H3)?

• How do changes in vital ecosystem services alter human outcomes (H4)?

• How do human behaviors and institutions respond to changes in the provision of ecosystem services (H5)?

• Which human actions influence the frequency, magnitude, or form of press and pulse disturbance regimes across ecosystems and what determines these actions (H6)?"

Wild ungulates at the heart of the functioning of agro-ecosystems

Agro-ecosystems are currently undergoing major changes resulting from both global factors (including climate change, globalization) and local mechanisms (including changes in agricultural practices). Productivist agriculture has reached its limits with yield ceilings, resource depletion, reduction of cultivated biodiversity, environmental degradation and the imperatives of climate change. The challenge now is to move from intensive, production-oriented agriculture to productive, sustainable agriculture, seeking both economic and environmental performance, based on the functionalities offered by ecosystems. The aim is to promote the establishment and maintenance of organisms providing ecosystem services, with a view to optimizing the natural processes that are conducive to agricultural productivity and the resilience of agro-ecosystems. In order to implement this agro-ecological transition supported by strong political will and societal demand, it is therefore necessary to better understand the role of biodiversity in the functioning of agro-ecosystems, and its provision of ecosystem services and disservices to agriculture.

A major challenge of the agro-ecological transition concerns the change of scales of agricultural management.

While agriculture has traditionally been managed at the scale of the agricultural plot, the latter does not constitute an isolated ecosystem within the landscape. It exchanges biotic and abiotic flows with the different natural and semi-natural compartments of the landscape, such as forests, meadows or hedgerows. In order to be able to take into account these flows (for example of carbon, nutrients or seeds) which play a key role in the functioning of agro-ecosystems, it is essential to consider the scale of the landscape mosaic by mobilising the concepts and tools of landscape ecology. In this context, the application of the emerging concept of meta-ecosystem (defined as a set of interactive ecosystems connected by flows of energy, materials and organisms across their boundaries) seems promising to better understand the emerging properties resulting from the spatial coupling of ecosystems within heterogeneous landscapes and to better take into account the diversity of modalities and nature of exchanges between ecosystems (such as water and atmospheric transfers, or animal movements). Indeed, while much attention has been paid to water transfers, the role of wildlife, as a mobile agent, has been largely neglected . Mobile agents are organisms that move between different compartments of the landscape to perform their various daily activities, and through their movements and behaviours will induce biotic and abiotic flows across the landscape, which can be translated into ecosystem services. A conceptual framework has recently been proposed to better understand the effects of landscape structure on the ecosystem services provided by mobile agents. However, this model, originally designed to study pollination, does not explicitly take into account the movements and behaviours of mobile agents across the landscape. The challenge is therefore to mobilize behavioural ecology concepts and tools to explicitly incorporate the movements and behaviours of mobile agents in this model and to better characterize and model the flows induced by mobile agents across the landscape and consequently the spatial distribution of ecosystem services within the landscape mosaic.

Within ordinary biodiversity, wild ungulates, as mobile agents, potentially play a key and unique role in the functioning of agro-ecosystems and can thus provide a number of ecosystem services and disservices to agriculture and forestry.

Indeed, following their recent population explosio, certain mainly forest-based European species, such as deer (Capreolus capreolus) or wild boar (Sus scrofa), have colonized anthropized open environments. Wild ungulates have thus become abundant and widespread in agro-ecosystems, with marked socio-economic consequences in terms of hunting income, damage to crops and forests, collisions with vehicles or transmission of diseases to livestock and zoonoses. Their impact on agro-ecosystems is all the more important as these species use a wide range of habitats, feed on a wide range of wild and cultivated plants and have good dispersal capacities. They are among the most mobile agents in agro-ecosystems, and thus could create unique long-distance flows across the landscape. Wild ungulates are true ecosystem engineers. Through their browsing, scraping, smearing and defiling, they affect the composition, structure and dynamics of vegetation, with a potential negative impact on forest and agricultural productivity. Through their trampling, defecation and even direct consumption by soil fauna, as well as their selective feeding, they also have an indirect effect on decomposition and litter quality and can determine local patterns of productivity and plant diversity. They can also have a significant, if not extremely important, long-term impact on nutrient fluxes and biogeochemical cycles within ecosystems including redistribution of nutrients (N and P) carried in their faeces from nutrient-rich areas (crops) where animals feed to poorer areas (forests) where animals rest and defecate. Finally, they play an important role in seed dispersal by epizoochory (attachment to fleece or hooves) and endozoochory (ingestion-defecation). 20-80% of temperate plant species would thus depend on frugivorous species for seed dispersal. Wild ungulates disperse by endozoochory, preferably species from open habitats and could contribute to modifying local plant distribution patterns, by favouring the intrusion of open habitats species in sufficiently lit forest areas (e.g. edges, clearings, cuttings). All these modifications can have cascading effects on primary production, wildlife habitats, food webs, and consequently on biodiversity, structure and dynamics of animal communities (arthropods, birds, small mammals). Although the impact of wild ungulates on the functioning of forest or savannah ecosystems is no longer in doubt, few studies have focused on their role in the functioning of agro-ecosystems and their provision of ecosystem services and disservices to agriculture and forestry, particularly in relation to the dissemination of seeds or nutrient flows.

The role of wild ungulates in the functioning of agro-ecosystems is largely determined by their spatial behaviour, which is influenced not only by landscape structure and connectivity but also by human activities (e.g. farming practices, hunting, traffic). In these heterogeneous agricultural landscapes, wild ungulates make differential use of different landscape compartments. Deer, for example, may feed in woodland edges and in fields and meadows where they find a nutrient-rich diet, but foraging in remaining woodlands remains marked and deer regularly require closed and sheltered habitats for rumination. Individuals thus move back and forth between different compartments of the landscape with potentially significant consequences on the flow of induced seeds and nutrients. Any substantial change in landscape structure, whether natural or anthropogenic, can have important consequences for the spatial functioning of wild ungulate populations, and hence their impact on the functioning of agro-ecosystems.

The objective of my project that I submitted to INRA was therefore to characterize the role of wild ungulates in the functioning of agro-ecosystems and the ecosystem services and disservices provided to agriculture and forestry. The aim is to study :

Axis 1. The impact of agricultural practices and landscape structure on the spatial behaviour of wild ungulates in agro-ecosystems ;

Axis 2. The consequences of these behaviours on nutrient fluxes and the dissemination of seeds between the different compartments of the landscape ;

Axis 3. Interactions of the services and disservices provided by wild ungulates with other ecosystem services provided in agro-ecosystems (bundles of services).

The results should allow the identification of land-use plans and agricultural practices to optimize the ecosystem services provided by wild ungulates while limiting disservices, for a sustainable and multifunctional management of the territory.

Anthropogenic Landscapes, Spatial Ecology and Activities of Brown Bears & Consequences on Pyrenean Ecosystems

Context: The return of large carnivores in a highly anthropized Europe

While many populations of large carnivores are still in drastic decline around the world due to persecution and destruction of their habitats by humans (Ripple et al. 2014), Europe is currently seeing a return of its grey wolf, brown bear and boreal lynx populations thanks to favourable legislation, reintroduction programmes, increasing populations of wild ungulates, as well as agricultural decline coupled with reforestation (Linnell et al. 2005; Chapron et al. 2014; Crete et al. 2020). Formerly present almost everywhere in Europe, both in plains and mountains, in the 1950s and 1970s these species were confined essentially to the wildest regions of Europe, thus conforming to the model of human-wildlife separation or land sparing (Chapron et al. 2014; Fisher et al. 2014; López-Bao et al. 2017; Crete et al. 2020). However, the recent increase in European populations of large carnivores is pushing them to progressively recolonise their historical range, areas that have been profoundly marked by the imprint of man over the last two centuries (Linnell et al. 2005). Given the importance of anthropisation and the fragmentation of European landscapes (Linnell et al. 2005; Selva et al. 2011; Venter et al. 2016), it seems difficult to aim for the recovery of viable populations of large carnivores in Europe (or populations "in a favourable conservation status" as desired by the Habitats-Fauna-Flora Directive), without moving to a model of coexistence or land sharing (Carter & Linnell 2016; López-Bao et al. 2017; Crete et al. 2020; Linnell et al. 2005, 2020).

But coexistence between large carnivores and humans is a real challenge. On one hand, transport infrastructures and soil artificialisation may constrain the movements of large carnivores (e.g. Basille et al. 2013; Tucker et al. 2018; Rio-Maior et al. 2019), or even constitute real barriers to their dispersal, limiting connectivity between sub-populations, gene flows and undermining the demo-genetic viability of small populations (e.g. Riley et al. 2006; Morales-González et al. 2020). Disturbance related to human activities may also affect habitat selection and activities of large carnivores and of their preys, and consequently affect their demographic parameters (e.g. Swenson et al. 1997; Martin et al. 2010; Ordiz et al. 2013; Morales-González et al. 2020). On other hand, the spatial overlap between large carnivores and human activities results in a wide range of interactions, classified either as ecosystem services or disservices, depending on the views, interests and values of the different stakeholders who perceive this relationship (Linnell et al. 2020). In particular, it generates conflicts related to the depredations caused by large carnivores on domestic livestock; damages that are often contrasted, according to an anthropocentric vision, with the numerous socio-economic or even ecological services provided by livestock farming (Linnel et al. 2005).

The interest in restoring viable populations of large carnivores and conserving these species is in fact questioned by some local stakeholders and is the subject of lively debate between naturalists, elected representatives, breeders, hunters, and the general public (Linnel et al. 2005). Reintroductions of large carnivores are generally motivated by the imperative to restore populations and more generally the functioning of ecosystems by precipitating trophic cascades that release species at the base of food webs (Wolf & Ripple 2018). But even within the scientific community, the interest in rewilding and more particularly the role of large carnivores in ecosystem functioning is a subject of controversy (Kuijper et al. 2016; Perino et al. 2019).

Issue: Controversy over the role of large carnivores in anthropized landscapes

Due to their position at the top of the food chain, large carnivores are often considered as keystone species (e.g. Linnell et al. 2000; del Rio et al. 2001; Helfield & Naiman 2006), i.e. they have a greater effect on community structure and ecosystems than expected given their biomass (Paine 1969, 1995; Power et al. 1996). Classically, tri-trophic cascades, i.e. the descending effects of large carnivores on the food chain via predation on large herbivores and the release of herbivory, have been put forward (Kuijper et al. 2016). However, the effects of large carnivores on ecosystems could be much more complex than initially envisaged.

Firstly, despite decades of research, formal evidence of the existence of such trophic cascades induced by large carnivores remains scarce (Polis et al. 2000; Ford & Goheen 2015; Ripple et al. 2016) and is often the subject of controversy due in particular to the lack of replication and control (see the case of the reintroduction of the wolf in Yellowstone or the disappearance of the dingo in Australia; Ford & Goheen 2015). In addition, it seems that trophic cascades that were once considered to be primarily the result of prey consumption effects (regulating the abundance of prey populations through predation) often turn out to be related to non-consumptive or behavioral effects of predation (related to the landscape of fear; Werner & Peacor 2003; Peckarsky et al. 2008; Suraci et al. 2016).

Furthermore, we lack knowledge on the functional role of large carnivores in anthropised ecosystems such as those encountered in Europe (Kuijper et al. 2016). However, in anthropogenic landscapes, humans could affect the role of large carnivores in ecosystems by acting as hyper keystone species (Worm & Paine 2016) and by influencing the density, behaviour, habitats and/or resource landscape of large carnivores, mesocarnivores and their prey (Kuijper et al. 2016). According to Kuijper et al. (2016), the potential of density-mediated trophic cascades may be limited in human-induced landscapes to unproductive areas where even small numbers of carnivores may impact prey density or in limited areas of the landscape where humans allow carnivores to reach ecologically functional densities. Thus, according to Linnell et al (2005), given the constraints on space, habitat and human tolerance, it seems unlikely that Europe will reach a stage where the abundance of carnivores and their prey is mainly determined by trophic interactions and non-human factors. However, the potential for behavioural meditative trophic cascades may be greater and more widespread, as even low densities of carnivores can affect prey behaviour (Kuijper et al. 2016).

Finally, the tri-trophic cascades are not the only ways in which large carnivores could influence ecosystems. Their cascade effects could also spread through their interactions with other (large, meso and small) carnivores via competition, predation or facilitation (Prugh et al. 2009; Ritchie & Johnson 2009; Prugh & Sivy 2020), but also parasites and scavengers . They could also act as mobile agents and generate biotic and/or abiotic flows across the landscape (Kremen et al. 2007). In particular, the ecological role of large carnivores with an omnivorous diet (e.g. brown bears) is much less well known than that of so-called carnivorous species. However, the type of diet of large carnivores could profoundly affect the role of these keystone species in ecosystems.

We might expect the cascading effects induced by large carnivores on the omnivorous diet to take more varied routes and to be able to maintain lower population densities and in landscapes that are more anthropogenic than those induced by large carnivores with a mainly carnivorous diet (which act a priori mainly via tri-trophic cascades).

Role of brown bears in ecosystems

The brown bear is often referred to as a keystone species due to its position at the top of the food chain (e.g., Ucarli 2011; Palazon 2017). By definition, this supposes that it has a greater effect on community structure and ecosystems than expected given its biomass (Paine 1969, 1995; Power et al. 1996). Classically, tri-trophic (carnivore-herbivore-plant) cascades have been put forward to justify this name in large carnivores (e.g. Kuijper et al. 2016; Suraci et al. 2016). Carnivores can indeed affect the abundance (lethal effects) and/or behavior (non lethal effects through landscape of fear) of large herbivores through their predation, resulting in a relaxation of herbivory and consequences for plant community structure, biodiversity, and ecosystem functioning (Ripple et al. 2001, 2014). However, recent studies show that the role of large carnivores in ecosystems is more complex than initially envisioned. Their cascading effects on ecosystems could also happen through their interactions with other (large, meso, and small) carnivores, through their predation on other carnivore species (intra-guild predation), a decrease in the abundance of prey available to other carnivores (interspecific competition by exploitation), and/or control of carnivore access to their usual prey through aggression, theft, and territoriality (interscpecific competition by inteference) (Brashares et al. 2010; Elmhagen et al. 2010; Prugh & Sivy 2020). The case of the brown bear is somewhat unique among large carnivores in that it is an omnivorous species. The proportion of animal preys in its diet is low. In the Pyrenees, ungulates (wild or domestic) and micro-mammals represent only 10 to 15% of its diet (Couturier 1954; Berducou et al. 1982). It therefore seems unlikely that the effects of brown bear predation would generate tri-trophic cascades as important as in other large carnivore species with essentially carnivorous diets (e.g. wolf, lynx). In such an omnivorous species, cascading effects on ecosystems would be expected to occur through much more varied pathways (e.g., endozoochory, myrmecophagy, herbivory, scavenging.

The brown bear (like other large carnivore species) is a facilitator species. Through its predations, it can promote the availability of carrion in the environment, which can be beneficial to other carnivore species that are occasionally scavengers (e.g., other large carnivores, mustelids, foxes), but also more generally to all scavenging species (e.g., insects, vultures, eagles, eagle owls, corvids, wild boars, jays, and even certain rodents).

The brown bear is considered an ecosystem engineer. Indeed, they create, modify, or maintain habitats by causing changes in the physical state of biotic or abiotic materials that directly or indirectly modulate the availability of resources for other species (Jones et al. 1997). In particular, wounds generated by bears to trees by debarking, either as part of feeding (to feed on the bast) or as part of marking, can provide breeding, refuge, and/or feeding sites for a variety of species (e.g., saproxylic insects, woodpeckers, saproxylic fungi, amphibians, and small mammals; Zyśk-Gorczyńska et al. 2015).

The brown bear is a perfect candidate to play the role of mobile-link species (also known as mobile agents, dispersal vectors or active subsidies; Lundberg & Moberg 2003; Kremen et al. 2007; Jeltsch et al. 2013; Earl & Zollner 2017). By their movements and behaviours, it can by definition generate biotic and abiotic flows across the landscape. Indeed, it has an opportunistic omnivorous diet with a clear vegetative dominance (~70-80% including notably dried fruits, fleshy fruits, grasses and fungi). It is very mobile and can travel more than 10 km per day. It occupies vast home ranges that can reach several hundred km² annually (especially in males). Finally, it uses different types of habitats for its daily needs (resting, feeding, movements) and seasonal needs (hibernation, reproduction, rearing of young). In particular, brown bears contribute to the long-distance dissemination of seeds by endozoochory (consumption of fruits or seeds - partial digestion - defecation) and/or epizoochory (attachment of seeds to the fleece) (e.g. Lalleroni et al. 2016 in the Pyrenees; Willson & Gende 2004; Nowak & Crone 2012; Harrer & Levi 2018; Karimi et al. 2018). Through cascading effects, it can thus influence the dynamics of plant and animal populations and communities. Many seeds indeed remain intact despite their passage through the digestive tract of bears, and the germination of some seeds may even be stimulated by passage through the bear digestive tract (e.g., blueberry seeds; Steyaert et al. 2019). As a consumer of fungi, brown bears could potentially also disperse mycorrhizal spores across the landscape. In addition, large carnivores may enrich some soils with nutrients through the deposition of feces, urine, and carcasses and participate in nutrient flows across the landscape. In North America, Grizzly bears, through their consumption of salmon, are responsible for significant nutrient flows between the ocean and terrestrial ecosystems (Hilderbrand et al. 1999). After capturing salmon in estuaries and streams, Grizzly bears typically move to shore to consume each fish on land, distributing the available carcass remains to necrophagous vertebrates and invertebrates up to several hundred meters from waterways (Gende et al. 2001; Gende & Quinn 2004). Carcass remains can in turn influence all trophic levels, from primary producers to large carnivores, in both terrestrial and aquatic ecosystems (Ben-David et al. 1998; Helfield & Naiman 2001, 2006; Hocking & Reynolds 2011). Finally, large carnivores can participate in the spread of parasites across the landscape, such as ticks attached to their fleeces. They may also affect disease prevalence (e.g., zoonotic and livestock diseases; Packer et al. 2003, Ostfeld & Holt 2004) by affecting the abundance and behavior of their prey (rodents, ungulates, birds) that potentially carry parasites and diseases (Packer et al. 2003, Ostfeld & Holt 2004).

Brown bears are considered an umbrella species (Ucarli 2011), requiring large areas covering different habitat types to survive. By protecting this species and its habitats, a range of other species living within these habitats can be protected as well (e.g., capercaillie).

The brown bear is finally a flagship species (Ucarli 2011), with a charismatic, emblematic character and high heritage value (especially in the Pyrenees). It serves as a symbol for the conservation and enhancement of biodiversity and ecosystems and can thus constitute an asset for economic and social development that is integrated into a rich and preserved environment.

Study species: The Brown Bear in the Pyrenean Anthroposystem

While they are important reservoirs of biodiversity providing numerous ecosystem services to humans, mountain ecosystems have been undergoing profound changes for nearly a century, due to ongoing changes in their uses (notably the explosion of tourism and nature sports) and their vulnerability to climate change (with temperatures rising on average higher than elsewhere) (IUCN France 2014, 2015). In the Pyrenees, the situation is particularly alarming, as highlighted in the OPCC report published in 2018 and entitled "Climate Change in the Pyrenees: Impacts, Vulnerabilities and Adaptation". The temperature rise of +1.2°C observed between 1949 and 2010 is 30% higher than the world average, and is reflected among other things by an upward shift of the different vegetation stages (+ 3 m / year of plants for more than 30 years; OPCC 2013). These climatic disturbances, combined with changes in use, constitute a serious threat to biodiversity and Pyrenean ecosystems, which are home to numerous endemic (including 150 to 200 plant species) and emblematic species (e.g. Capercaillie, Pyrenean Desman, Brown Bear, Iberian Ibex, Isard, Bearded Vulture), as well as areas of interest (e.g. old-growth forests which account for 2% of the forested area). But their impacts are not limited to the biophysical sectors. All the services provided by Pyrenean ecosystems and the socio-economic sectors that depend on them (e.g. tourism, agro-pastoralism, energy, natural hazards) could also suffer the consequences (OPCC 2018).

Within Pyrenean biodiversity, the brown bear holds a very special place due to its patrimonial character, the conservation status (critically endangered) and vulnerability of its population (isolated, highly inbred and fragmented), its role as an umbrella species (its spatial requirements and habitat needs are such that the protection of the species de facto implies that of all other species present, and thus by extension the safeguarding of entire sections of ecosystems and landscapes) and a keystone species of large carnivore with an omnivorous diet (with a high potential for trophic cascades and mobile agents), and the high political-socio-economic stakes linked to its presence (charismatic, crystallizing by its depredations the difficulties of agropastoralism). The conservation and strengthening of this emblematic species in the Pyrenees is thus the subject of lively debate. What would happen if the ursine population were to disappear from the massif? What could limit the viability of this population in the short, medium and long term? What are the impacts of anthropic disturbances on this population and how can these impacts be mitigated? How do bears affect the biodiversity and ecosystems of the Pyrenees ? What services and disservices can humans derive from them? Could bears promote the resilience of the Pyrenean ecosystems to climate change ? There is thus a real societal demand and a real political questioning on the interest of conserving and strengthening the Pyrenean Brown Bear population beyond its intrinsic value.

In order to meet this demand and promote the coexistence between human activities and large carnivores, it seems essential to better understand the ecological, societal and economic role played by the Pyrenean brown bear in the Pyrenees, as well as its interactions with humans. To do this, it would be relevant to adopt an integrative conceptual framework adapted to the study of the complex dynamics of interactions between man and nature. That of socio-ecological systems has largely proved its worth in the framework of Workshop Zones, human-environment observatories or LTER sites and has proven its operational efficiency in resolving conflicts linked to interactions between social and ecological systems (Liu et al. 2007; US-LTER 2007; Bretagnolle et al. 2019). Through this framework, the aim is to redefine ecosystems by explicitly including humans in food webs as : (i) an agent modifying the environment and affecting species and their role in ecosystems, as well as (ii) a beneficiary of goods and services provided by nature.

The aim of this project is thus to carry out an action-research anchored in the Pyrenean territory and aiming at the sustainability and resilience of the socio-ecological systems of this massif, with a special focus on the problem of brown bear conservation in the Pyrenees.

This project aims to study the spatio-temporal dynamics of the Pyrenean socio-ecological system under the prism of the Brown bear using an integrative and interdisciplinary approach. The aim is in particular to characterize the role of the Brown bear in the functioning of Pyrenean ecosystems, and to assess how human activities can affect this role and what the consequences are in terms of the provision of ecosystem services and disservices and the perception of bears by humans. This information is essential to better conserve and sustainably manage the Pyrenean Brown bear population and its habitats, and more broadly the biodiversity of Pyrenean mountain ecosystems, in coexistence with human activities and in a context of global changes and the gradual arrival of other large carnivores (colonisation by wolves and the lynx reintroduction project in Catalonia).