REVIEW SUMMARY

REWILDING

Rewilding complex ecosystemsAndrea Perino Henrique M Pereira Laetitia M Navarro Neacutestor FernaacutendezJames M Bullock Silvia Ceausu Ainara Corteacutes-Avizanda Roel van KlinkTobias Kuemmerle Angela Lomba Guy Persquoer Tobias Plieninger Joseacute M Rey BenayasChristopher J Sandom Jens-Christian Svenning Helen C Wheeler

BACKGROUNDRapid global change is creat-ing fundamental challenges for the persistenceof natural ecosystems and their biodiversityConservation efforts aimed at the protection oflandscapes have had mixed success and thereis an increasing awareness that the long-termprotection of biodiversity requires inclusionof flexible restoration along with protectionRewilding is one such approach that has beenboth promoted and criticized in recent yearsProponents emphasize the potential of rewild-ing to tap opportunities for restoration whilecreating benefits for both ecosystems andsocieties Critics discuss the lack of a con-sistent definition of rewilding and insufficientknowledge about its potential outcomes Othercriticisms arise from the mistaken notion thatrewilding actions are planned without consid-ering societal acceptability and benefits Herewe present a framework for rewilding actionsthat can serve as a guideline for researchersand managers The framework is applicableto a variety of rewilding approaches rangingfrom passive to trophic rewilding and aims

to promote beneficial interactions betweensociety and nature

ADVANCES The concept of rewilding hasevolved from its initial emphasis on protectinglarge connected areas for large carnivore con-servation to a process-oriented dynamic ap-proach On the basis of concepts from resilienceand complexity theory of social-ecological sys-tems we identify trophic complexity stochasticdisturbances and dispersal as three critical com-ponents of natural ecosystem dynamics Wepropose that the restoration of these processesand their interactions can lead to increasedself-sustainability of ecosystems and shouldbe at the core of rewilding actions Building onthese concepts we develop a framework todesign and evaluate rewilding plans Alongsideecological restoration goals our frameworkemphasizes peoplersquos perceptions and exper-iences of wildness and the regulating andmaterial contributions from restoring natureThese societal aspects are important outcomesand may be critical factors for the success of

rewilding initiatives (see the figure) We fur-ther identify current societal constraints onrewilding and suggest actions tomitigate them

OUTLOOK The concept of rewilding chal-lenges us to rethink the way wemanage natureand to broaden our vision about how naturewill respond to changes that society brings both

intentionally and unin-tentionally The effects ofrewilding actions will bespecific to each ecosystemand thus a deep under-standing of the processesthat shape ecosystems is

critical to anticipate these effects and to takeappropriatemanagement actions In additionthe decision of whether a rewilding approach isdesirable should consider stakeholdersrsquoneeds andexpectations To this end structured restorationplanningmdashbasedonparticipatoryprocesses involv-ing researchers managers and stakeholdersmdashthat includes monitoring and adaptive manage-ment can be usedWith the recent designation of2021ndash2030 as the ldquodecade of ecosystem restora-tionrdquo by the United Nations General Assemblypolicy- anddecision-makers couldpush rewildingtopics to the forefront of discussions about howto reach post-2020 biodiversity goals

RESEARCH

Perino et al Science 364 351 (2019) 26 April 2019 1 of 1

The list of author affiliations is available in the full article onlineCorresponding author Email andreaperinoidivde (AP)hpereiraidivde (HMP)Cite this article as A Perino et al Science 364 eaav5570(2019) DOI 101126scienceaav5570

Dispersal

Stochastic disturbancesTrophic complexity

Environmental risks

Disp

ersa

l-lim

iting

infra

stru

ctur

e

Human-wildlife conflicts

MaterialRegulatingNonmaterial

Bene

fitLo

ss

Ecological state Contributions from natureSocietal outcomes can be assessed by mapping positive and negative impacts on nonmaterial regulating and material contributions from nature

Restoration of ecological processes can positively influence their interactions ndash eg species diversity and trophic complexity can be increased if dispersal to new ecosystems is possible

Rewilding actions and outcomes are framed by societal and ecologi-cal context Rewilding can be assessed by representing the state ofecosystems in a three-dimensional space where each dimensioncorresponds to an ecological process The difference in volume betweenthe restored (yellow pyramid) and the degraded ecosystem (orangepyramid) is a proxy for the effects of rewilding on the self-sustainability of

the ecosystem The dashed line within the yellow pyramid representsthe societal boundaries that determine to what extent ecologicalprocesses can be restored Rewilding actions can help push societalboundaries toward the ecological potential (orange arrows) bypromoting societal support and opportunities for people to experiencethe autonomy of ecological processes in enjoyable ways

TOMORROWrsquoS EARTHRead more articles onlineat scimagTomorrowsEarth

ON OUR WEBSITE

Read the full articleat httpdxdoiorg101126scienceaav5570

on March 3 2020

httpsciencesciencem

agorgD

ownloaded from

REVIEW

REWILDING

Rewilding complex ecosystemsAndrea Perino12 Henrique M Pereira123 Laetitia M Navarro12Neacutestor Fernaacutendez12 James M Bullock4 Silvia Ceausu56 Ainara Corteacutes-Avizanda378Roel van Klink1 Tobias Kuemmerle9 Angela Lomba3 Guy Persquoer110Tobias Plieninger1112 Joseacute M Rey Benayas13 Christopher J Sandom14Jens-Christian Svenning56 Helen C Wheeler15161718

The practice of rewilding has been both promoted and criticized in recent years Benefitsinclude flexibility to react to environmental change and the promotion of opportunitiesfor society to reconnect with nature Criticisms include the lack of a clear conceptualizationof rewilding insufficient knowledge about possible outcomes and the perception thatrewilding excludes people from landscapes Here we present a framework for rewilding thataddresses these concernsWe suggest that rewilding efforts should target trophic complexitynatural disturbances and dispersal as interacting processes that can improve ecosystemresilience and maintain biodiversityWe propose a structured approach to rewilding projectsthat includes assessment of the contributions of nature to people and the social-ecologicalconstraints on restoration

Shifting societal and environmental condi-tions including land-use change and increas-ing demand for resources are acceleratingbiodiversity loss and ecosystemdegradation(1ndash4) The associated loss of many impor-

tant ecological processes (5 6) can decrease thecomplexity and resilience of ecosystems by ham-pering their capacity to recover from perturba-tions (7ndash9) Although responses to the biodiversitycrisismdashespecially the establishment of protectedareasmdashhave reduced biodiversity loss in someinstances (10ndash12) reports of ineffective protectedareas (13) and ongoing declines of threatenedspecies (14) indicate that conservation strategiesmust go beyond current efforts (15 16)A growing body of literature emphasizes the

need for novel process-oriented approaches torestoring ecosystems in our rapidly changingworld (4 17ndash19) Dynamic and process-orientedapproaches focus on the adaptive capacity ofecosystems (4) and the restoration of ecosystemprocesses promoting biodiversity rather thanaiming to maintain or restore particular eco-system states characterized by predefined speciescompositions or particular bundles of ecosystemservices Such approaches recognize ecosystemsas dynamic systems (20) whose future develop-ment cannot always be predicted (21 22)

Rewilding is one such approach to restorationThis strategy aims to restore self-sustaining andcomplex ecosystems with interlinked ecologicalprocesses that promote and support one anotherwhile minimizing or gradually reducing humaninterventions (23ndash25) Rewilding also emphasizesthe emotional experience and perception of wildnature andwild ecosystemswithout human inter-vention (26) Although conventional restorationprojects often aim to minimize human interven-tion many scientists and practitioners considersome level of management as critical to replaceecosystemprocesses that have been lost because ofhuman activities or tomaintain important aspectsof cultural landscapes (27) Such managementoften focuses on selected processes via preciselydefined actions targeting concrete goals [egman-agement of Satoyama landscapes in Japan (28)]Rewilding on the contrary recognizes andworkswith complexity and autonomy as ecosystem-inherent characteristics and acknowledges theirdynamic unpredictable nature (29)Despite the potential for rewilding to address

pressing restoration challenges critics havepointed out several shortcomings that have asyet hampered the application of rewilding prin-ciples Criticism includes a lack of a consistentdefinition of rewilding (30) and insufficient

knowledge about the possible outcomes of re-wilding endeavors (31) In addition concernshave been raised about rewilding activities beingplanned in amanner that excludes people fromlandscapes rather than being designed withlocal support (32)Here we articulate a conceptual framework

for rewilding projects that addresses the afore-mentioned criticisms We start by briefly review-ing the history of the rewilding concept from itsinitial emphasis on protecting large connectedareas for carnivore conservation (33) to the diver-sity of rewilding concepts today (25) We proposea framework to design and evaluate rewildingplans that integrates the current variety of rewild-ing approaches Our framework draws on ecolog-ical theory to identify three interacting ecologicalprocesses that promote the self-organization ofecosystems and therefore should be the focus ofrewilding actions For each of these processes wereview ecological knowledge and identify rewild-ing actions that can assist the restoration of self-sustaining resilient ecosystems (Fig 1) Notablythese actions will vary depending on the societalcontext Rewilding can occur spontaneously ifhumans withdraw from landscapesmdashfor exam-ple after agricultural abandonment (34ndash36) orin areas that have become inhospitable as a resultof armed conflict (37ndash39) or environmental catas-trophes such as the Chernobyl disaster (40 41)In other cases rewilding projects are driven byactive choices about how societies want to ex-perience nature (42) and the extent towhich theycan accept an autonomy of natural processes Inthese cases the feasibility of rewilding projectsalso depends on material nonmaterial and reg-ulating contributions from nature (Fig 2) Wediscuss how rewilding projects need to accountfor social-ecological dynamics by addressingpeoplersquos preferences as well as human effectson ecosystems Finally we apply our frameworkto a set of ongoing rewilding projects and illus-trate how interactions among the key processescan be promoted to increase both ecosystemresilience and societal benefits

A brief history of the rewilding concept

Rewilding as it was originally conceived 20 yearsago (33) referred to ldquothe scientific argument forrestoring big wilderness based on the regulatoryroles of large predatorsrdquo (33) that could act askeystone species andmaintain the resilience anddiversity of terrestrial ecosystems through top-down control (33 43) The protection and restora-tion of ldquolarge strictly protected core reserves

RESEARCH

Perino et al Science 364 eaav5570 (2019) 26 April 2019 1 of 8

1German Centre for Integrative Biodiversity Research (iDiv) Halle-Jena-Leipzig Leipzig Germany 2Institut fuumlr Biologie Martin-Luther-University Halle-Wittenberg Halle Germany 3CIBIO(Research Centre in Biodiversity and Genetic Resources)ndashInBIO (Research Network in Biodiversity and Evolutionary Biology) Universidade do Porto Vairatildeo Portugal 4NERC Centre for Ecologyand Hydrology Wallingford UK 5Section for Ecoinformatics and Biodiversity Department of Bioscience Aarhus University Aarhus Denmark 6Center for Biodiversity Dynamics in a ChangingWorld (BIOCHANGE) Aarhus University Aarhus Denmark 7Animal Ecology and Demography Unit IMEDEA (CSIC-UIB) Balearic Islands (Mallorca) Spain 8Department of Conservation BiologyEstacioacuten Bioloacutegica de Dontildeana (CSIC) Seville Spain 9Geography Department and Integrative Research Institute for Transformations in Human-Environment Systems (IRI THESys) HumboldtUniversity of Berlin Berlin Germany 10Department of Economics and Department of Ecosystem Services Helmholtz-Zentrum fuumlr Umweltforschung UFZ Leipzig Germany 11Faculty of OrganicAgricultural Sciences University of Kassel Kassel Germany 12Department of Agricultural Economics and Rural Development University of Goumlttingen Goumlttingen Germany 13Department of LifeSciences University of Alcalaacute Alcalaacute de Henares Spain 14School of Life Sciences University of Sussex Brighton UK 15Department of Biology Anglia Ruskin University Cambridge UK 16CentredrsquoEacutecologie Fonctionnelle et Evolutive Centre National de la Recherche Scientifique Paris France 17Department of Arctic and Marine Biology UiT Norges Arktiske Universitet Tromsoslash Norway18Department of Biology Chemistry and Geography Universiteacute du Quebec agrave Rimouski Rimouski Quebec CanadaCorresponding author Email andreaperinoidivde (AP) hpereiraidivde (HMP)

on March 3 2020

httpsciencesciencem

agorgD

ownloaded from

connectivity and keystone speciesrdquo (33) were thecentral characteristics of this first definition ofrewilding Although the conservation of largecarnivores and their habitats is still an importantaspect of rewilding (25 44) the concept hasevolved from this original idea to include a rangeof diverse approaches (25) Trophic rewildingperhaps the closest to the original concept ad-vocates the reintroduction of missing keystonespecies such as large carnivores and large her-bivores Trophic rewilding often promotes theuse of functional replacementsmdashie the intro-duction of non-native species as ecological proxiesfor species that became extinct centuries or mil-lennia ago (25 32 44) A particular type of tro-phic rewilding is Pleistocene rewilding whichaims to restore ecosystems that include and areshaped by populations of megafauna extirpatedsince the Late Pleistocene taking a long-termevolutionary perspective on biodiversity and eco-systems (44) In contrast ecological (or passive)rewilding emphasizes the passive managementof ecological succession in abandoned landscapes

Passive rewilding actions include the creation ofno-hunting areas low-intervention forestry ma-nagement setting aside agricultural land the re-moval of dispersal barriers and the restorationof natural flood regimes (22 25 34)The ecosystem features that rewilding aims

to restore are characteristic of but not restrictedto wilderness areas (45 46) Instead we refer towildness which is the autonomy of natural pro-cesses (46 47) that can occur in a variety of set-tings and across spatial scales The restoration ofwildness rather than wilderness is thus the keygoal of rewilding efforts Broadening the originaldefinition of rewilding and articulating the re-storation of wildness rather than wilderness asits central goal makes rewilding applicable acrossspatial scales and adaptable to a wide range ofsocietal and landscape contexts fromurban greenspaces to abandoned agricultural landscapes (29)

A theoretical framework for rewilding

In many ecosystems complexity and resilienceare maintained by trophic complexity natural

disturbances and dispersal (48 49) (Fig 1)Human activities often lead to degradation inone or more of these ecological processes Re-wilding aims to restore these three ecologicalprocesses to foster complex and self-organizingecosystems that require minimum human man-agement in the long term (50) If missing ordegraded ecosystem processes are not expectedto recover (on policy-relevant time scales) with-out assistance rewilding may encompass initialinterventions sometimes followed by continuousminimal management In the following para-graphs we explain each of the processes in detailelaborate on how interactions among them canpromote biodiversity and ecosystem resilienceand illustrate how rewilding can be used to restoreand promote such interactions

Trophic complexity

Species at higher trophic levels are often highlyconnected and functionally important to ecosys-tems (Fig 1) (51) Large-bodied herbivores exertstrong influences on the diversity and abundanceof other taxa such as birds small mammals in-sects (52 53) and plants (54 55) These effectsoccur through direct pathways such as the pro-visioning of dung and carrion (56) or the fa-cilitation of dispersal (54 55) but also throughthemodification of the physical environment forinstance by grazing and trampling or the build-ing of dams by beavers (53 57) Large carnivorescan through predation affect population sizesand behavior of herbivores and create spatio-temporal heterogeneity in these processes In theabsence of top-down control by carnivores highdensities of large herbivores canhave detrimentaleffects on the abundance and diversity of otherspecies groups (52 53)Humans cause changes in species composition

and alter species interactions through huntingharvesting or planting selected species in agri-culture and forestry (Fig 1A) Large vertebratesare particularly susceptible to human-driven de-faunation owing to their body size long reproduc-tive cycles and highmetabolic demands leadingto the need for large foraging ranges (58ndash62)Thus even where large vertebrates are still pres-ent in human-dominated landscapes they mightnot be able to exert the top-down control theyhave in wild ecosystems owing to their reduceddensities (63 64) Selective defaunation of toppredators and large herbivores can result in tro-phic cascading effects and greater susceptibility ofecosystems to collapse (51 65)Rewilding can enhance trophic complexity

through a variety of actions that depend on thecharacteristics of the ecosystem Passive rewild-ing measures can for example include the crea-tion of no-hunting areas Where spontaneousrecolonization is unlikely the restoration of tro-phic complexity may also be achieved by trans-locating species Introductions of ecologicalreplacements can be an option if species havegone extinct globally (44) However such replace-ments can entail unforeseeable uncertainties andecological risks and should be assessed withcaution (25) Rewilding can also be supported by

Perino et al Science 364 eaav5570 (2019) 26 April 2019 2 of 8

During disturbance Recovery from disturbanceA

B

Trophiccomplexity DispersalDisturbance

Fig 1 Promoting interactions among ecosystem processes enhances resilience of rewildingareas (A) Intensively managed areas are often characterized by decreased trophic complexityDispersal barriers between ecosystems impede the movement of individuals particularly at highertrophic levels Natural disturbances are often suppressed or altered in magnitude and frequencypotentially leading to even larger disturbance events Impoverished trophic networks dispersalbarriers and deterministic disturbances can hamper recovery of depressed populations (open nodesin the trophic webs) after major disturbance events (B) Rewilded areas have restored complextrophic webs with functional roles for top predators (red nodes) herbivores (yellow nodes) and primaryproducers (blue nodes) Improved connectivity among habitats allows for dispersal of species at alltrophic levels Frequent disturbance events occur in the landscape Dispersal among habitats aidsrecovery of ecosystems after disturbance events by allowing recolonization and population recovery ofaffected species Large vertebrates in complex ecosystems often act as dispersal agents for plants andcan introduce stochasticity into a system eg through predation or grazing

RESEARCH | REVIEWon M

arch 3 2020

httpsciencesciencemagorg

Dow

nloaded from

Perino et al Science 364 eaav5570 (2019) 26 April 2019 3 of 8

Fig 2 Restored ecological processes and their influence oncontributions from natureThe ecological state of each case study isrepresented in a three-dimensional space with one axis for each ecologicalprocess of our framework (trophic complexity dispersal and stochasticdisturbances) The initial ecological state is represented by the orangepyramids whereas the yellow pyramids represent the ecological stateafter rewilding actions The bar plots indicate the number of contributionsto people (42) that are positively or negatively affected by rewildingactions (A) Rewetting of a river branch in the Leipziger Auwald (Germany)led to increases in flood-tolerant species and an overall increase in speciesrichness in several taxa Management actions increased the provision ofnonmaterial services (eg opportunities for learning and inspiration)and regulating services (eg habitat creation and maintenance) Impactson material services are negligible as the project neither affects largeagricultural areas nor substantially improves nature-based incomeopportunities (B) Nonmanagement a hunting-ban and reintroductions

improved trophic complexity and stochastic disturbance in the SwissNational Park Management actions promoted economic prosperity(positive material contributions) and agricultural abandonment (negativematerial contributions) The park provides nonmaterial and regulatingcontributions (eg opportunities for nature experiences habitat creationand maintenance) (C) Reintroductions of mammals to Tijuca NationalPark (Brazil) improved ecological interactions The parkrsquos urban locationlimits the restoration potential of all three processes Managementactions may increase material contributions (ie income generationthrough ecotourism) Nonmaterial contributions (eg supporting identi-ties or maintenance of options) can potentially emerge from community-based projects (D) Land abandonment protection and reintroductionsled to the recovery of the large mammal community in the Chernobylexclusion zone (Belarus) Positive regulating nonmaterial and materialcontributions include habitat creation and maintenance as well asopportunities for learning inspiration and wildlife tourism

RESEARCH | REVIEWPHOTOS(C

)BRIAN

GRATWIC

KEW

IKIM

EDIA

COMMONS(D

)MAXMISCHEKA

DOBESTOCK

on March 3 2020

httpsciencesciencem

agorgD

ownloaded from

activities to promote coexistence between peopleand wildlifemdasheg through compensation schemesfor crop or livestock damage (66 67)

Stochastic disturbances

Natural disturbances often occur in a stochasticmanner at different locations magnitudes andfrequencies enhancing spatial and temporal het-erogeneity in ecosystems (48) These perturbationscan trigger reorganization and reconfiguration ofecosystems (68) and may lead to increased eco-system complexity They promote coexistenceas there is often a trade-off in speciesrsquo compe-titive abilities and resilience to events such asfires floods or pest outbreaks (68) Species thatare able to survive disturbances act as biologicallegacies that promote recovery and reorganiza-tion (eg seed banks or small mammals surviv-ing a fire) (48)In human-dominated landscapes natural dis-

turbances are often suppressed (eg fire sup-pression or flood regulation) or altered in theirmagnitude and frequency (Fig 1A) which maylead to even larger and potentially devastatingdisturbance events (eg large wildfires ratherthan smaller andmore frequent ones) Insteadstochastic disturbances are replaced by predict-able and constant disturbances [eg use of fer-tilizers and irrigation to maintain constant inputsto ecosystems or annual soilmobilization toweedout competing species (48)] Because these deter-ministic disturbances often act in the same placeover a long period of time without a chance forthe affected ecosystem to recover and reorganize(68) sensitive species may be lost (1) Moreoverhuman efforts to repair damage after natural dis-turbance events can remove biological legacies(48 68) and lead to additional perturbations thathinder natural regeneration and reorganizationprocesses (69) For example salvage logging toremove dead trees after wind throw or pest out-breaks often eliminates important resources andhabitats for saproxylic beetles or cavity-nestingspecies (70)Rewilding actions aim to release ecosystems

from continued and controlled anthropogenicdisturbances to allow for natural variability andsources of stochasticity (71) (Fig 1B) Mowing ofgrassland can be reduced or replaced by naturalgrazing Dams can be removed or their manage-ment modified to restore natural flood regimes(72) Logging can be replaced by allowing naturalfire and pest regimes

Dispersal

Populations depend on dispersal among habitatsto avoid overcrowding (73) intraspecific com-petition and loss of genetic diversity (74) Theexchange of individuals from different popula-tions can increase gene flow mitigate inbreed-ing and hence lead to more viable populations(75) Habitat degradation or anthropogenic dis-persal barriers reduce habitat connectivity anddispersal ability (Fig 1A)A rewilding approach includes the improve-

ment of connectivity within and among ecosys-tems to promote dispersal Whereas connectivity

efforts often focus on corridors alone a multi-scale approach should seek to identify and linkopportunities ranging from local features suchas hedgerows to support birds or insects (76) tolarge-scale corridors which allow recolonizationby large mammals over long distances Con-nectivity can also be improved by removing orincreasing the permeability of dispersal barriers(Fig 1B) such as roads dams or fences The per-meability of unsuitable habitat particularly homo-geneous agricultural areas can be improved by theintroduction of natural landscape elements (77)

Integrating across ecological processes

The three ecological processes can influence andpromote one another (Fig 1) Disturbances canfor example promote habitat heterogeneity andincrease resource availability for less competitivespecies and may therefore lead to an increasein species diversity (78) High levels of dispersalamong habitats aid recovery of ecosystems after(major) disturbance events by allowing recoloni-zation and recovery of affected populations (Fig1B) Large vertebrates present in complex eco-systems often act as dispersal agents for plants(54 55) and can introduce stochasticity into asystem eg through predation or grazing (79)Therefore the restoration of one of these pro-cesses may positively influence the functionalitylevels of the two other processes (Fig 1B) In-teractions among the processes can increase eco-system resilience by jointly promoting for examplefunctional redundancy or recolonizationRewilding efforts can be assessed by repre-

senting ecosystems in their degraded and restoredstates in a three-dimensional pyramid-shapedspacewhere each axis corresponds to an ecologicalprocess and the faces represent the interactionbetween processes (Fig 2) During the restora-tion of a process the respective vertices of thepyramid move farther from the origin and thevolume of the pyramid increases The differencein volume between the restored and degradedecosystem is thus a proxy for the effect of re-wilding on the resilience of the ecosystem Be-cause the processes interact restoring only onedimension but leaving the other two unaddressedtypically corresponds to a smaller improvementthan that facilitated by restoring the three di-mensions simultaneously (eg the change in thevolume of the pyramid is very small when one ofthe axes is fully restored but the other two axesremain highly degraded)

Rewilding as a societal choice

Ecosystems cannot be assessed separately fromhuman societies (80) All areas that are candi-dates for rewilding are influenced by peopleandor have a history of use Consequently anyrewilding project can affect local livelihoods andwell-being Societal changes can influence eco-systems in positive or negative ways and viceversa and the trajectories of ecosystems are oftendefined by human decisions that focus on thedelivery of certain resources and ecosystem ser-vices (67 81) Considering and managing for in-teractions between ecosystems and people while

assessing and communicating the benefits ofrewilding for society (Fig 2) can incentivizeactions that benefit both ecosystems and society(67) thereby increasing the acceptance and suc-cess of rewilding endeavorsThe restoration of the three ecosystem pro-

cesses can positively affect peoplersquos lives invarious ways Rewilding plays an important rolefor nonmaterial contributions of nature and rela-tional values of biodiversity (82) A growing bodyof literature concludes that exposure to green ornatural spaces can reduce stress levels increasepositive emotions and cognitive function encour-age physical activity and facilitate social cohesionin humans (83ndash86) In particular wilderness ex-periences provide an opportunity for ecotherapyto promote psychological resilience in childrenand adolescents (87) and personal transformationand self-fulfillment in adults (88) Moreover thesatisfaction that people may gain from knowingthat a species or ecosystem is thriving (89 90) canreach societies in great geographical distanceto an actual rewilding site The presence of char-ismatic or symbolic species or landscapes caninspire spiritual artistic and technological de-velopment (42) Far-ranging andmigrating speciestraveling on dispersal pathways may motivatenature-based activities such as bird-watching(42)Witnessing natural processes associatedwithchildhood experiences such as the migration ofswallows or cranes can promote a sense of placeand rootedness and may be the basis for nar-ratives rituals and celebrations that shape cul-tural identity (42)Economic benefits of rewildingmay arise from

opportunities for nature-based economies andalternative sources of income based on non-material contributions from nature [eg recrea-tional activities (42 91 92)] Furthermore naturaldisturbance events can trigger innovation andchange in social-ecological systems (93) Rewild-ing promotes other regulating services andnature-based solutions such as regulation of climate airquality pollination and dispersal of seeds (42 94)Improved dispersal potential and trophic com-plexity may prevent the depletion of materialcontributions from nature (42) such as econom-ically relevant natural resources (eg wildlifegame) not only in the areas undergoing rewild-ing but also in surrounding areasHowever rewilding can also have undesired

consequences for people Natural disturbancessuch as fires or floods may threaten humansand human infrastructure (95) Humanndashwildlifeconflictsmdashfor example crops damaged by largeherbivores or livestock killed by large predators(96)mdashare becoming more frequent and moresevere where these animals are reintroducedor their populations recover (97) Additionallyconcerns about the loss of traditional culturallandscapes including their unique natural andcultural heritage are growing in Europe andother regions (91 98 99) Particular unease hasbeen expressed regarding impacts on farmlandbiodiversity and on cultural ecosystem servicesmdashfor instance aesthetic values (100) sense of place(101) and a general ldquoerasurerdquo of human history

Perino et al Science 364 eaav5570 (2019) 26 April 2019 4 of 8

RESEARCH | REVIEWon M

arch 3 2020

httpsciencesciencemagorg

Dow

nloaded from

and involvement with the land and its flora andfauna (32)In sum the relationship of peoplewithwildness

in nature is and has always been characterizedby sets of paradoxes (102) These range from con-tradictory views of wildness in nature ascribedto prehistoric peoples as a ldquoconstant threat to[human] life and livelihoodrdquo versus the ldquoprimarysource of life and livelihoodrdquo to contemporarycontradictory perceptions as ldquoa potentially dan-gerous alienating and challenging placerdquo ver-sus ldquoa potentially peaceful refuge to relax andconveniently enjoyrdquo (102) This range of emo-tions emphasizes that well-planned rewildingprojects that mitigate possible conflicts have thepotential to maximize positive experiences andbeneficial contributions from nature

Applying the framework

A structured and participatory approach to re-wilding is important to ensure that all stake-holders have a clear understanding of the goalsmanagement options desirable outcomes andassociated risks (103) The first step of a rewild-ing project should be an analysis of the ecologicalstatus of the focus area by identifying missingandor degraded components Paleoecologicaldatamdashfor example on past vegetation changemegafauna presence or fire dynamicsmdashas wellas information on land-use histories should beconsidered in such analyses (4)In the second stepmanagers should assess the

ecological viability of different management op-tions and potential synergies among those To-gether with key stakeholders (conservationistsfarmers hunters and the general public) man-agers should identify social-ecological constraints(such as infrastructure hindering dispersal emerg-ing humanndashwildlife conflicts or risks associatedwith the restoration of natural disturbances) andevaluate benefits and disadvantages associatedwith the rewilding interventionThe third step is the implementation of re-

wilding actions via an adaptive management ap-proach This includes monitoring of the differentinterventions ideally using a before-after-control-impact (BACI) approach (104) that considersboth ecological and societal outcomes Resultsof this monitoring may lead to adjustments inongoing rewilding interventions or raise theneed for further management actions and de-cisions The implementation phase should beaccompanied by a communication strategy thatinvolves affected communities in decisions aswell as outreach activities that inform the widerpublic about the outcomes of rewilding Theseshould ideally be offered via an array of op-portunities for nature experiences (guided toursthrough the rewilding area nature educationtools and opportunities for leisure activities)Additionally managers may seek to develop op-portunities for sustainable business opportunitiesto increase the acceptance of rewilding amongstakeholdersOur stepwise approach can also be applied for

passive rewilding projects In such cases there isno deliberate decision to initiate a project but

insteadmanagers can take advantage of ongoingsocial-ecological dynamics (eg farmland aban-donment) If such an opportunity is identifiedthe first step will involve an assessment of theongoing passive rewilding dynamics associatedrisks and benefits and potential impediments tothose dynamics The second step will focus onidentifying options to support those dynamicsand mitigate threats This step will often involvethe consolidation of ongoing nonintervention(eg establishment of no-hunting arrangementsor protected areas) or themitigation of emergingconflicts As in active rewilding projects thethird step involves adaptive management mon-itoring and outreach activitiesWe now describe the stepwise application of

our framework with four rewilding case studiesspanning a range of scales ecosystem types anddegrees of intervention (Fig 2) As will becomeapparent the development of a rewilding projectis rarely a linear process Owing to the adaptivenature of our approach some of the steps will becarried out repeatedly andor in parallel

Restoration of the natural floodregime in the Leipziger Auwald cityforest Germany

The Leipziger Auwald is an alluvial forest sur-rounding and crossing the city of Leipzig inGermany Since the middle of the 19th centuryflood suppression and changes in forest man-agement have led to a well-documented shift intree community composition with increasingdominance of sycamore (Acer pseudoplatanus)Norway maple (Acer platanoides) and commonash (Fraxinus excelsior) mainly at the expense ofhornbeam (Carpinus betulus) and oak (Quercusrobur) (105) In this state connectivity betweenthe Auwaldrsquos waterbodies is severely diminishedso active management is necessary to restore theflood regime (Fig 2A)After identifying flood disturbance as a major

missing component of this ecosystem city man-agers initiated yearly experimental flooding ofa pilot area in the early 1990s (106) Results ofconcomitantmonitoring confirmed the effective-ness and suitability of this management actionFlooding led to an increase of flood-tolerantspecies such as oak and hornbeam and a decreaseor local extinction of some plant species that areintolerant to flooding but had become dominantafter flooding had been suppressed (eg sycamoreand Norway maple) (106) At the same time(re)colonization by moisture-tolerant slug spe-cies and several ground beetle species associ-ated with alluvial forest systems was observed(106) The findings of this long-term experimentthus inform the projectrsquos implementation phasein which the natural flood regime is restoredin several drained branches of the Luppe River(Lebendige Luppe project) (72) (Fig 2A)The implementation phase is accompanied by

an extensive outreach strategy that offers severalopportunities for the public to engage with theecosystem in the Auwald Examples includemul-timedia teaching material to support environ-mental education and tools (eg magnifying

glasses landing nets and maps) for interactiveexperiments that allow children to learn aboutthe ecology and topography of the alluvial forestand explore its flora and fauna A local conserva-tion nongovernmental organization (NGO) or-ganizes excursions to inform residents aboutongoing management activities and regular pub-lic discussion forums offer the opportunity toengage actively in the project Two concomitantresearch programs assess the ecological outcomesof the project and monitor and evaluate the ac-ceptance and perception of natural processes inthe Auwald respectively (107)

Nonintervention policy in the SwissNational Park

Established in 1914 the Swiss National Park isthe oldest national park in Europe and the largestprotected area in Switzerland (108) In 1909 theparkrsquos foundersmdashbotanists and naturalists con-cerned with the widespread development oftourism infrastructure threatening the arearsquos uni-que flora and faunamdashidentified the region aroundthe Pass dal Fuorn as a suitable target area owingto its remoteness and species richness (108)Making space for natural processes and con-

ducting research on their development are cen-tral missions of the parkrsquos management (108)The establishment of the park and managementdecisions were advised by cartographers andnaturalists with extensive knowledge of the areaand its ecosystems (109) The protection statusof the area was secured by a lease agreementnegotiated with the local municipalities andfinanced through the foundation of the SwissFederation of Nature ConservationSince its establishment the National Park has

been subject to a strict nonmanagement ap-proach and has been fully protected from humanactivities such as hunting agriculture and forestryTrophic complexity was promoted through tar-geted reintroductions of ibex (Capra ibex) in 19201923 and 1926 and bearded vultures (Gypaetusbarbatus) from 1991 to 2007 (110) Natural dis-turbances are not managed and dispersal po-tential is high formost species (Fig 2B) Ecosystemdevelopment has been monitored continuouslyand many of the monitoring schemes have beenin place for decades (109) Conflicts with localcommunities were mitigated via selected activemanagement measures For example public dis-content over sapling damage caused by red deer(Cervus elaphus) was alleviated by organizinghunting events outside the borders of the park(109) The nonmanagement approach has resultedin the recovery of large populations of red deerchamois (Rupicapra rupicapra) ibex and roedeer (Capreolus capreolus) species that werenearly extinct or very rare in Switzerland whenthe park was established (111) The increased reddeer population density has resulted in higherplant species richness in subalpine grassland (112)Additionally wolves (Canis lupus) and brownbears (Ursus arctos) have recently been sightedsuggesting the imminent recolonization of thearea by large predators Socioeconomic studiesshow that the park attracts ~150000 visitors per

Perino et al Science 364 eaav5570 (2019) 26 April 2019 5 of 8

RESEARCH | REVIEWon M

arch 3 2020

httpsciencesciencemagorg

Dow

nloaded from

year contributing substantially to the economicprosperity of the region (109 113 114)

Restoring ecological interactionsin the Tijuca National Park Rio deJaneiro city Brazil

The Atlantic Forest of Brazil is a globally im-portant biodiversity hotspot However most ofthe protected areas containing Atlantic Forestremnants have been defaunated (115) One ofthese remnants is the TijucaNational Park in Riode Janeiro During the 17th and 18th centuriesdeforestation for agricultural purposes and hunt-ing pressure led to severe losses of native faunaBecause the forest is completely surrounded byurban infrastructure the affected animal speciescould not fully recover after the areawas reforestedin the 19th century (116) anddispersal of mammalspecies to other ecosystems is still inhibitedThe REFAUNA Project was established in

2012 to restore the mammal community viagradual reintroductions of species that have dis-appeared from the Atlantic Forest (116) Tijucawas considered suitable for first reintroductionsbecause its relatively small size and urban loca-tionwould allow for easymonitoring and controlof the released animals (116) Researchers iden-tified two native locally extinct candidate spe-cies the red-rumped agouti (Dasyprocta leporina)and the howlermonkey (Alouatta guariba) bothof which were expected to promote ecologicalinteractions Agoutis are important dispersers oflarge seeded plants (117) and increase seed survivalby transporting them to locations with lower den-sities of conspecific tree species Howlermonkeysinfluence dung beetle abundance and the de-composition of howler dung by the beetles can en-hance nutrient cycling and soil fertilization (118)Concomitant monitoring revealed that the

presence of agoutis and howler monkeys en-hanced ecological interactions in the park Agoutisbroadened their diet and improved the dispersaland germination success of several large-seededplants By interacting with the dung beetle com-munity howler monkeys promoted the dispersalof large seeds with likely positive effects on forestregeneration (116) (Fig 2C) Although Tijuca isBrazilrsquos most popular national park (119) thereis little emotional connection between the parkand people living in adjacent communities (120)To improve the linkage between the park andlocal communities the park administration hasinstalled a park council through which represent-atives of governmental institutions NGOs andthe private sector aim to reach satisfactory man-agement decisions for all stakeholders (121) Acommunity-based cooperative project has trainedlocals as tourist guides and offers tours throughthe park and a neighboring favela Additionallythe cooperative runs a restaurant that offers localcuisine prepared with products growing in theforest and community gardens (122 123)

Ecosystem and wildlife recovery in theChernobyl exclusion zone Belarus

Themeltdown of the nuclear reactor in Chernobylon 26 April 1986 resulted in massive contam-

ination especially in the immediate surroundingarea (124ndash126) The evacuation of the entirelocal population within a 30-km exclusion zonearound the reactor and the most strongly con-taminated areas outside this zone resulted in theabandonment of ~1400 km2 of agricultural land(40 41) The breakdown of the Soviet Union withwidespread outmigration and an additional 36of all farmland abandoned in Belarus andUkrainefurther lowered human pressure in the regionsurrounding the Chernobyl site (41)Two years after the meltdown the Belarusian

part of the exclusion zone and adjacent areaswere turned into the strictly protected 1300 km2

Polesie State Radioecological Reserve In 1993the reserve was extended by 850 km2 making itthe largest nature reserve in Belarus (127) Man-agement of the exclusion zone on both sides of theborder has since followed a paradigm of minimalto no intervention Targeted reintroductions ofEuropean bison (Bison bonasus) in the PolesieState Radioecological Reserve and of Przewalskirsquoshorses (Equus ferus przewalskii) in the Ukrainianexclusion zone to restore trophic interactionsin the Chernobyl area were exceptions to thispassive approach Recognizing the growing eco-logical and conservation value of the Chernobylarea the Ukrainian government established the2300-km2 Chornobyl Radiation and EcologicalBiosphere Reservemdashan almost 5000 km2 contig-uous rewilding area in the heart of EasternEuropemdashin 2016 (128) Management activities inthe biosphere reserve aim to recover biodiversityand ecosystem resilience and includemonitoringof the ecological medical and radiation status ofthe area as well as educational activities (128)The region now harbors the entire portfolio

of extant European large carnivores [wolf lynx(Lynx lynx) and brown bear] large herbivores[European bison wild horse moose (Alces alces)red deer roe deer and wild boar (Sus scrofa)] arich mesopredator community [eg Europeanbadger (Meles meles) raccoon dog (Nyctereutesprocyonoides) and red fox (Vulpes vulpes)] andkey ecosystem engineers such as the Eurasianbeaver (Castor fiber) The Chernobyl exclusionzone is the only area where these species in-teract in sizable numbers in a large wildernesscomplex and can thus be considered one of themost iconic natural experiments on rewilding inrecent history

The way forward

Rewilding directly aims to restore ecological func-tions instead of particular biodiversity composi-tional states Therefore the effects of rewildingmay be indirect and unexpected Consequentlythe development of sound rewilding plans re-quires a deep understanding of the interactingecosystem processes that lead to resilience andthe socioeconomic context in which rewildingoccurs Interdisciplinary training of scientistsand practitioners is necessary to develop suchunderstanding Moreover objective evidence-based assessments of rewilding initiatives areneeded to make rewilding projects fully account-able to funders the public and the research com-

munity A recently proposedmethod to assess theprogress of rewilding projects through a combi-nation of expert opinion and monitoring data(129) is a step toward this goalUnfortunately current landscape management

and conservation policies do not provide sufficientopportunities for rewilding to be implementedon a broader scale For instance the EuropeanUnionrsquos common agricultural policy incentivizesagricultural activities in low-production areasimpeding opportunities for rewilding (130) Re-storation policies often focus on the safeguard-ing of current or historical conditions (130) andthe protection of certain species and habitats(24 130 131) Therefore the successful contribu-tion of rewilding to national and internationalbiodiversity goals depends on policy changesthat shift the conservation focus toward restor-ing the ecological processes identified in ourframework (131)Discussions on post-2020 biodiversity strat-

egies by the signatory countries of the Conven-tion on Biological Diversity are currently beinginitiated and the United Nations General Assem-bly has recently declared 2021ndash2030 the ldquodecadeof ecosystem restorationrdquo (132) We believe thatrewilding provides one of the possible pathwaystoward the vision in which ldquoBy 2050 biodiversityis valued conserved restored and wisely usedmaintaining ecosystem services sustaining ahealthy planet and delivering benefits essentialfor all peoplerdquo (133) Perhaps innovative policychanges that favor rewilding can add to the cur-rent momentum for novel approaches to resto-ration (19 134) For instance Aichi BiodiversityTarget 15 which aims to restore 15 of degradedecosystems by 2020 could be revised to recog-nize rewilding as amajor approach to ecologicalrestoration An ambitious positive target of in-creasing wildness across the globe by 2030 couldbe a truly inspiring goal infusing energy andpublic support into global biodiversity policies

REFERENCES AND NOTES

1 R Dirzo P H Raven Global state of biodiversity and lossAnnu Rev Environ Resour 28 137ndash167 (2003) doi 101146annurevenergy28050302105532

2 ldquoEcosystems and human well-being Biodiversity synthesisrdquoMillenium Ecosystem Assessment (World ResourcesInstitute 2005)

3 J A Foley et al Global consequences of land useScience 309 570ndash574 (2005) doi 101126science1111772pmid 16040698

4 A D Barnosky et al Merging paleobiology with conservationbiology to guide the future of terrestrial ecosystemsScience 355 eaah4787 (2017) doi 101126scienceaah4787pmid 28183912

5 B J Cardinale et al Biodiversity loss and its impact onhumanity Nature 486 59ndash67 (2012) doi 101038nature11148 pmid 22678280

6 N Pettorelli et al Satellite remote sensing of ecosystemfunctions Opportunities challenges and way forward RemoteSens Ecol Conserv 4 71ndash93 (2017) doi 101002rse259

7 C S Holling Resilience and stability of ecological systemsAnnu Rev Ecol Syst 4 1ndash23 (1973) doi 101146annureves04110173000245

8 M Scheffer S Carpenter J A Foley C Folke B WalkerCatastrophic shifts in ecosystems Nature 413 591ndash596(2001) doi 10103835098000 pmid 11595939

9 T H Oliver et al Declining resilience of ecosystem functionsunder biodiversity loss Nat Commun 6 10122 (2015)doi 101038ncomms10122 pmid 26646209

Perino et al Science 364 eaav5570 (2019) 26 April 2019 6 of 8

RESEARCH | REVIEWon M

arch 3 2020

httpsciencesciencemagorg

Dow

nloaded from

10 S Chape J Harrison M Spalding I Lysenko Measuring theextent and effectiveness of protected areas as anindicator for meeting global biodiversity targets PhilosTrans R Soc B 360 443ndash455 (2005) doi 101098rstb20041592 pmid 15814356

11 C L Gray et al Local biodiversity is higher inside thanoutside terrestrial protected areas worldwideNat Commun 7 12306 (2016) doi 101038ncomms12306pmid 27465407

12 R M Pringle Upgrading protected areas to conserve wildbiodiversity Nature 546 91ndash99 (2017) doi 101038nature22902 pmid 28569807

13 W F Laurance et al Averting biodiversity collapse in tropicalforest protected areas Nature 489 290ndash294 (2012)doi 101038nature11318 pmid 22832582

14 C A Hallmann et al More than 75 percent decline over27 years in total flying insect biomass in protected areasPLOS ONE 12 e0185809 (2017) doi 101371journalpone0185809 pmid 29045418

15 L P Koh et al Species coextinctions and the biodiversitycrisis Science 305 1632ndash1634 (2004) doi 101126science1101101 pmid 15361627

16 S L Pimm P Raven Extinction by numbers Nature403 843ndash845 (2000) doi 10103835002708pmid 10706267

17 L Thomas J Middleton Guidelines for Management Planningof Protected Areas vol 10 of Best Practice Protected AreaGuidelines (IUCN 2003)

18 A Valiente-Banuet et al Beyond species lossThe extinction of ecological interactions in a changingworld Funct Ecol 29 299ndash307 (2015) doi 1011111365-243512356

19 E S Higgs et al Keep ecological restoration open andflexible Nat Ecol Evol 2 580 (2018) doi 101038s41559-018-0483-9 pmid 29403073

20 D Lindenmayer et al A checklist for ecological managementof landscapes for conservation Ecol Lett 11 78ndash91 (2008)pmid 17927771

21 R T Corlett Restoration reintroduction and rewilding in achanging world Trends Ecol Evol 31 453ndash462 (2016)doi 101016jtree201602017 pmid 26987771

22 B Cantrell L J Martin E C Ellis Designing autonomyOpportunities for new wildness in the AnthropoceneTrends Ecol Evol 32 156ndash166 (2017) doi 101016jtree201612004 pmid 28108135

23 J Lorimer et al Rewilding Science practice and politicsAnnu Rev Environ Resour 40 39ndash62 (2015) doi 101146annurev-environ-102014-021406

24 P Jepson A rewilding agenda for Europe Creating a networkof experimental reserves Ecography 39 117ndash124 (2016)doi 101111ecog01602

25 N Fernaacutendez L M Navarro H M Pereira RewildingA Call for boosting ecological complexity in conservationConserv Lett 10 276ndash278 (2017) doi 101111conl12374

26 G Monbiot Feral Searching for Enchantment on the Frontiersof Rewilding (Penguin 2013)

27 C Queiroz R Beilin C Folke R Lindborg Farmlandabandonment Threat or opportunity for biodiversityconservation - a global review Front Ecol Environ 12288ndash296 (2014) doi 101890120348

28 K Katoh S Sakai T Takahashi Factors maintaining speciesdiversity in satoyama a traditional agricultural landscapeof Japan Biol Conserv 142 1930ndash1936 (2009)doi 101016jbiocon200902030

29 J Prior E Brady Environmental aesthetics and rewildingEnviron Values 26 31ndash51 (2017) doi 103197096327117X14809634978519

30 D Nogueacutes-Bravo D Simberloff C Rahbek N J SandersRewilding is the new Pandorarsquos box in conservationCurr Biol 26 R87ndashR91 (2016) doi 101016jcub201512044 pmid 26859272

31 D R Rubenstein D I Rubenstein From Pleistocene totrophic rewilding A wolf in sheeprsquos clothing Proc Natl AcadSci USA 113 E1 (2016) doi 101073pnas1521757113pmid 26676584

32 D Joslashrgensen Rethinking rewilding Geoforum 65 482ndash488(2015) doi 101016jgeoforum201411016

33 M E Souleacute R Noss Rewilding and biodiversityComplementary goals for continental conservationWild Earth 8 18ndash28 (1998)

34 L M Navarro H M Pereira Rewilding abandonedlandscapes in Europe Ecosystems 15 900ndash912 (2012)doi 101007s10021-012-9558-7

35 S Estel et al Mapping farmland abandonment andrecultivation across Europe using MODIS NDVI time seriesRemote Sens Environ 163 312ndash325 (2015) doi 101016jrse201503028

36 P Meyfroidt F Schierhorn A V Prishchepov D MuumlllerT Kuemmerle Drivers constraints and trade-offsassociated with recultivating abandoned cropland in RussiaUkraine and Kazakhstan Glob Environ Change 37 1ndash15(2016) doi 101016jgloenvcha201601003

37 T Hanson Biodiversity conservation and armed conflictA warfare ecology perspective Ann N Y Acad Sci1429 50ndash65 (2018) doi 101111nyas13689pmid 29683517

38 M Baumann V C Radeloff V Avedian T KuemmerleLand-use change in the Caucasus during and after theNagorno-Karabakh conflict Reg Environ Change 151703ndash1716 (2015) doi 101007s10113-014-0728-3

39 M Baumann T Kuemmerle The impacts of warfare andarmed conflict on land systems J Land Use Sci 11 672ndash688(2016) doi 1010801747423X20161241317

40 T G Deryabina et al Long-term census data revealabundant wildlife populations at Chernobyl Curr Biol25 R824ndashR826 (2015) doi 101016jcub201508017pmid 26439334

41 P Hostert et al Rapid land use change after socio-economicdisturbances The collapse of the Soviet Union versusChernobyl Environ Res Lett 6 045201 (2011) doi 1010881748-932664045201

42 S Diacuteaz et al Assessing naturersquos contributions to peopleScience 359 270ndash272 (2018) doi 101126scienceaap8826pmid 29348221

43 J Terborgh et al in Continental Conservation Design andManagement Principles for Long-Term Regional ConservationNetworks M E Souleacute J Terborgh Eds (Island Press 1999)pp 39ndash64

44 J-C Svenning et al Science for a wilder AnthropoceneSynthesis and future directions for trophicrewilding research Proc Natl Acad Sci USA113 898ndash906 (2016) doi 101073pnas1502556112pmid 26504218

45 Definition of ldquowildernessrdquo by Oxford English Dictionaryhttpsenoxforddictionariescomdefinitionwilderness

46 R L Chapman Ecological restoration restoredEnviron Values 15 463ndash478 (2006) doi 103197096327106779116096

47 Definition of ldquowildnessrdquo by Oxford English Dictionaryhttpsenoxforddictionariescomdefinitionwildness

48 J Bengtsson et al Reserves resilience and dynamiclandscapes AMBIO 32 389ndash396 (2003) doi 1015790044-7447-326389 pmid 14627367

49 T Elmqvist et al Response diversity ecosystem change andresilience Front Ecol Environ 1 488ndash494 (2003)doi 1018901540-9295(2003)001[0488RDECAR]20CO2

50 K N Suding K L Gross G R Houseman Alternativestates and positive feedbacks in restoration ecologyTrends Ecol Evol 19 46ndash53 (2004) doi 101016jtree200310005 pmid 16701225

51 R V Soleacute J M Montoya Complexity and fragility inecological networks Proc Biol Sci 268 2039ndash2045 (2001)doi 101098rspb20011767 pmid 11571051

52 C N Foster P S Barton D B Lindenmayer Effects of largenative herbivores on other animals J Appl Ecol 51 929ndash938(2014) doi 1011111365-266412268

53 R van Klink F van der Plas C G E T van NoordwijkM F WallisDeVries H Olff Effects of large herbivoreson grassland arthropod diversity Biol Rev CambPhilos Soc 90 347ndash366 (2015) doi 101111brv12113pmid 24837856

54 D H Janzen Dispersal of small seeds by big herbivoresFoliage is the fruit Am Nat 123 338ndash353 (1984)doi 101086284208

55 C Bello et al Defaunation affects carbon storage in tropicalforests Sci Adv 1 e1501105 (2015) doi 101126sciadv1501105 pmid 26824067

56 P S Barton S A Cunningham D B LindenmayerA D Manning The role of carrion in maintaining biodiversityand ecological processes in terrestrial ecosystems Oecologia171 761ndash772 (2013) doi 101007s00442-012-2460-3pmid 23007807

57 R van Klink M F WallisDeVries Risks and opportunities oftrophic rewilding for arthropod communities Philos TransR Soc B 373 20170441 (2018) doi 101098rstb20170441 pmid 30348868

58 W J Ripple et al Status and ecological effects of the worldrsquoslargest carnivores Science 343 1241484 (2014)doi 101126science1241484 pmid 24408439

59 W J Ripple et al Collapse of the worldrsquos largest herbivoresSci Adv 1 e1400103 (2015) doi 101126sciadv1400103pmid 26601172

60 E S Bakker et al Combining paleo-data and modernexclosure experiments to assess the impact of megafaunaextinctions on woody vegetation Proc Natl AcadSci USA 113 847ndash855 (2016) doi 101073pnas1502545112 pmid 26504223

61 M Cardillo et al Multiple causes of high extinction risk inlarge mammal species Science 309 1239ndash1241 (2005)doi 101126science1116030 pmid 16037416

62 R Dirzo et al Defaunation in the Anthropocene Science345 401ndash406 (2014) doi 101126science1251817pmid 25061202

63 I Dorresteijn et al Incorporating anthropogenic effectsinto trophic ecology Predator-prey interactions in ahuman-dominated landscape Proc Biol Sci 28220151602 (2015) doi 101098rspb20151602pmid 26336169

64 D P J Kuijper et al Paws without claws Ecological effectsof large carnivores in anthropogenic landscapes ProcBiol Sci 283 20161625 (2016) doi 101098rspb20161625 pmid 27798302

65 J A Dunne R J Williams N D Martinez Network structureand biodiversity loss in food webs Robustness increases withconnectance Ecol Lett 5 558ndash567 (2002) doi 101046j1461-0248200200354x

66 J Persson G R Rauset G Chapron Paying for anendangered predator leads to population recoveryConserv Lett 8 345ndash350 (2015) doi 101111conl12171

67 S Ceaușu R A Graves A K Killion J-C SvenningN H Carter Governing trade-offs in ecosystem servicesand disservices to achieve human-wildlife coexistenceConserv Biol 101111cobi13241 (2019) doi 101111cobi13241 pmid 30350889

68 J F Franklin et al Threads of continuity Conserv Pract 18ndash17 (2000) doi 101111j1526-46292000tb00155x

69 D Lindenmayer S Thorn S Banks Please do not disturbecosystems further Nat Ecol Evol 1 31 (2017)doi 101038s41559-016-0031 pmid 28812609

70 S Thorn et al Impacts of salvage logging on biodiversityA meta-analysis J Appl Ecol 55 279ndash289 (2018)doi 1011111365-266412945 pmid 29276308

71 D Kulakowski et al A walk on the wild side Disturbancedynamics and the conservation and management ofEuropean mountain forest ecosystems For Ecol Manage388 120ndash131 (2017) doi 101016jforeco201607037pmid 28860677

72 J Putkunz in Der Leipziger Auwald - ein dynamischerLebensraum Tagungsband zum 5 Leipziger Auensymposiumam 16 April 2011 UFZ-Bericht no 062011 C WirthA Reiher U Zaumlumer H D Kasperidus Eds (Helmholtz-Zentrum fuumlr Umweltforschung 2011) pp 31ndash37

73 K E Moseby G W Lollback C E Lynch Too much of agood thing successful reintroduction leads to overpopulationin a threatened mammal Biol Conserv 219 78ndash88 (2018)doi 101016jbiocon201801006

74 T N Wasserman S A Cushman A S ShirkE L Landguth J S Littell Simulating the effects ofclimate change on population connectivity of Americanmarten (Martes americana) in the northern RockyMountains USA Landsc Ecol 27 211ndash225 (2012)doi 101007s10980-011-9653-8

75 R C Lacy Importance of genetic variation to the viability ofmammalian populations J Mammal 78 320ndash335 (1997)doi 1023071382885

76 J M Rey Benayas J M Bullock ldquoVegetation restorationand other actions to enhance wildlife in Europeanagricultural landscapesrdquo in Rewilding EuropeanLandscapes H M Pereira L M Navarro Eds(Springer 2015) pp 127ndash142

77 T Merckx H M Pereira Reshaping agri-environmentalsubsidies From marginal farming to large-scale rewildingBasic Appl Ecol 16 95ndash103 (2015) doi 101016jbaae201412003

78 A Corteacutes-Avizanda R Jovani M Carrete J A DonaacutezarResource unpredictability promotes species diversityand coexistence in an avian scavenger guild A fieldexperiment Ecology 93 2570ndash2579 (2012) doi 10189012-02211 pmid 23431588

Perino et al Science 364 eaav5570 (2019) 26 April 2019 7 of 8

RESEARCH | REVIEWon M

arch 3 2020

httpsciencesciencemagorg

Dow

nloaded from

79 J A Estes et al Trophic downgrading of planet EarthScience 333 301ndash306 (2011) doi 101126science1205106pmid 21764740

80 F Berkes C Folke Eds Linking Social and EcologicalSystems Management Practices and Social Mechanisms forBuilding Resilience (Cambridge Univ Press 1998)

81 G C Daily Ed Naturersquos Services Societal Dependence onNatural Ecosystems (Island Press 1997)

82 S Diacuteaz et al The IPBES Conceptual FrameworkmdashConnectingnature and people Curr Opin Environ Sustain 14 1ndash16(2015) doi 101016jcosust201411002

83 L M Fredrickson D H Anderson A qualitative explorationof the wilderness experience as a source of spiritualinspiration J Environ Psychol 19 21ndash39 (1999)doi 101006jevp19980110

84 I Markevych et al Exploring pathways linking greenspace tohealth Theoretical and methodological guidance EnvironRes 158 301ndash317 (2017) doi 101016jenvres201706028pmid 28672128

85 K Engemann et al Childhood exposure to green space -A novel risk-decreasing mechanism for schizophreniaSchizophr Res 199 142ndash148 (2018) doi 101016jschres201803026 pmid 29573946

86 K Engemann et al Residential green space in childhoodis associated with lower risk of psychiatric disordersfrom adolescence into adulthood Proc Natl AcadSci USA 116 5118ndash5193 (2019) doi 101073pnas1807504116

87 A S Masten M J Reed ldquoResilience in developmentrdquo inHandbook of Positive Psychology S J Lopez C R SnyderEds (Oxford Library of Psychology Oxford Univ Press ed 22011) pp 74ndash88

88 L Naor O Mayseless How personal transformation occursfollowing a single peak experience in nature Aphenomenological account J Humanist Psychol 1011770022167817714692 (2017) doi 1011770022167817714692

89 J V Krutilla Conservation reconsidered Am Econ Rev 57777ndash786 (1967)

90 R R Alexander Modelling species extinction The case fornon-consumptive values Ecol Econ 35 259ndash269 (2000)doi 101016S0921-8009(00)00198-1

91 R T Corlett The role of rewilding in landscape design forconservation Curr Landsc Ecol Rep 1 127ndash133 (2016)doi 101007s40823-016-0014-9

92 J C Jobse L Witteveen J Santegoets D J StobbelaarldquoPreparing a new generation of wilderness entrepreneursrdquo inRewilding European Landscapes H M Pereira L M NavarroEds (Springer 2015) pp 191ndash204

93 L H Gunderson in Navigating Social-Ecological SystemsBuilding Resilience for Complexity and Change F BerkesJ Colding C Folke Eds (Cambridge Univ Press 2003)

94 ldquoThe IUCN Programme 2013ndash2016rdquo (IUCN 2012)95 M G Turner Disturbance and landscape dynamics in a

changing world Ecology 91 2833ndash2849 (2010)doi 10189010-00971 pmid 21058545

96 N Bauer A Wallner M Hunziker The change of Europeanlandscapes Human-nature relationships public attitudestowards rewilding and the implications for landscapemanagement in Switzerland J Environ Manage 902910ndash2920 (2009) doi 101016jjenvman200801021pmid 18848747

97 A Treves in Wildlife and Society The Science of HumanDimensions (Island Press 2009)

98 J Fischer T Hartel T Kuemmerle Conservation policy intraditional farming landscapes Conserv Lett 5 167ndash175(2012) doi 101111j1755-263X201200227x

99 T Plieninger et al Exploring ecosystem-change and societythrough a landscape lens Recent progress in Europeanlandscape research Ecol Soc 20 5 (2015) doi 105751ES-07443-200205

100 U Schirpke F Timmermann U Tappeiner E Tasser Culturalecosystem services of mountain regions Modelling theaesthetic value Ecol Indic 69 78ndash90 (2016) doi 101016jecolind201604001 pmid 27482152

101 F Houmlchtl S Lehringer W Konold ldquoWildernessrdquo What itmeans when it becomes a realitymdasha case study from the

southwestern Alps Landsc Urban Plan 70 85ndash95 (2005)doi 101016jlandurbplan200310006

102 K Arts A Fischer R Van der Wal The promise of wildernessbetween paradise and hell A cultural-historical exploration ofa dutch National Park Landsc Res 37 239ndash256 (2012)doi 101080014263972011589896

103 C R Margules R L Pressey Systematic conservationplanning Nature 405 243ndash253 (2000) doi 10103835012251 pmid 10821285

104 R H Green Sampling Design and Statistical Methods forEnvironmental Biologists (Wiley 1979)

105 D Haase J Glaumlser Determinants of floodplain forestdevelopment illustrated by the example of the floodplainforest in the District of Leipzig For Ecol Manage 258887ndash894 (2009) doi 101016jforeco200903025

106 K Richter H Teubert in Der Leipziger Auwald - eindynamischer Lebensraum Tagungsband zum 5 LeipzigerAuensymposium am 16 April 2011 UFZ-Bericht no 062011C Wirth A Reiher U Zaumlumer H D Kasperidus Eds(Helmholtz-Zentrum fuumlr Umweltforschung 2011) pp 45ndash50

107 Lebendige Luppe - Sozialwissenschaftliche Begleitungavailable at httpslebendige-luppedeindexphparticle_id=19

108 1904-1914 - Swiss National Parc available at wwwnationalparkchenaboutabout-usfoundation-and-development1904-1914

109 H Lozza in Jahrbuch des Vereins zum Schutz der Bergwelt(Verein zum Schutz der Bergwelt eV 2014) vol 79Jahrgang pp 69ndash78

110 Bundesversammlung der SchweizerischenEidgenossenschaft Bundesgesetz uumlber den SchweizerischenNationalpark im Kanton Graubuumlnden (Nationalparkgesetz)(1980)

111 J Senn W Suter Ungulate browsing on silver fir (Abies alba)in the Swiss Alps Beliefs in search of supporting dataFor Ecol Manage 181 151ndash164 (2003) doi 101016S0378-1127(03)00129-4

112 M Schuumltz A C Risch E Leuzinger B O KruumlsiG Achermann Impact of herbivory by red deer (Cervuselaphus L) on patterns and processes in subalpinegrasslands in the Swiss National Park For Ecol Manage 181177ndash188 (2003) doi 101016S0378-1127(03)00131-2

113 N Backhaus C Buser M Adamec D Jorio M SpeichWirtschaftliche Auswirkungen des Sommertourismus imUNESCO Biosphaumlrenreservat Val Muumlstair Parc Naziunal(Schriftenreihe Humangeographie vol 27 GeographischesInstitut Abteilung Humangeographie University of Zuumlrich2013)

114 F Knaus N Backhaus ldquoTouristische Wertschoumlpfung inSchweizer Paumlrkenrdquo (Swiss Academies Factsheets vol 9no 3 2014)

115 M Galetti A S Pires P H S Brancalion F A S FernandezReversing defaunation by trophic rewilding in empty forestsBiotropica 49 5ndash8 (2017) doi 101111btp12407

116 F A S Fernandez et al Rewilding the Atlantic ForestRestoring the fauna and ecological interactions of a protectedarea Perspect Ecol Conserv 15 308ndash314 (2017)doi 101016jpecon201709004

117 P A Jansen et al Thieving rodents as substitutedispersers of megafaunal seeds Proc Natl Acad Sci USA109 12610ndash12615 (2012) doi 101073pnas1205184109pmid 22802644

118 E Nichols et al Ecological functions and ecosystem servicesprovided by Scarabaeinae dung beetles Biol Conserv 1411461ndash1474 (2008) doi 101016jbiocon200804011

119 E Viveiros de Castro T Beraldo Souza B ThapaDeterminants of Tourism Attractiveness in the National Parksof Brazil Parks 21 51ndash62 (2015) doi 102305IUCNCH2014PARKS-21-2EVDCen

120 M M Carreiro W C Zipperer Co-adapting societal andecological interactions following large disturbances in urbanpark woodlands Austral Ecol 36 904ndash915 (2011)doi 101111j1442-9993201002237x

121 J-P Briot P Guyot M Irving F Barros N FrydmanN Giambasi B Zeigler Eds ldquoParticipatory simulation forcollective management of protected areas for biodiversity

conservation and social inclusionrdquo (Society for Modeling andSimulation International 2007)

122 O A Barros M E Melo From Myth to Reality TheExperience of Sustainable Tourism in The Vale EncantadoCommunity in Tijuca Forest Rio de Janeiro Brazil FieldActions Sci Rep special issue 3 (2011) httpfactsreportsrevuesorg1131

123 A Nidumolu ldquoVale Encantado An emerging example forsustainable communitiesrdquo RioOnWatch - Community Report(2015)

124 A P Moslashller T A Mousseau Biological consequencesof Chernobyl 20 years on Trends Ecol Evol 21200ndash207 (2006) doi 101016jtree200601008pmid 16701086

125 ldquoEnvironmental consequences of the Chernobyl accidentand their remediation Twenty years of experiencerdquo(Radiological Assessment Reports Series InternationalAtomic Energy Agency 2006)

126 K Baverstock D Williams Chernobyl An overlooked aspectScience 299 44b (2003) doi 101126science299560344b pmid 12512509

127 About the Polesie State Radioecological Reservewwwzapovednikbyenabout

128 Conservation Optimization and Management of Carbonand Biodiversity in the Chornobyl Exclusion Zonehttpchornobyl-gefcomentasks-of-the-reservehtml

129 A Torres et al Measuring rewilding progress Philos TransR Soc B 101098rstb20170433 (2018) doi 101098rstb20170433

130 N Pettorelli et al Making rewilding fit for policy J Appl Ecol55 1114ndash1125 (2018) doi 1011111365-266413082

131 L M Navarro H M Pereira ldquoTowards a European policy forrewildingrdquo in Rewilding European Landscapes H M PereiraL M Navarro Eds (Springer 2015) pp 205ndash223

132 United Nations General Assembly ldquoResolution adopted bythe General Assembly on 1 March 2019 United NationsDecade on Ecosystem Restoration (2021ndash2030)rdquo ARES73284 (2019) wwwunorgengasearchview_docaspsymbol=ARES73284

133 D Cooper ldquoScenarios for the 2050 vision for biodiversityrdquo(Convention on Biological Diversity Montreal Canada 2010)

134 G M Mace et al Aiming higher to bend the curve ofbiodiversity loss Nat Sustain 1 448ndash451 (2018)doi 101038s41893-018-0130-0

ACKNOWLEDGMENTS

AP thanks M Marselle I Rosa A Torres C Meyer and J Hinesfor valuable comments on earlier drafts of this manuscript andthree anonymous reviewers whose comments helped tosubstantially improve this article Funding AP HMP LMNGP RvK and NF are supported by the German Centre forIntegrative Biodiversity Research (iDiv) Halle-Jena-Leipzig fundedby the German Research Foundation (FZT 118) J-CS and SCthank the Carlsberg Foundation (Semper Ardens projectMegaPast2Future grant CF16-0005 to J-CS) and the VILLUMFONDEN (VILLUM Investigator project grant 16549 to J-CS) foreconomic support AL was supported by the Portuguese Scienceand Technology Foundation (FCT) through grant SFRHBPD807472011 and the FARSYD project (PTDCAAG-REC50072014-POCI-01-01-0145-FEDER-016664) ACA was supported by aViccedilent Munt postdoctoral contract from the Juan de la CiervaIncorporacioacuten (IJCI-2014-20744) of the Spanish Ministry ofEconomy and Competitiveness and by a postdoctoral contract ofthe Vicepresidencia y Consejeriacutea de Innovacioacuten Investigacioacuten yTurismo of the Govern de les Illes Balears (PD0392017)JMB acknowledges funding from CEH NC project NEC06895JMRB was supported by the Spanish Ministry of Economy andCompetitiveness (grant CGL2014-53308-P) and the Madridgovernment (project S2013MAE-2719 REMEDINAL-3) HCW wassupported by the Belmont Forum Competing interests HMPis a former supervisory board member of the Rewilding EuropeFoundation CJS is the director of Wild Business Ltd J-CSis a past or present advisory board member of two rewildingprojects in Denmark

101126scienceaav5570

Perino et al Science 364 eaav5570 (2019) 26 April 2019 8 of 8

RESEARCH | REVIEWon M

arch 3 2020

httpsciencesciencemagorg

Dow

nloaded from

Rewilding complex ecosystems

Christopher J Sandom Jens-Christian Svenning and Helen C WheelerCorteacutes-Avizanda Roel van Klink Tobias Kuemmerle Angela Lomba Guy Peer Tobias Plieninger Joseacute M Rey Benayas Andrea Perino Henrique M Pereira Laetitia M Navarro Neacutestor Fernaacutendez James M Bullock Silvia Ceausu Ainara

DOI 101126scienceaav5570 (6438) eaav5570364Science

this issue p eaav5570Scienceimplementing it broadly and in a way that considers ongoing human interaction

review this concept and present a framework foret allands and wildness to landscapes we have altered Perino coexist with and leave space for other species The practice of rewilding has emerged as a method for returning wildimpacts on wilderness If there is any hope of retaining a biodiverse planetary system we must begin to learn how to

Humans have encroached upon a majority of Earths lands The current extinction crisis is a testament to humanFacilitating wildness

ARTICLE TOOLS httpsciencesciencemagorgcontent3646438eaav5570

CONTENTRELATED

filecontenthttpsciencesciencemagorgcontentsci36464383401fullhttpsciencesciencemagorgcontentsci3646438326fullhttpsciencesciencemagorgcontentsci3646438eaav3506full

REFERENCES

httpsciencesciencemagorgcontent3646438eaav5570BIBLThis article cites 118 articles 17 of which you can access for free

PERMISSIONS httpwwwsciencemagorghelpreprints-and-permissions

Terms of ServiceUse of this article is subject to the

is a registered trademark of AAASScienceScience 1200 New York Avenue NW Washington DC 20005 The title (print ISSN 0036-8075 online ISSN 1095-9203) is published by the American Association for the Advancement ofScience

Science No claim to original US Government WorksCopyright copy 2019 The Authors some rights reserved exclusive licensee American Association for the Advancement of

on March 3 2020

httpsciencesciencem

agorgD

ownloaded from