Environmental Research Letters

TOPICAL REVIEW • OPEN ACCESS

Fostering natural forest regeneration on former agricultural land througheconomic and policy interventionsTo cite this article: Robin L Chazdon et al 2020 Environ. Res. Lett. 15 043002

View the article online for updates and enhancements.

This content was downloaded from IP address 182.2.132.208 on 13/05/2020 at 08:08

https://doi.org/10.1088/1748-9326/ab79e6

Environ. Res. Lett. 15 (2020) 043002 https://doi.org/10.1088/1748-9326/ab79e6

TOPICAL REVIEW

Fostering natural forest regeneration on former agricultural landthrough economic and policy interventions

Robin LChazdon1,2, David Lindenmayer3,Manuel RGuariguata4, RenatoCrouzeilles2,5,6,JoséMaría ReyBenayas7 andElena LazosChavero8

1 Tropical Forests and People ResearchCentre, University of the SunshineCoast, 90 SippyDownsDrive, SippyDowns,QLD4556,Australia

2 International Institute for Sustainability, EstradaDonaCastorina 124, Rio de Janeiro, Brazil3 Sustainable Farms, Fenner School of Environment and Society, TheAustralianNational University, Canberra, ACT 2601, Australia4 Center for International Forestry Research, Av. LaMolina 1895, Lima, Peru5 International Institute for Sustainability Australia, Canberra, ACT, 2602, Australia6 Mestrado Profissional emCiências doMeioAmbiente, Universidade Veiga de Almeida, 20271-901 Rio de Janeiro, Brazil7 Departamento deCiencias de laVida—UnidadDocente de Ecología, Universidad deAlcalá, E-28805, Alcalá deHenares, Spain8 Instituto de Investigaciones Sociales, UNAM,CiudadUniversitaria, Coyoacán, 04510México

E-mail: rchazdon@usc.edu.au

Keywords: forest restoration, forest transition, land abandonment, out-migration, forest succession, reforestation, agriculturalintensification

AbstractUnder suitable conditions, deforested land used for agricultural crops or pastures can revert to forestthrough the assisted or unassisted process of natural regeneration. These naturally regenerating forestsconserve biodiversity, provide awide array of ecosystem goods and services, and support ruraleconomies and livelihoods. Based on studies in tropical and temperate forest ecosystems, wesummarize cases where natural regeneration is occurring in agricultural landscapes around theworldand identify the socio-ecological factors that favor its development and affect its qualities, outcomesand persistence.We describe how the economic and policy context creates barriers for thedevelopment, persistence, andmanagement of naturally regenerating forests, including perverseoutcomes of policies intended to enhance protection of native forests.We concludewithrecommendations for specific economic and policy interventions at local, national, and global scalesto enhance forest natural regeneration and to promote the sustainablemanagement of regrowthforests on former agricultural landwhile strengthening rural communities and economies.

1. Introduction

When crop fields and pastures that earlier replacednative forests are left unused, the process of naturalregeneration—also known as secondary succession,old-field succession, forest regrowth, spontaneousrestoration or passive restoration—often leads to thedevelopment of a new forest system that graduallyregains many properties of the previous forest ecosys-tem (Cramer et al 2008, Chazdon 2014). During thisprocess, native vegetation regenerates in several ways,including by seeds shed in response to burning, fromseeds in the soil or newly deposited by wind or byanimals, from resprouting rootstocks, or by vegetativepropagules (Duncan and Chapman 1999, Pignataroet al 2017). In this context, natural regeneration of

forests is both an ecological process as well as atransition from agricultural to forest land use and landcover. The nature of forest regeneration on formeragricultural land defines a distinct ecological, socialand policy context that contrasts with selective loggingand associated silvicultural treatments in naturalforestsmanaged for timber production.

A forest undergoing natural regeneration follow-ing agricultural land use is a socio-ecological system intransition (Lambin and Meyfroidt 2010). Wheresocio-economic and biophysical conditions are favor-able, this system is likely to recover the structuralproperties, species composition and socio-ecologicalfunctions of the prior forest ecosystem (Filotas et al2014, Ghazoul et al 2015, Ghazoul and Chazdon2017). Unfavorable conditions, however, can push the

OPEN ACCESS

RECEIVED

30October 2019

REVISED

21 February 2020

ACCEPTED FOR PUBLICATION

25 February 2020

PUBLISHED

15April 2020

Original content from thisworkmay be used underthe terms of the CreativeCommonsAttribution 4.0licence.

Any further distribution ofthis workmustmaintainattribution to theauthor(s) and the title ofthework, journal citationandDOI.

© 2020TheAuthor(s). Published by IOPPublishing Ltd

https://doi.org/10.1088/1748-9326/ab79e6mailto: rchazdon@usc.edu.auhttps://crossmark.crossref.org/dialog/?doi=10.1088/1748-9326/ab79e6&domain=pdf&date_stamp=2020-04-15https://crossmark.crossref.org/dialog/?doi=10.1088/1748-9326/ab79e6&domain=pdf&date_stamp=2020-04-15http://creativecommons.org/licenses/by/4.0http://creativecommons.org/licenses/by/4.0http://creativecommons.org/licenses/by/4.0

system towards an alternate steady state where activeinterventions are required to restore a forest ecosys-tem (Suding et al 2004). Increasing land-use intensity,weed infestations, and lack of seed dispersal, canstrongly modify recovery trajectories, including spe-cies composition (Goldsmith et al 2011, Jakovac et al2015, 2016).

Naturally regenerating forests on former agri-cultural land can provide solutions for conservation ofbiodiversity, mitigation of, and adaptation to, climatechange, and multiple ecosystem goods and services(Houghton et al 2015, Locatelli et al 2015, Wilson et al2017, Jones et al 2019, Matos et al 2019, Pugh et al2019). Similar benefits can be provided by active forestrestoration (e.g. deliberate planting) and diverse formsof reforestation, but at significantly higher costs(Bullock et al 2011). For millennia, naturally regenerat-ing forests in shifting cultivation systems were a nexusfor food production and forest management (Hernán-dez-X et al 1995, Chazdon 2014). Recent expansion ofintensified and mechanized agricultural systems, how-ever, has often displaced traditional smallholder agri-culture, putting natural regeneration of forests in limbowith regard to land management policies, environ-mental regulations, and restoration targets (Wood et al2016,Martin et al 2018, Rasmussen et al 2018).

In preparation for the UN Decade of EcosystemRestoration (2021–2030), it is timely to considerwhere and how naturally regenerating forests on landpreviously used for crops or grazing can contribute tomassively up-scaling efforts to restore degraded andlost ecosystems to conserve biodiversity, combat cli-mate change, enhance food security, and protect watersupplies in a social, economic, and ecologically effec-tive manner (Chazdon and Brancalion 2019). Bastinet al (2019) estimated that 9 million km2 of restoredwoodlands and forests globally could be ecologicallysuitable areas for reforestation (including naturalregeneration). However, the benefits and feasibility ofrecovering forests to this extent have not been fullyevaluated (Chazdon and Brancalion 2019), nor do wehave a clear vision of the potential or feasibility of nat-ural regeneration to replenish native forests at thismassive scale.

Natural regeneration can occur spontaneouslywithout human intervention after the cessation of pre-vious land use, or the recovery process can be assistedin a variety of ways to overcome existing limitations(hereafter termed assisted natural regeneration).Assisted natural regeneration interventions may noteffectively overcome limitations, thus requiring activerestoration using site preparation and tree planting(Holl and Aide 2011, Holl et al 2018). Continuousplantings, cluster plantings (Saha et al 2016), andplanting islands or corridors of native trees are effec-tive ways to actively restore forests and to encouragetheir development through subsequent natural regen-eration (Holl 2017, Levy-Tacher et al 2019).

Despite many social and ecological obstacles, for-ests are regenerating in many regions worldwide(Hecht et al 2014) (figure 1). Throughout the world,woodlands and forests are returning following aban-donment of small-scale agriculture (Li and Li 2017). Itis time to recognize the many values of naturallyregenerating forests and to place this land-use changefirmly within the context of forward-looking environ-mental policies to create multi-functional landscapesthat sustain people and nature. Post-agricultural forestregeneration occurs within the context ofmultiple-uselandscapes, requiring attention to a wide range ofsocial as well as ecological issues, as highlighted by therecent IPBES global assessment (Díaz et al 2019). Thistask is reinforced by the fact that recent global meta-analyses related to forest restoration have found thatrecovery levels of biodiversity, forest structure andfunction indicators are similar or greater for passiverestoration than for active restoration in the long term,in spite of highly variable results among primary stu-dies (Crouzeilles et al 2017, Meli et al 2017, Jones et al2018).

Here, we review the social and ecological impor-tance of naturally regenerating forests on former agri-cultural land in temperate and tropical forest biomes.We summarize available information regarding whereforests are regenerating in agricultural landscapes, andexplore the conditions that influence their develop-ment and persistence. Finally, we examine specificcases where economic and regulatory policies posi-tively or negatively influence natural regeneration. Weconclude with recommendations for specific eco-nomic and policy interventions to enhance naturalregeneration in the context of international, national,and sub-national forest restoration targets.

Our review draws attention to the pervasive eco-nomic and policy contexts that currently influence(positively and negatively) natural regeneration of for-ests around the world. Given the global urgency andambition for large-scale forest restoration, our synth-esis provides a starting point for policy-level discus-sions and for developing approaches to enhancenatural regeneration on former agricultural land inways that promote long-term recovery while provid-ing economic benefits to rural residents.

2. Searchmethods

The articles featured in this review were selectedlargely through thematic literature searches and refer-ence list checking in addition to an extensive biblio-graphy on these topics accumulated from our activeresearch in this field. We searched published, peer-reviewed literature, emphasizing papers publishedsince 2015, using a wide variety of terms including‘land abandonment,’ ‘farm abandonment’, ‘foresttransition’, ‘secondary vegetation’, ‘forest expansion’,‘reforestation’, ‘regrowth’, ‘rewilding’, and ‘passive

2

Environ. Res. Lett. 15 (2020) 043002 RLChazdon et al

restoration’ in combination with ‘temperate’ and‘tropical’, and additional terms for specific geographicregions to uncover literature from Europe, Asia,Africa, and the Americas. We also used more specia-lized terms such as ‘enrichment planting’, ‘sustainablemanagement’, ‘remittances’, and ‘out-migration’ incombination with ‘forest regeneration’, ‘naturalregeneration’, and ‘secondary forests’. We eliminatedpapers that focused on silvicultural interventions inlogged forests or that focused on natural regenerationin the understory of plantations.

3. Environmental and socio-economicimportance of naturally regeneratingforests

3.1. Biodiversity recovery in naturally regeneratingforests and landscapesNatural regeneration of forests is an intersectionpoint for conservation and restoration goals (Arroyo-Rodriguez et al 2017, Chazdon 2019). Studies ofnaturally regenerating forests show gradual recoveryof native species compared to reference forests, butoutcomes vary widely and species composition recov-ery is significantly slower than species richness

(Chazdon et al 2009, Navarro and Pereira 2015,Acevedo-Charry and Aide 2019, Matos et al 2019,Rozendaal et al 2019). Agricultural land use can have acenturies-long legacy on the biodiversity and produc-tivity of forest ecosystems derived from old-fieldsuccession (Isbell et al 2019). During the first 40 yearsof natural regeneration in temperate areas across theglobe, organism abundance and diversity levelsattained 133% and 82%, respectively, of referenceforest levels (Meli et al 2017). A meta-analysis of 147studies in tropical regenerating forests found thatspecies richness of amphibians, reptiles, birds, andmammals recovered after approximately 40 years, butrecovery of species composition was considerablyslower, particularly for forest specialists (Acevedo-Charry andAide 2019). In Central Spain, Cruz-Alonsoet al (2019) reported recovery levels with respect toreference forests of 103% for woody species richness,45% for tree biomass, 39% for frugivore-dispersedshrub abundance, and 96% for tree functional disper-sion for a variety of secondary forests after 50 years ofagricultural abandonment. In lowland Latin America,tree species richness showed rapid recovery (mean of54 years) in naturally regenerating forests, but recoveryof species composition may require several centuries

Figure 1.Documented cases of large-scale natural forest regeneration in agricultural landscapes in temperate, tropical, and subtropicalforest regions of theworld. These illustrative studies vary in the time frames of analysis and in themetrics used to report changes innaturally regenerating forest cover.Herewe report changes as net gain in natural forest cover or percentage increases based on regionalland use and land cover assessments. In all cases, only natural regeneration is reported, and gains in planted tree cover are excluded.Tropical and subtropical forest biomes are indicated by dark green shading, whereas temperate and boreal forest biomes are indicatedas light green.

3

Environ. Res. Lett. 15 (2020) 043002 RLChazdon et al

(Rozendaal et al 2019). Natural regeneration inAustralian subtropical woodlands provides valuablehabitat for reptile and bird communities (Bowen et al2009, Bruton et al 2013). In tropical regions, recoveryof biodiversity and forest structure can be 34%–56%and 19%–56% higher, respectively, in naturallyregenerating forests than in actively restored forests(Crouzeilles et al 2017).

Biological legacies in the landscape (sensu Franklinet al 2000), i.e. the living organisms that survive a cata-strophic disturbance, contribute to and are created bynaturally regenerating forests, with spatial context andprior land use strongly influencing the future trajec-tory of communities and ecosystems (Bengtsson et al2003, Johnstone et al 2016). A meta-analysis based onnatural regeneration studies in 135 landscapes in tem-perate and tropical forest regions showed that theextent of forest cover in the landscape is the mostimportant predictor of landscape variability in recov-ery of biodiversity, a measure inversely related to eco-logical restoration success (Crouzeilles et al 2019).Restorable areas in landscapes (1×1 km pixel) withmore than 27% forest cover showed low levels of var-iation in biodiversity recovery, and encompass a totalof 238 M ha, 38% of the temperate and tropical forestregions of the world (Crouzeilles et al 2019). Theseareas present lower risks (higher predictability) forbiodiversity recovery through natural regeneration. Incontrast, landscapes with less than 6% forest covershowed high levels of variation in recovery, and arebetter candidates for active restoration or reforestationinterventions (Crouzeilles et al 2019).

At a landscape scale, naturally regenerating forestscan cost-effectively contribute to the conservation andrestoration of biodiversity through the creation of buf-fer zones, establishment of biological corridors andstepping stones in an agricultural matrix, and recoveryof disturbed areas within protected areas (Guevara et al2005, Evans et al 2017, Newmark et al 2017). Forestfragmentation could be reduced by 44% in the Brazi-lian Atlantic Forest if the 210 000 km2 of land with ahigh capacity for spontaneous and assisted naturalregeneration were left to recover (Crouzeilles et al2020). In temperate agricultural southern Australia,shelterbelts composed of natural regeneration can actas critical habitats for a range of native biota while pro-tecting crops from wind and storm damage and redu-cing erosion (Lindenmayer et al 2016). Naturallyregenerating forests can support markedly differentassemblages compared to planted forests and oldgrowth temperate woodland (Lindenmayer et al 2012).Secondary forests in the Brazilian Amazon show highlevels of landscape-scale diversity and contribute tohabitat heterogeneity (Solar et al 2015). In a frag-mented landscape in Central Amazonia, naturalregeneration of deforested areas between remnant frag-ments promoted the conservation of birds (Stoufferet al 2011), dung beetles (Quintero and Roslin 2005,Bitencourt et al 2019), and bats (Rocha et al 2018). In

Europe, agricultural land abandonment is the majordriver of population expansion of large herbivores andcarnivores (Perino et al 2019).

Effects of climate change on forest regenerationare a major concern (Bastin et al 2019). Because colo-nizing species are adapted to local conditions, and toother colonizing taxa (Chazdon 2014), naturallyregenerating forests are more resilient to drought, dis-ease, windstorms, or heavy rainfall than single-speciestree plantations (Jactel et al 2017). Droughts andtemperature increases associated with climate changecan influence rates and quality of vegetation recoveryin naturally regenerating forests and in other typesof restored forests (Anderson-Teixeira et al 2013,Locatelli et al 2015,Uriarte et al 2016a, 2016b).

3.2. Naturally regenerating forests as sources ofecosystem servicesRecovery of ecosystem functions exhibits similarpatterns between naturally regenerating and plantedforests (Meli et al 2017). At a global scale, forestsregenerating on land historically cleared for agricul-ture or timber clear-cuts constitute a significant globalcarbon sink (Pan et al 2011, Griscom et al 2017,Houghton and Nassikas 2017). Pugh et al (2019)estimated that regenerating forest stands (< 140 yearold) encompassed 61.5% of the 42.8 million km2 offorests globally in 2010. From2001 to 2010, the carbonsink from regenerating forests (1.3 Pg yr−1) consti-tuted 60.5% of the global forest carbon sink of2.15 Pg yr−1. Carbon sinks in regenerating forests arelocated mostly in deciduous broadleaf and evergreenconiferous forests in temperate zones, whereasmost ofworld’s remaining old-growth forest stands are in themoist tropics and boreal Siberia (Pugh et al 2019).Chazdon et al (2016b) estimated a total of 2.9 millionkm2 of regenerating forests (

biodiversity (Queiroz et al 2014), alteration of waterflows (Bonnesoeur et al 2019, Evaristo andMcDonnell2019), and even a loss of cultural landscapes and tradi-tional land management techniques when humanmigration rates are high (Lasanta et al 2017). Naturalregeneration also can lead to increase of animal popu-lations that negatively affect agricultural productivityand human health (Byg et al 2017). These concernsalso apply to active restoration and reforestation inter-ventions, however, and underscore the need for broadstakeholder engagement in decisions regarding man-agement of landscape-scale interventions.

3.3. Economic benefits of naturally regeneratingforestsNatural regeneration can bring direct and indirecteconomic benefits to local residents and communities.Under supportive policies and market development,natural regeneration can enhance, diversify, andincrease long-term productivity of agricultural sys-tems (Peltier et al 2014), including silvopastoralsystems (Hoosbeek et al 2016,Kremen andMerenlender2018). In the temperate woodlands of south-easternAustralia, natural regeneration is a key component ofintegrating enhanced agricultural production andbiodi-versity conservation (Lindenmayer et al 2018). Naturallyregenerating woodlands can act as shelterbelts forprotecting livestock and thereby promoting lambingsuccess as well as weight gain in cattle (Cleugh 2003).Areas of naturally regenerated rainforest that occurwithin oil palm plantations have been shown to supportlarge numbers of native animals and plants (Azhar et al2014).

Over a 20 year period, the economic benefits ofnatural regeneration can compensate for the opportu-nity costs of foregoing agricultural use of these lands(Strassburg et al 2016). For example, reduction of sedi-ment loads through regeneration of abandoned pas-tures in the Paraitinga River Basin of São Paulo State inBrazil was estimated to reduce costs of dredging sedi-ments out of the river by US$1.17 million annually,and would avoid additional costs of water purification(Strassburg et al 2016). Natural regeneration also cancreate income streams from community-based eco-tourism, which brings financial returns to local resi-dents in addition to providing conservation benefitsfor wildlife and provision of ecosystem services (Stemet al 2003, Bray 2016).

Compared to natural regeneration, direct eco-nomic returns from commercial tree plantations andtree planting are higher and more predictable for tim-ber products in the short-term (Baral et al 2016).Indirect economic benefits from natural regenerationcan be substantial, however. Through retention ofnutrients in buffer strips and hedgerows, which canarise from natural regeneration, crop yields can beenhanced. Hedgerows bordering agricultural crop-lands in the temperate regions of the world retain 69%

of nitrogen, 67% of phosphorous, and 91% of sedi-ments of run-off (Van Vooren et al 2017). In theHumbo community-based natural regeneration pro-ject in Ethiopia, assisted natural regeneration broughtsocial and economic benefits to participating commu-nities who collected wild fruits, firewood and fodder(Wolde et al 2016).

A major advantage of natural regeneration as anecological restoration approach is the substantiallyreduced implementation costs compared to treeplanting (Brancalion et al 2016, Cruz-Alonso et al2019). In Atlantic Forest landscapes with relativelyhigh forest cover, where natural regeneration is mostlikely, costs of site preparation and tree planting arereduced by 38% (Molin et al 2018). Because of theselower costs, considerably larger areas can be restoredusing assisted natural regeneration approaches comparedto widespread tree planting (Chazdon and Guariguata2016). In Minas Gerais State, Brazil, Nunes et al (2017)projected that spontaneous and assisted natural regenera-tion could effectively restore 15 000 km2 of forest over20 years. Across the entire Atlantic Forest region ofBrazil, 210 000 km2 of degraded lands can potentiallybe restored through assisted natural regeneration, redu-cing implementation costs by US$ 90.6 billion (77%)compared to active restorationmethods (Crouzeilles et al2020).

4.Where andwhy forests are growing back

4.1. Global indicators of natural forest regenerationfrom satellite imageryDespite many technical advances such as fine-scalesatellite imagery (including LIDAR), we still lack anaccurate and systematic assessment of where forestsare naturally regenerating around the world, largelydue to challenges in distinguishing between areas ofnative forest and tree plantations and to high rates ofreclearance of regenerating forests (Rudel 2005, Asneret al 2009, Vieira et al 2014, Chazdon et al 2016a, Reidet al 2019). Net increases in tree cover (includingplanted and unplanted tree cover) detected fromsatellite imagery in boreal and temperate biomes from2000 to 2010 can largely be explained by naturalregeneration of forests on abandoned agriculturallands (FAOandUNCCD2015).

Global scale analysis of satellite imagery from 1982to 2016 revealed that tree cover is changing in dra-matic ways across major geographic regions, with treecover gain attributed to both natural regeneration aswell as the establishment of tree plantations (Songet al 2018). A tree cover increase of 15% in the EasternUnited States was attributed to natural regeneration(Song et al 2018). The greatest increases in tree coverwere in Eastern Europe (35%), including EuropeanRussia and Carpathian montane forests (Song et al2018). In Eastern Europe, tree cover gain was attrib-uted to natural forest regeneration on abandoned

5

Environ. Res. Lett. 15 (2020) 043002 RLChazdon et al

agricultural land following the collapse of the formerSoviet Union (Potapov et al 2015, Rudel et al 2016,Buitenwerf et al 2018). Political changes in EasternEurope and land-use subsidies in the European Unionto set aside marginal agricultural areas in regions withsteep slopes to limit food production and avoid sur-pluses (Common Agricultural Policy reforms) led toabandonment of farmland from 1998 to 2008 (Lasantaet al 2017).

4.2. Natural forest regeneration in EuropeAbandonment of agriculture in mountainous regionsin Europe led to both expansion of plantations andnatural regeneration in many countries over the lastcentury, accompanied by rural out-migration andintensification of agriculture in lowland regions(Benayas et al 2007, Sitzia et al 2010, Cruz-Alonso et al2019). In Italy, forest cover increased by 87% since theend of World War II, with the greatest areas of forestregrowth in lowland areas, where abandonment offarmland and the loss of traditional rural landscapeshas occurred as a result of industrialization, urbaniza-tion, and agricultural intensification elsewhere(Camarretta et al 2018).Within the Basilicata region ofsouthern Italy, approximately 70 154 ha of forestregenerated on abandoned agricultural lands andpastures from 1984 to 2010 (Mancino et al 2014). InSpain, natural forest regeneration represented around2/3 of the increase in tree cover between 2000 and2010 (Vallejo et al 2014). A land-use dynamics modelpredicted that between 100 000 and 290 000 km2 ofagricultural land in Europewill be abandoned between2000 and 2030 (Verburg and Overmars 2009). Muchof this new tree cover is expected to result fromnaturalregeneration (Thers et al 2019).

4.3. Natural forest regeneration in the tropics andsubtropicsAnalyses of sequential satellite imagery and groundsurveys reveal many areas around the world wheretropical and subtropical forests are naturally regener-ating following agricultural land use at scales ofhundreds of km2 or greater (figure 1). In severalregions of Africa, farmer managed natural regenera-tion is occurring on former croplands and grazinglands (Smale et al 2018, Lembani et al 2019). Thisapproach has transformed an estimated 70 000 km2 ofdenuded dryland forest landscapes into productiveagroforestry parklands in Niger alone (Smale et al2018). Nanni et al (2019) identified 15 regions ofsustained natural regeneration of forests in LatinAmerica and the Caribbean between 2001 and 2014.Combined, these regions covered 2.2 M km2, repre-senting 11% of the region’s land area. One of theseregions was the tropical Andes, where 5000 km2 ofwoody vegetation regrew over this period (Aide et al2019), associated with a decline in rural populationand out-migration to urban areas.

Brazil’s Atlantic Forest is another region with sig-nificant natural regeneration (Nanni et al 2019). Forestcover increased by 102% in the Paraiba Valley of SãoPaulo, Brazil from 1962 to 2011, dominated by naturalregeneration on abandoned cattle pastures (Lira et al2012, da Silva et al 2017, Calaboni et al 2018). Theseland-use changes appear to be driven by agriculturalexpansion and intensification on the most suitableagricultural lands, which encouraged abandonment ofmarginal agricultural lands. Similar trends applyacross the entire Atlantic Forest Region of Brazil. Inthis region with 755 000 km2 of deforested land,27 000 km2 of forest regenerated naturally from 1996to 2015, and a predictivemodel estimated that another28 000 km2 could regrow between 2015 and 2035without human assistance (Crouzeilles et al 2020).Using assisted natural regeneration methods, an addi-tional 188 000 km2 of Atlantic Forest in Brazil has thepotential to be restored (Crouzeilles et al 2020).

Natural regeneration also occurs in regions thatare still undergoing net deforestation. In Brazil’s arc ofdeforestation in Pará State, naturally regenerating for-ests are increasing dramatically following abandon-ment of cattle pastures. Across the Brazilian Amazon,natural regeneration increased five-fold over the lastthree decades, exceeding 150 000 km2 in 2012 (Aguiaret al 2016). Extensive areas of natural regeneration inAmazonia are often observed in areas close to largeremnant patches of forest and low intensity of landuse (Jakovac et al 2015, Lennox et al 2018). Along a1000 km stretch of the BR-163 highway, natural regen-eration adjacent to forests contributed to 85% and70% in Pará and Mato Grosso, respectively, of all for-est regrowth detected between 1985 and 2012 (Mülleret al 2016). Absolute rates of natural regeneration werestrongly dependent on the overall amount of defor-ested area, with higher rates in Pará (maximumof 50%of deforested area) on former pastures with lowermanagement intensity compared to Mato Grosso(maximum of 25% of deforested area) where capital-intensive cropland and pasture systems dominate(Müller et al 2016). In the Brazilian Amazon, Conradoda Cruz et al (2020) identified 405 forest restorationprojects in 191municipalities between 1950 and 2017,forest restoration techniques used in descending orderof importance were seedling planting, agroforestrysystems, assisted natural regeneration, and naturalregeneration.

Compared to subtropical and temperate zones,natural regeneration on former agricultural land in thetropics tends to be a more recent phenomenon, wherenet forest loss is still occurring (Song et al 2018). Tro-pical secondary forests are younger (mean of 18 year)compared to temperate deciduous forests (mean of52 year) and coniferous evergreen forests (mean of72 year) (Pugh et al 2019). In Latin America, casesof forest gain through natural regeneration from 2001to 2014 fell into five main clusters that reflect topo-graphic features and related aspects of agro-ecological

6

Environ. Res. Lett. 15 (2020) 043002 RLChazdon et al

marginality, climate change, rural population decline,and increased urbanization (Nanni et al 2019).Broader analysis of global patterns showed that dis-tinct regional contexts have given rise to significantcases of net reforestation (Li and Li 2017).

4.4. Local, landscape and regional drivers of naturalforest regenerationNatural regeneration reflects myriad drivers and con-texts of land-use change. Environmental factors thatcan influence natural regeneration include soil quality,the presence of weedy or invasive species that arrestthe natural regeneration process, or inadequate seeddispersal that restricts colonization of native species(Rey Benayas et al 2008). Observational studies haveshown that the loss of primates and birds negativelyinfluences forest regeneration (Gardner et al 2019). Inthese cases, interventions are needed to control weeds,enrich natural regeneration, and enhance seed disper-sal. Natural regeneration also can be assisted bycontrolling or eliminating grazing livestock and pre-venting wildfires (Fischer et al 2009, Gardner et al2019). The diversity of local tree regeneration can besupplemented by enrichment planting of importantlocal species or non-invasive commercial species for

later harvesting of timber and non-timber products(Paquette et al 2009,Maier et al 2018) (Box 1).

In some areas of Australia, natural regeneration isoccurring as a result of reduced grazing pressure eitherthrough deliberate limiting of grazing pressure ordrought-related destocking (Fischer et al 2009, Geddeset al 2011). In other contexts, deliberate interventionsto assist natural regeneration include establishment ofprotected areas (Von Thaden et al 2018), fenced enclo-sures on farms (Mekuria et al 2018), reforestation onprivate lands in compliance with mandatory restora-tion policies (Brancalion et al 2016), and voluntaryactions to enhance conservation values in amenitylandscapes (Stelling et al 2018).

The ecological determinants of natural regenera-tion have been investigated in a variety of contexts(Chazdon 2014) and provide the basis for land-useplanning within farms, landscapes, and munici-palities. Planning the location of naturally regeneratedareas relative to other parts of farms such as grazingpaddocks, watercourses, and rocky outcrops is criticalto effectively integrate agricultural production withareas of natural regeneration (Lindenmayer et al2016). In agricultural landscapes, patches of forestregeneration are more likely to be found adjacent toexisting old-growth forest remnants (Sloan et al 2016),and natural regeneration is more likely to occur andhave better biodiversity outcomes in landscapes withmore forest cover (Crouzeilles et al 2016, 2020).Deforested areas on steep slopes with less intensiveprior land use and close to forest remnants are themost likely to regenerate spontaneously (Rezende et al2015, Molin et al 2018). A systematic review of driversof tropical forest cover expansion through naturalregeneration found that proximity to forest remnants,steep slopes, high forest cover at the landscape scaleand proximity to watercourses were the most impor-tant biophysical factors (Borda-Niño et al 2020). Nat-ural regeneration is often associated with poor soilquality or other proxies of agricultural marginality(Arroyo-Mora et al 2005), but this trend is not uni-versally observed (Sloan et al 2016).

Important socio-economic factors associated withtropical forest cover expansion through natural regen-eration were inclusion in protected areas, distance toroads, and distance to population centers (Borda-Niño et al 2020). Land tenure regimes also significantlyaffected recovery of woody natural regeneration inMexico. Municipalities dominated by communal landtenure showed the largest increases in forest coverfrom 2001 to 2010 inmoist forest, dry forest, and con-iferous forest biomes (Bonilla-Moheno et al 2013).

In Mesoamerica and South America, agriculturalabandonment is associated with the expectation ofincreasing economic opportunities from jobs innearby cities, ecotourism operations, or industrialzones, and is often accompanied by out-migrationfrom rural areas (Hecht et al 2015). Similar trendsoccur in Nepal, where levels of international

Box 1.Management of natural forest regeneration in theAmericantropics.

Management of naturally regenerated forests on former agricultural

land for commercial products is relatively uncommon. Based on

studies in the Latin American tropics and subtropics, however, we

know that these forests holdmuch potential formanagement for

timber and non-timber products (Kammesheidt 2002). Theseyoung forests are rich sources of a wide variety of products such as

medicine, ornamental plants, food, timber, and fuel (ChazdonandCoe 1999, Guariguata 1999, Souza et al 2016). Experimentalstudies inCentral Amazonia andCosta Rica show a high potential

for enhancing growth and survival of timber species in naturally

regenerating tropical forests though creating canopy gaps,

removing understory vegetation andmanipulating leaf litter

(Mesquita 2000,Dupuy andChazdon 2008). In Puerto Rico,widespread naturally regenerating forests contain high densities

of trees suited for timber and non-timber products, although

many forests are still too young to support extractive activities

(Forero-Montaña et al 2019).Managed natural regeneration incoastal areas of Brazil’s Atlantic Forest also showed high diversity

and abundance of useful species, including two endemic species,

and is providing economic benefits to smallholders (Souza et al2016). In a study of two 33 year old naturally regenerating forestsin Brazil’s Atlantic Forest, onemanaged through enrichment

planting and one unmanaged, Fantini et al (2019) found thatselective harvesting could produce valuable timber fromplanted

and unplanted species, while permitting growth for future har-

vests. Small-scalemanagement of secondary forest in this region

has the potential to produce sufficientmerchantable timber to

become an incentive for land owners tomaintain and recover for-

est on their farms. Enrichment of young regenerating forests with

native palm species used for commercial fruit production and

timber species generated an economically viable production

model over a 30 year period in theAtlantic Forest of southeastern

São Paulo, Brazil (Maier et al 2018).

7

Environ. Res. Lett. 15 (2020) 043002 RLChazdon et al

outmigration are high (Oldekop et al 2018). Interna-tional outmigration in Nepal was associated with sub-stantial increases in local forest cover due to farmlandabandonment and subsequent natural regeneration offorest (Oldekop et al 2018).

The social and cultural costs of rural migrationmay be high, including exploitation and increasingpoverty (Garcia-Barrios et al 2009, Hecht et al 2015).In some areas, influx of remittances following out-migration can partially compensate for losses of agri-cultural labor, sustaining some traditional farmingactivities in these areas (Ospina et al 2019). The influxof remittances varies greatly, however, depending onexternal economic and political conditions. Forinstance, remittances accounted for approximately25% of Nepal’s Gross Domestic Product in 2013(Oldekop et al 2018).

5. Economic and policy barriers to naturalforest regeneration

Soil degradation (often caused by intensive and long-term land-use), climate harshness, and low levels ofneighboring natural forest cover are major impedi-ments to natural regeneration around the world(Jakovac et al 2015, Sato et al 2019). Aside from thesebiophysical constraints, natural regeneration facesadditional socio-economic and jurisdictional barriers.In the following paragraphs, we focus on barriers tonatural regeneration due to regulations, policies andglobal economic trends that favor intensifiedmodes ofcommodity production, restrictive forest conservationmeasures, and large-scale tree monocultures. Thesebarriers also pose challenges to the widespread adop-tion of active forest restoration approaches involvingplanting of native tree species. In the tropics, intensiveagricultural production systems for palm oil, soy-beans, sugarcane, pineapples, and other crops requireremoval of trees from parts of the landscape thathinder mechanized or intensive production, such asflat areas in lowlands, where young patches of naturalregeneration are frequently eliminated (Sayer et al2012, Shaver et al 2015).

Additional barriers stem from the ‘invisibility’ ofnatural regeneration in the context of reforestationand forest restoration. Decision-makers, resourcemanagement agencies, farmers, and restoration prac-titioners tend to overlook natural restoration-basedapproaches for at least six reasons. First, large-scalerestoration initiatives are often conceived solelythrough tree planting (Chazdon and Uriarte 2016,Biggs 2018, Hua et al 2018). Second, farmers view earlystages of natural regeneration as undesirable andmessy, or as a sign of poor land management (Zahawiet al 2014). Third, limited knowledge is available toguide policies and actions to target where naturalregeneration could potentially occur, to estimate howmuch area could be regenerated, and how long it takes

to deliver specific social and environmental outcomes(Uriarte and Chazdon 2016). Fourth, there is a lack ofsound economic projections and business modelsbased on natural regeneration to evaluate socio-eco-nomic effectiveness (Ding et al 2017). Fifth, naturalregeneration has not been considered an activityrequiring human agency and therefore cannot beenforced as a policy. And sixth, in some countries,agrarian reform laws obligate farmers to cultivate land,and state authorities can confiscate uncultivated landor declare fallow land as ‘unutilized or degraded land’to be used for other purposes (Ferguson 2014, Duang-jai et al 2015).

In commodity production landscapes, naturalregeneration in suitable areas presents high opportunitycosts and requires that landowners receive appropriatefinancial compensation to transform agricultural landinto natural forest. Payments for environmental servicesto landowners in Costa Rica are USD $125/ha/yr fora 16 year contract to establish a native tree speciesplantation, but only USD $39/ha/yr for a 5 yearcontract for protecting natural regeneration (Porras andChacón-Cascante 2018). Given the choice, landownersfavor clearing young secondary forest to establish treeplantations or for growing commodity crops overregenerating native forest (Shaver et al 2015). Naturallyregenerating forests can actually support a highabundance of commercial tree species (Box 1), but treescan take several decades to reach commercial size(Forero-Montaña et al 2019). The economic valueof naturally regenerating forests is often considerablylower than a commercial forestry-style plantation,agroforestry system, or cropfield.

Older stages of natural regeneration and primaryforests are now legally protected from clearing inmany countries, but early growth stages are rarely pro-tected and are commonly (and sometimes legally)cleared to make way for crop or cattle production.Outside of protected areas, young stages of naturalregeneration are highly vulnerable to being re-cleared(Schwartz et al 2017, Reid et al 2019). In the BrazilianAmazon, 42 040 km2 of secondary forests derivedfromnatural regeneration of abandoned pastures wereconverted into other types of land cover between 2010and 2014 (Carvalho et al 2019). From 2008 to 2014,deforestation of secondary forests in Brazilian Amazo-nia became decoupled from deforestation of primaryforests, suggesting a trend toward pasture manage-ment based on reclearance of young forests (Wang et al2020). In Costa Rica, recent expansion of pineappleand other crops largely replaced pasture, exotic andnative tree plantations, and secondary forests, as 1986legislation strictly prohibits clearance of primary for-est (Shaver et al 2015). Environmental legislation tendsto look backward rather than forward, emphasizingprotection of historical conditions (preventing loss ofprimary forests) rather than ensuring the futurepotential for landscape-scale restoration and forest

8

Environ. Res. Lett. 15 (2020) 043002 RLChazdon et al

connectivity, which strongly influence future levels ofbiodiversity and ecosystem services.

Naturally regenerated forest on former agri-cultural land is generally poorly mapped for planningand decision-making purposes. Forest gain is rarelydisaggregated into its components of natural and plan-ted forests. The importance of natural regeneration isalso easily overlooked because often it is not shown onamap (figure 1). Estimates of deforestation in the Bra-zilian Amazon using the Brazilian national satellite-based deforestation monitoring system PRODES donot include deforestation of secondary forests. Yet,clearing of secondary forests and woodlands for agro-industrial pastures, plantations, and small-scale agri-cultural activities contributes significantly to forestloss in some areas. One exception is the TerraClassland-use mapping system used in Brazil that classifiessecondary forest, pasture with woody regeneration,and regeneration with pasture as distinct categories(Almeida et al 2016). This approach revealed that19.2% of previously deforested areas in Mato GrossoState, Brazil in 2008 were undergoing naturalregeneration.

Even when owners of small properties allow nat-ural regeneration and manage native forests on theirland, they are often legally prevented from managingthe young forest or selectively harvesting timber andnon-timber products. For example, once naturalregeneration reaches a stage when it is legally definedas forest in Mexico, harvesting restrictions and hightransaction costs reduce the economic benefitsreceived by small farmers (Román-Dañobeytia et al2014). Forest law in Bhutan stipulates that planted for-ests on private and communal property are consideredprivate property, and thus do not require state author-ization to harvest. But trees and forests establishednaturally, either on public or private land, is nationalforest patrimony and require a management plan andauthorization prior to utilization (Sears et al 2018a),moreover, timber harvested from natural forests issubject to taxation. In lowland Peru, a local market forthe pioneer treeGuazuma crinitamakes natural regen-eration economically profitable. But there are no fea-sible national regulatory mechanisms for low-incomesmallholder farmers to harvest timber from fallow for-ests that are cyclically cleared for agricultural use(Sears et al 2018b) and current legislation restricts thesale of timber from these systems.

Sectoral and jurisdictional policies also hampernatural regeneration. For example, land use planningin Peru falls under the mandate of the Ministry ofEnvironment, yet it is the Ministry of Agriculture thatgoverns land use change by issuing titles and permits.As a result, the Ministry of Environment has poorleverage to support conservation of natural regenera-tion in spite of implementing carbon-based paymentsand related incentives (Kowler et al 2016). Conflictingmandates across government sectors in the context of

who governs forest restoration interventions (whichinclude natural regeneration) are in fact widespreadacross most Latin American countries (Schweizer et al2020). In southern Australia, large patches of old growthwoodland on agricultural land are generally excludedfrom clearing under legislation, whereas natural regener-ating (regrowth) woodland is rarely protected and oftensubject to widespread clearing (https://environment.nsw.gov.au/questions/is-land-clearing-permitted), lead-ing to the loss of key habitats for biodiversity, especiallyduringdroughtperiods (Lindenmayer et al2019).

6. Policy options andmanagementinnovations to favor natural forestregeneration

Natural regeneration occurs under specific biophysi-cal, socio-economic and cultural conditions. How-ever, in most cases, it is the result of an unintentionalconsequence of other processes, such as rural out-migration, changes in commodity prices and exportpolicies, abandonment of agriculture on hilly or steeptopography that preclude mechanization and agricul-tural intensification, land abandonment due todroughts, or government restrictions on agriculturalland use on private or common property. Naturalregeneration occurs intentionally when previouslydeforested areas are newly incorporated into state-managed protected areas or partially deforested pri-vate land purchased with the intention of conservingand restoring native forests (Algeet-Abarquero et al2015), or when communities decide to promoteregeneration to form community-managed forestreserves that provide forest products and otherbenefits to local livelihoods (Levy-Tacher et al 2019).These cases illustrate different socio-economic, cul-tural, and political drivers and impacts. Compared totemperate and boreal zones, approaches for manage-ment of naturally regenerated forests in tropicalregions are poorly developed, particularly on formeragricultural land. Yet there is much potential forsilvicultural interventions in temperate and tropicalregenerating forests to promote management fortimber, non-timber products, carbon storage, andrecreational, cultural and educational activities (Levy-Tacher et al 2012, Cojzer et al 2014) (Box 1).

Policy changes could be more achievable now, ascapabilities have advanced to permit identification ofspecific areas where natural regeneration of forests isfeasible and beneficial to both the environment andlivelihoods. Natural regeneration is increasinglyrecognized as an important natural solution to tack-ling climate change (Chazdon et al 2016b, Griscomet al 2017), but its drivers and limitations need to beclearly identified. In cases where the major limitationsare socio-economic rather than biophysical, innova-tions in policies and economic incentives at multiple

9

Environ. Res. Lett. 15 (2020) 043002 RLChazdon et al

https://environment.nsw.gov.au/questions/is-land-clearing-permittedhttps://environment.nsw.gov.au/questions/is-land-clearing-permitted

levels will be needed to reach the scale needed torestore native forests around theworld.

Holistic land-use planning and spatial prioritizationapproaches can help ensure that native forests continueto regrow and persist without compromising food, fuel,or fiber production (Chazdon and Brancalion 2019).However, policies andmechanisms to empower holisticsolutions—including expansion of agroforestry andsilvopastoral systems—are underdeveloped (KremenandMerenlender 2018, Chazdon andBrancalion 2019).Economic and policy incentives will be needed aseconomies and markets transition from those drivingfurther degradation of native forests to restoration andenhancement of native forests (Boillat et al 2017). Wenow have the capacity to identify specific target areaswhere natural regeneration is beneficial and feasible(Brancalion et al 2019, Crouzeilles et al 2020), which canfacilitate policy changes. Further development of thesetargeted approaches will need to be accompanied byinnovative policies at multiple levels to reach the scaleneeded to restore native forests around the world byharnessing the power ofnature (table 1).

Nurturing a forest transition—particularly wherenatural regeneration is promoted— presents immensepolicy and institutional challenges (Sloan 2015). Thesechallenges are not insurmountable, but will requirefurther research and innovations in policy and govern-ance. For example, innovative institutional and policyapproaches in Costa Rica supporting agriculturalintensification, forest protection, and payments forenvironmental services contributed to a forest trans-ition process that led to overall environmental benefits(Jadin et al 2016) including native species plantationsand natural regeneration of forests (Calvo-Alvaradoet al 2019). We encourage a focus on creatingmulti-functional landscapes where forest regrowthis compatible with agricultural production andsustainable rural livelihoods, by rejecting narrowsectoral mandates that spawned conflicts betweenconservation, production, and land rights (Kremenand Merenlender 2018). Sustainable intensificationof agriculture and land sharing are key goals to

Table 1. Suggested interventions at the international, national, andsub-national scales to encourage natural regeneration tomeetnational and global forest restoration targets.

International scale

Create appropriate land type definitions. TheUnitedNations Stra-

tegic Plan for Forests (2017–2020) has a global goal of increasingforest area by 3%worldwide. The FAOdefinition of forest used

does not distinguish between native forests andmonoculture

plantations composed of exotic species. This goal should bemod-

ified to also include an aim to increase native forest area specifi-

cally through native tree plantings or assisted natural

regeneration

Produce a globalmap at a 30 m resolution spatial scale of natural

regeneration potential. Thismap should be based on the histor-

ical distribution of existing areas of natural regeneration (exclud-ing plantations), environmental (topography, proximity toremnant forests, and river systems) and socio-economic factors(prior land use, land distribution, poverty index, inequity index,commodity production, forestry, shifting cultivation, and human

migration dynamics). Thismap can show also the expected ecolo-gical outcomes fromnatural regeneration to reduce uncertainty

andmanage risk of low-cost forest restoration

Leverage the 2021–2030UNDecade on EcosystemRestoration, the

UNFrameworkConvention onClimate Change agenda and the

UNConvention on Biological Diversity to call for actions to

enhance the long-termpersistence of native forests (includingnatural regeneration) for biodiversity conservation, climatemiti-gation and adaptation, and hydrological regulation

National scale

Create a globalmap and national-scalemaps of natural regeneration

capacity in assessments of national-level restoration opportu-

nities. Identify restoration opportunities that are suitable for

unassisted or assisted natural regeneration

Increase efforts tomap and classify naturally regenerated forests that

include biophysical and socio-ecological land use dimensions

(Boillat et al 2017)Rebalance national forestmanagement policies to emphasize local

decision-making and to permit local or regional governance of

management policies for harvesting timber and non-timber pro-

ducts fromnaturally regenerated forests, including harvesting of

small diameter timber species and non-timber products while

creating new income streams

Develop a national programof enrichment planting of trees with

local commercial value or ecological value forwildlife to enhance

diversity andmanagement of secondary forest patches on private

farms or community-managed land

Train and build capacity for environmental and restoration profes-

sionals to becomenatural regeneration extension agents who

advise landowners and communities regarding prioritization of

areas, assisted natural regeneration techniques, and sustainable

management practices

Develop businessmodels for assisted natural regenerationwith

input from local communities (Maier et al 2018)Sub-national scale

Encourage landowners in areas suitable for natural regeneration to

wait 1–2 years prior to planting trees to assess whether the rate of

natural regeneration is sufficient, a policy currently applied in

states in Brazil (Brancalion et al 2016) as a good predictor oflonger term recovery (Holl et al 2018)

Stimulate ‘local forest’movements. Develop ‘adopt a forest’ pro-

grams for local communities and schools, supported byNGOs,

local government agencies, and local businesses partners. Provide

incentives for local stewardship and valuation of regenerating for-

ests and their importance for providing ecosystem services that

benefit local communities. Use local regenerating forests for cul-

tural, educational, and capacity building programs

Table 1. (Continued.)Respect, encourage and foster local community decisions for nat-

ural regeneration to achieve sustainablemanagement (Levy-Tacher et al 2019)

In areas appropriate for natural regeneration, apply the same value

in payments for environmental service programs for natural

regeneration as for tree plantations and reduce theminimumarea

requirement so smallholders can qualify and benefit from these

programs

Leverage theUNSustainableDevelopment goal of healthy rural live-

lihoods to create attractive options for small farmers to retain

land ownershipwhile earning off-farm income and increasing

native forest cover on their properties. Provide incentives via tax

credits and conservation or restoration easements to protect land

ownershipwhile enhancing native forest cover and increasing

conservation values

10

Environ. Res. Lett. 15 (2020) 043002 RLChazdon et al

promote food security and wellbeing of smallholdersin rural landscapes (Latawiec et al 2018, Liao andBrown 2018).

7. Conclusions: toward a sustainable ruralresurgence in forest landscapes

This review brings out on an emergent theme regard-ing the driving forces that operate from regional toglobal scales to influence natural regeneration: expan-sion of intensified and mechanized agriculture inlowlands and, in many cases, associated abandonmentof agricultural land, primarily in steep ormountainousareas that are poorly suited for this mode of agricul-tural expansion. In some cases, natural regeneration isassociated with rural outmigration or remittanceeconomies.

Reestablishing native forest cover does not have torequire mass exodus of families and decline of rurallivelihoods or traditions. We urge new ways of think-ing about how natural regeneration, coupled withother solutions, may promote a rural resurgencewhere communities and local economies thrive alongwith expansion of native forests. One challenge forpolicy initiatives that promote natural regeneration isto address the social costs and drivers of rural out-migration. Enhancing natural regeneration of nativeforests is not a viable option for forest restoration ifthese changes fail to provide benefits for rural resi-dents and forests are short-lived (Chazdon andBrancalion 2019).

In the new era of restoration, rural livelihoods canbe re-envisioned through new opportunities createdby growing native forests and trees in agriculturallandscapes. For example, the Sustainable Rural Devel-opment Program of Rio de Janeiro State, Brazil hasnow become public policy involving community-based rural development in micro-watersheds withthe support of rural organizations and decision-makers at local, municipal, and regional levels (Hissaet al 2019). Rural communities can become the stew-ards of community-managed forests that providelocal, regional and global benefits. Forests of all kindscan contribute to prosperity (Miller and Hajjar 2020),a healthier society (Colfer 2012), and mitigate climatechange (Griscom et al 2017). In many regions, youthand employable adults are leaving rural areas andabandoning a future relationship with land and withforests (Paudel et al 2014).We still have time to changethese trends and promote rural resurgence based onproactive and integrated land management and land-scape-scale restoration, where forests and new genera-tions of people have room to grow and prospertogether.

Acknowledgments

We thank the International Institute for Sustainabilityin Rio de Janeiro and the United States Agency forInternational Development (USAID) for the financialand logistical support for the workshop where thisreview was conceived. MRG acknowledges fundingfrom the CGIAR Program on Forests, Trees andAgroforestry. JMRB acknowledges funding from theREMEDINAL project (TE-CM S2018/EMT-4338)funded by theMadrid AutonomousGovernment.

Data availability statement

Any data that support the findings of this study areincludedwithin the article.

References

Acevedo-CharryO andAide TM2019Recovery of amphibian,reptile, bird andmammal diversity during secondary forestsuccession in the tropicsOikos 128 1065–78

Aguiar A P et al 2016 Land use change emission scenarios:anticipating a forest transition process in the BrazilianAmazon?Glob. Change Biol. 22 1821–40

Aide TM,GrauHR,Graesser J, Andrade‐NuñezM J, Aráoz E,Barros A P, Campos‐CerqueiraM,Chacon‐Moreno E,FranciscoCC and Espinoza R 2019Woody vegetationdynamics in the tropical and subtropical andes from2001 to2014: satellite image interpretation and expert validationGlob. Change Biol. 25 2112–26

Algeet-AbarqueroN, Sánchez-Azofeifa A, Bonatti J andMarchamaloM2015 Land cover dynamics in osa region,costa Rica: secondary forest is here to stayReg. Environ.Change 15 1461–72

AlmeidaCAD, CoutinhoAC, Esquerdo J CDM,AdamiM,Venturieri A, Diniz CG,DessayN,Durieux L andGomesAR2016High spatial resolution land use and land covermappingof the Brazilian Legal Amazon in 2008 using Landsat-5/TMandMODIS dataActa Amazonica 46 291–302

Anderson-Teixeira K J,Miller AD,Mohan J E,Hudiburg TW,Duval BD andDelucia EH 2013Altered dynamics of forestrecovery under a changing climateGlob. Change Biol. 192001–21

Arroyo-Mora J P, Sanchez-Azofeifa GA, Rivard B, Calvo J C andJanzenDH2005Dynamics in landscape structure andcomposition for theChorotega region, Costa Rica from1960to 2000Agric. Ecosyst. Environ. 106 27–39

Arroyo-Rodriguez V,Melo F,Martinez-RamosM, Bongers F,ChazdonR,Meave J, NordenN, Santos B, Leal I andTabarelliM 2017Multiple successional pathways in human-modified tropical landscapes: new insights from forestsuccession, forest fragmentation and landscape ecologyresearchBiol. Rev. 92 326–40

AsnerGP, Rudel TK, Aide TM,Defries R and EmersonR 2009Acontemporary assessment of change in humid tropical forestsConserv. Biol. 23 1386–95

Azhar B, LindenmayerDB,Wood J, Fischer J andZakariaM2014Ecological impacts of oil palm agriculture on nativemammalrichness and feeding guilds in PeninsularMalaysiaBiodiversity Conservation 23 1175–91

BaralH,GuariguataMR andKeenanR J 2016A proposedframework for assessing ecosystem goods and services fromplanted forestsEcosyst. Serv. 22 260–8

Bastin J-F, Finegold Y,Garcia C,MolliconeD, RezendeM,RouthD,ZohnerCMandCrowther TW2019The global treerestoration potential Science 365 76–9

11

Environ. Res. Lett. 15 (2020) 043002 RLChazdon et al

https://doi.org/10.1111/oik.06252https://doi.org/10.1111/oik.06252https://doi.org/10.1111/oik.06252https://doi.org/10.1111/gcb.13134https://doi.org/10.1111/gcb.13134https://doi.org/10.1111/gcb.13134https://doi.org/10.1111/gcb.14618https://doi.org/10.1111/gcb.14618https://doi.org/10.1111/gcb.14618https://doi.org/10.1007/s10113-014-0714-9https://doi.org/10.1007/s10113-014-0714-9https://doi.org/10.1007/s10113-014-0714-9https://doi.org/10.1590/1809-4392201505504https://doi.org/10.1590/1809-4392201505504https://doi.org/10.1590/1809-4392201505504https://doi.org/10.1111/gcb.12194https://doi.org/10.1111/gcb.12194https://doi.org/10.1111/gcb.12194https://doi.org/10.1111/gcb.12194https://doi.org/10.1016/j.agee.2004.07.002https://doi.org/10.1016/j.agee.2004.07.002https://doi.org/10.1016/j.agee.2004.07.002https://doi.org/10.1111/brv.12231https://doi.org/10.1111/brv.12231https://doi.org/10.1111/brv.12231https://doi.org/10.1111/j.1523-1739.2009.01333.xhttps://doi.org/10.1111/j.1523-1739.2009.01333.xhttps://doi.org/10.1111/j.1523-1739.2009.01333.xhttps://doi.org/10.1007/s10531-014-0656-zhttps://doi.org/10.1007/s10531-014-0656-zhttps://doi.org/10.1007/s10531-014-0656-zhttps://doi.org/10.1016/j.ecoser.2016.10.002https://doi.org/10.1016/j.ecoser.2016.10.002https://doi.org/10.1016/j.ecoser.2016.10.002https://doi.org/10.1126/science.aax0848https://doi.org/10.1126/science.aax0848https://doi.org/10.1126/science.aax0848

Benayas J,Martins A,Nicolau J and Schulz J 2007Abandonment ofagricultural land: an overview of drivers and consequencesCABRev.: Perspect. Agric., Veterinary Sci., NutritionNat.Resour. 2 1–14

Bengtsson J, AngelstamP, Elmqvist T, EmanuelssonU, FolkeC,IhseM,Moberg F andNyströmM2003Reserves, resilienceand dynamic landscapesAMBIO: J. Human Environ. 32389–96

BiggsD 2018Clearing, ‘wasting,’ and regreening: an environmentalhistory of bare hills in central Vietnam J. Asian Stud. 771037–58

Bitencourt B S,Da Silva PG,Morato E F andDe LimaYG2019Dung beetle responses to successional stages in the amazonrainforestBiodiversity Conservation 28 2745–61

Boillat S, Scarpa FM, Robson J P, Gasparri I, Aide TM,Aguiar A PD,Anderson LO, BatistellaM, FonsecaMGandFutemmaC2017 Land system science in Latin America:challenges and perspectivesCurr. Opin. Environ. Sustain. 2637–46

Bonilla-MohenoM,RedoD J, Aide TM,ClarkML andGrauHR2013Vegetation change and land tenure inMexico: acountry-wide analysis LandUse Policy 30 355–64

BonnesoeurV, Locatelli B,GuariguataMR,Ochoa-TocachiB F,VanackerV,MaoZ, StokesA andMathez-Stiefel S-L 2019Impacts of forests and forestation onhydrological services inthe andes: a systematic reviewForest Ecol.Manage.433 569–84

Borda-NiñoM,Meli P andBrancalion PHS 2020Drivers oftropical forest cover increase: a systematic review LandDegrad.Dev.Accepted (https://doi.org/10.1002/ldr.3534)

BowenME,McalpineCA, Seabrook LM,House APNandSmithGC2009The age and amount of regrowth forest infragmented brigalow landscapes are both important forwoodland dependent birdsBiol. Conservation 142 3051–9

Brancalion PH, SchweizerD,GaudareU,Mangueira J R,Lamonato F, Farah FT,NaveAGandRodrigues RR 2016Balancing economic costs and ecological outcomes of passiveand active restoration in agricultural landscapes: the case ofBrazilBiotropica 48 856–67

Brancalion PH et al 2019Global restoration opportunities intropical rainforest landscapes Sci. Adv. 5 eaav3223

BrayDB2016Muir and pinchot in the sierra norte of oaxaca:governance of forestmanagement and forest recovery inpueblosmancomunadosWorldDev. Perspect. 4 8–10

BrutonM J,McalpineCA andMaronM2013Regrowthwoodlandsare valuable habitat for reptile communitiesBiol.Conservation 165 95–103

Buitenwerf R, Sandel B,Normand S,Mimet A and Svenning J C2018 Land surface greening suggests vigorous woodyregrowth throughout European semi‐natural vegetationGlob. Change Biol. 24 5789–801

Bullock JM, Aronson J,NewtonAC, Pywell R F andRey-Benayas JM2011Restoration of ecosystem services andbiodiversity: conflicts and opportunitiesTrends Ecol. Evol. 26541–9

BygA,Novo P,DinatoM,Moges A, Tefera T, Balana B,Woldeamanuel T andBlackH 2017Trees, soils, andwarthogs–distribution of services and disservices fromreforestation areas in southern Ethiopia Forest Policy Econ. 84112–9

Calaboni A, Tambosi L, Igari A, Farinaci J,Metzger J P andUriarteM2018The forest transition in São Paulo, Brazil:historical patterns and potential drivers Ecol. Soc. 23

Calvo-Alvarado J, Jiménez V,Calvo-ObandoA andCastilloM2019Current perspectives on forest recovery trends inGuanacaste,Costa Rica Int. Forestry Rev. 21 425–31

CamarrettaN, Puletti N, ChiavettaU andCorona P 2018Quantitative changes of forest landscapes over the lastcentury across Italy Plant Biosyst.—Int. J. Dealing Asp. PlantBiol. 152 1011–9

CarvalhoR, AdamiM, Amaral S, Bezerra FG andDeAguiar APD2019Changes in secondary vegetation dynamics in a contextof decreasing deforestation rates in Pará, Brazilian AmazonAppl. Geogr. 106 40–9

ChazdonR andBrancalion P 2019Restoring forests as ameans tomany ends Science 365 24–5

ChazdonRL 2014 SecondGrowth: The Promise of Tropical ForestRegeneration inAnAge of Deforestation (Chicago, IL:University of Chicago Press)

ChazdonRL 2019Towardsmore effective integration of tropicalforest restoration and conservationBiotropica 51 463–72Accepted

ChazdonRL, Brancalion PH, Laestadius L, Bennett-Curry A,BuckinghamK,KumarC,Moll-Rocek J, Vieira I CG andWilson S J 2016aWhen is a forest a forest?Forest conceptsand definitions in the era of forest and landscape restorationAmbio 45 538–50

ChazdonRL andCoe FG 1999 Ethnobotany ofwoody species insecond-growth, old-growth, and selectively logged forests ofnortheasternCosta RicaConserv. Biol. 13 1312–22

ChazdonRL andGuariguataMR2016Natural regeneration as atool for large-scale forest restoration in the tropics: prospectsand challengesBiotropica 48 716–30

ChazdonRL, Peres CA,DentD, Sheil D, LugoAE, LambD,StorkNE andMiller S E 2009The potential for speciesconservation in tropical secondary forestsConserv. Biol. 231406–17

ChazdonRL andUriarteM2016Natural regeneration in thecontext of large-scale forest and landscape restoration in thetropicsBiotropica 48 709–15

ChazdonRL et al 2016bCarbon sequestration potential of second-growth forest regeneration in the Latin American tropics Sci.Adv. 2 e1501639

CleughH2003Trees for Shelter—AGuide toUsingWindbreaks onAustralian Farms (Canberra: Rural Industries Research andDevelopment Corporation)

CojzerM,Diaci J and Brus R 2014Tending of young forests insecondary succession on abandoned agricultural lands: anexperimental study Forests 5 2658–78

Colfer C J P 2012HumanHealth and Forests: AGlobal Overview ofIssues, Practice, and Policy. (London: Earthscan)

CramerVA,Hobbs R J and Standish R J 2008What’s new about oldfields?Land abandonment and ecosystem assemblyTrendsEcol. Evol. 23 104–12

Crouzeilles R, Barros F S,Molin PG, FerreiraMS, Junqueira AB,ChazdonRL, LindenmayerDB, Tymus J RC,Strassburg B BN andBrancalion PHS 2019Anew approachtomap landscape variation in forest restoration success at theglobal scale J. Appl. Ecol. 56 2675–86

CrouzeillesR,CurranM,FerreiraMS, LindenmayerDB,GrelleCEandBenayas JMR2016Aglobalmeta-analysis on theecological drivers of forest restoration successNat.Commun.7

Crouzeilles R, FerreiraMS, ChazdonRL, LindenmayerD,Sansevero J B B,Monteiro L, IribarremA, Latawiec A andStrassburg B BN2017 Ecological restoration success is higherfor natural regeneration than for active restoration in tropicalforests Sci. Adv. 3 e1701345

Crouzeilles R et al2020Achieving cost-effective landscape-scale forestrestoration through targetednatural regenerationConservationLett.Accepted (https://doi.org/10.1111/conl.12709)

Cruz-AlonsoV, Ruiz-Benito P, Villar-Salvador P andRey-Benayas JM2019 Long‐term recovery ofmultifunctionality inmediterranean forests depends onrestoration strategy and forest type J. Appl. Ecol. 56 745–57

DaCruzDC, Rey Benayas JM, Ferreira GC, Santos S R andSchwartzG 2020An overview of forest loss and restoration inthe BrazilianAmazonNewForests (https://doi.org/10.1007/s11056-020-09777-3)

Da Silva R F B, BatistellaM,Moran E F and LuD2017 Land changesfostering Atlantic Forest transition in Brazil: evidence fromthe ParaíbaValleyProf. Geogr. 69 80–93

Díaz S et al 2019 Summary for policymakers of the global assessmentreport of the Intergovernmental science-policy platformonbiodiversity and ecosystem services, Bonn, IPBES Secretariat,p 39 (https://ipbes.net/sites/default/files/inline/files/ipbes_global_assessment_report_summary_for_policymakers.pdf)

12

Environ. Res. Lett. 15 (2020) 043002 RLChazdon et al

https://doi.org/10.1079/PAVSNNR20072057https://doi.org/10.1079/PAVSNNR20072057https://doi.org/10.1079/PAVSNNR20072057https://doi.org/10.1579/0044-7447-32.6.389https://doi.org/10.1579/0044-7447-32.6.389https://doi.org/10.1579/0044-7447-32.6.389https://doi.org/10.1579/0044-7447-32.6.389https://doi.org/10.1017/S002191181800089Xhttps://doi.org/10.1017/S002191181800089Xhttps://doi.org/10.1017/S002191181800089Xhttps://doi.org/10.1017/S002191181800089Xhttps://doi.org/10.1007/s10531-019-01791-yhttps://doi.org/10.1007/s10531-019-01791-yhttps://doi.org/10.1007/s10531-019-01791-yhttps://doi.org/10.1016/j.cosust.2017.01.015https://doi.org/10.1016/j.cosust.2017.01.015https://doi.org/10.1016/j.cosust.2017.01.015https://doi.org/10.1016/j.cosust.2017.01.015https://doi.org/10.1016/j.landusepol.2012.04.002https://doi.org/10.1016/j.landusepol.2012.04.002https://doi.org/10.1016/j.landusepol.2012.04.002https://doi.org/10.1016/j.foreco.2018.11.033https://doi.org/10.1016/j.foreco.2018.11.033https://doi.org/10.1016/j.foreco.2018.11.033https://doi.org/10.1002/ldr.3534https://doi.org/10.1016/j.biocon.2009.08.005https://doi.org/10.1016/j.biocon.2009.08.005https://doi.org/10.1016/j.biocon.2009.08.005https://doi.org/10.1111/btp.12383https://doi.org/10.1111/btp.12383https://doi.org/10.1111/btp.12383https://doi.org/10.1126/sciadv.aav3223https://doi.org/10.1016/j.wdp.2016.11.014https://doi.org/10.1016/j.wdp.2016.11.014https://doi.org/10.1016/j.wdp.2016.11.014https://doi.org/10.1016/j.biocon.2013.05.018https://doi.org/10.1016/j.biocon.2013.05.018https://doi.org/10.1016/j.biocon.2013.05.018https://doi.org/10.1111/gcb.14451https://doi.org/10.1111/gcb.14451https://doi.org/10.1111/gcb.14451https://doi.org/10.1016/j.tree.2011.06.011https://doi.org/10.1016/j.tree.2011.06.011https://doi.org/10.1016/j.tree.2011.06.011https://doi.org/10.1016/j.tree.2011.06.011https://doi.org/10.1016/j.forpol.2017.06.002https://doi.org/10.1016/j.forpol.2017.06.002https://doi.org/10.1016/j.forpol.2017.06.002https://doi.org/10.1016/j.forpol.2017.06.002https://doi.org/10.5751/ES-10270-230407https://doi.org/10.1505/146554819827906825https://doi.org/10.1505/146554819827906825https://doi.org/10.1505/146554819827906825https://doi.org/10.1080/11263504.2017.1407374https://doi.org/10.1080/11263504.2017.1407374https://doi.org/10.1080/11263504.2017.1407374https://doi.org/10.1016/j.apgeog.2019.03.001https://doi.org/10.1016/j.apgeog.2019.03.001https://doi.org/10.1016/j.apgeog.2019.03.001https://doi.org/10.1126/science.aax9539https://doi.org/10.1126/science.aax9539https://doi.org/10.1126/science.aax9539https://doi.org/10.1111/btp.12678https://doi.org/10.1111/btp.12678https://doi.org/10.1111/btp.12678https://doi.org/10.1007/s13280-016-0772-yhttps://doi.org/10.1007/s13280-016-0772-yhttps://doi.org/10.1007/s13280-016-0772-yhttps://doi.org/10.1046/j.1523-1739.1999.98352.xhttps://doi.org/10.1046/j.1523-1739.1999.98352.xhttps://doi.org/10.1046/j.1523-1739.1999.98352.xhttps://doi.org/10.1111/btp.12381https://doi.org/10.1111/btp.12381https://doi.org/10.1111/btp.12381https://doi.org/10.1111/j.1523-1739.2009.01338.xhttps://doi.org/10.1111/j.1523-1739.2009.01338.xhttps://doi.org/10.1111/j.1523-1739.2009.01338.xhttps://doi.org/10.1111/j.1523-1739.2009.01338.xhttps://doi.org/10.1111/btp.12409https://doi.org/10.1111/btp.12409https://doi.org/10.1111/btp.12409https://doi.org/10.1126/sciadv.1501639https://doi.org/10.3390/f5112658https://doi.org/10.3390/f5112658https://doi.org/10.3390/f5112658https://doi.org/10.1016/j.tree.2007.10.005https://doi.org/10.1016/j.tree.2007.10.005https://doi.org/10.1016/j.tree.2007.10.005https://doi.org/10.1111/1365-2664.13501https://doi.org/10.1111/1365-2664.13501https://doi.org/10.1111/1365-2664.13501https://doi.org/10.1038/ncomms11666https://doi.org/10.1126/sciadv.1701345https://doi.org/10.1111/conl.12709https://doi.org/10.1111/1365-2664.13340https://doi.org/10.1111/1365-2664.13340https://doi.org/10.1111/1365-2664.13340https://doi.org/10.1007/s11056-020-09777-3https://doi.org/10.1007/s11056-020-09777-3https://doi.org/10.1080/00330124.2016.1178151https://doi.org/10.1080/00330124.2016.1178151https://doi.org/10.1080/00330124.2016.1178151https://ipbes.net/sites/default/files/inline/files/ipbes_global_assessment_report_summary_for_policymakers.pdfhttps://ipbes.net/sites/default/files/inline/files/ipbes_global_assessment_report_summary_for_policymakers.pdfhttps://ipbes.net/sites/default/files/inline/files/ipbes_global_assessment_report_summary_for_policymakers.pdf

DingH, Altamirano J C, AnchondoA, Faruqi S, VerdoneM,WuA,ZamoraR, ChazdonR andVergaraW2017Roots ofProsperity: The Economics and Finance of Restoring Land(Washington, DC:World Resources Institute)

DuangjaiW, Schmidt-Vogt D and Shrestha RP 2015 Farmers’ landuse decision-making in the context of changing land andconservation policies: a case study of doimae salong inchiang rai province, Northern Thailand LandUse Policy 48179–89

DuncanR S andChapmanCA1999 Seed dispersal and potentialforest succession in abandoned agriculture in tropical AfricaEcol. Appl. 9 998–1008

Dupuy JMandChazdonRL 2008 Interacting effects of canopy gap,understory vegetation and leaf litter on tree seedlingrecruitment and composition in tropical secondary forestsForest Ecol.Manage. 255 3716–25

Evans L J, Goossens B andAsnerGP 2017Underproductiveagriculture aids connectivity in tropical forests Forest Ecol.Manage. 401 159–65

Evaristo J andMcDonnell J J 2019Global analysis of streamflowresponse to forestmanagementNature 570 455–61

Fantini AC, SchuchC, Siminski A and Siddique I 2019 Small-scalemanagement of secondary forests in the brazilian atlanticforest Floresta e Ambiente 26

FAO&GlobalMechanismOf TheUNCCD2015 Sustainablefinancing for forest and landscape restoration: opportunities,challenges and theway forwardDiscussion Paper (Rome: FAOandUNCCD)

Ferguson JM2014The scramble for the waste lands: trackingcolonial legacies, counterinsurgency and internationalinvestment through the lens of land laws in Burma/MyanmarSingap. J. Tropical Geogr. 35 295–311

Filoso S, BezerraMO,Weiss KC and PalmerMA2017 Impacts offorest restoration onwater yield: a systematic review PLoSOne 12 e0183210

Filotas E et al 2014Viewing forests through the lens of complexsystems scienceEcosphere 5 art1

Fischer J, Stott J, Zerger A,WarrenG, SherrenK and Forrester R I2009Reversing a tree regeneration crisis in an endangeredecoregionProc. Natl Acad. Sci. 106 10386–91

Forero-Montaña J,Marcano-VegaH, Zimmerman JK andBrandeis T J 2019 Potential of second-growthNeotropicalforests for forestry: the example of Puerto Rico Forests, TreesLivelihoods 28 1–16

Franklin J F, LindenmayerDB,Macmahon J A,Mckee A,Magnuson J, PerryDA,WaideR and FosterDR 2000Threads of continuityConservation Biol. Pract. 1 8–16.

Garcia-Barrios L, Galvan-Miyoshi YM,Valdivieso-Perez I A,MaseraOR, BoccoG andVandermeer J 2009Neotropicalforest conservation, agricultural intensification, and ruralout-migration: theMexican experienceBioscience 59 863–73

Gardner C J, Bicknell J E, Baldwin-CantelloW, StreubigM J andDavies ZG 2019Quantifying the impacts of defaunation onnatural forest regeneration in a globalmeta-analysisNat.Commun. 10 4590

Geddes L S, Lunt ID, Smallbone LT andMorgan JW2011Old fieldcolonization by native trees and shrubs following land usechange: could this beVictoria’s largest example of landscaperecovery?Ecol.Manage. Restor. 12 31–6

Ghazoul J, Burivalova Z, Garcia-Ulloa J andKing LA 2015Conceptualizing forest degradationTrends Ecol. Evol. 30622–32

Ghazoul J andChazdonRL 2017Degradation and recovery inchanging forest landscapes: amultiscale conceptualframeworkAnnu. Rev. Environ. Resour. 42 161–88

GoldsmithGR, Comita L S andChua SC2011 Evidence for arrestedsuccessionwithin a tropical forest fragment in SingaporeJ. Tropical Ecol. 27 323–6

GriscomBW,Adams J, Ellis PW,HoughtonRA, LomaxG,MitevaDA, SchlesingerWH, ShochD, Siikamäki J V andSmith P 2017Natural climate solutions Proc. Natl Acad. Sci.114 11645–50

GuariguataMR1999 Early response of selected tree species toliberation thinning in a young secondary forest inNortheasternCosta Rica Forest Ecol.Manage. 124 255–61

Guevara S, Laborde J and Sanchez-RiosG 2005The trees the forestleft behind Interciencia 30 595

Hecht S, YangAL, Basnett B S, PadochC andPelusoNL 2015People inMotion, Forests in Transition: Trends inMigration,Urbanization, andRemittances and their Effects on TropicalForests (Bogor, Indonesia: CIFOR) (https://doi.org/10.17528/cifor/005762)

Hecht S B,MorrisonKD and PadochC 2014The Social Lives ofForests: Past, Present, and Future ofWoodland Resurgence(Chicago, IL: University of Chicago Press)

Hernández-X E, Levy-Tacher S I and Bello-Baltazar Y E 1995 Laroza-tumba-quema enYucatán ed EHernández-X et al Lamilpa en Yucatán: un sistema de producción agrícolatradicional (México: Colegio de)

HissaHR,Alves FilhoNT,CostaM, StrauchG, Bassi L andAssis, R. L. D E 2019 Sustainable rural development in Rio deJaneiro state: the Rio rural program Strategies and Tools for aSustainable Rural Rio de Janeiro (Berlin: Springer)

Holl KD2017Restoring tropical forests from the bottomup Science355 455–6

Holl KD andAide TM2011When andwhere to actively restoreecosystems? Forest Ecol.Manage. 261 1558–63

Holl KD, Reid J L, Oviedo‐Brenes F, Kulikowski A J andZahawi RA2018Rules of thumb for predicting tropical forest recoveryAppl. Veg. Sci. 21 669–77

HoosbeekMR, RemmeRP andRuschGM2016Trees enhance soilcarbon sequestration and nutrient cycling in a silvopastoralsystem in south-westernNicaraguaAgroforestry Syst. 92 1–11

HoughtonR, Byers B andNassikas AA 2015A role for tropicalforests in stabilizing atmospheric CO2Nat. Clim. Change 51022–3

HoughtonR andNassikas AA2017Global and regional fluxes ofcarbon from land use and land cover change 1850–2015Glob.Biogeochem. Cycles 31 456–72

Hua F,Wang L, Fisher B, ZhengX,WangX,DouglasWY, TangY,Zhu J andWilcoveD S 2018Tree plantations displacingnative forests: the nature and drivers of apparent forestrecovery on former croplands in SouthwesternChina from2000 to 2015Biol. Conservation 222 113–24

Isbell F, TilmanD, Reich PB andClarkAT 2019Deficits ofbiodiversity and productivity linger a century afteragricultural abandonmentNat. Ecol. Evol. 3 1533–8

JactelH, Bauhus J, Boberg J, BonalD, Castagneyrol B, Gardiner B,Gonzalez-Olabarria J R, Koricheva J,MeurisseN andBrockerhoff EG 2017Tree diversity drives forest standresistance to natural disturbancesCurr. Forestry Rep. 3223–43

Jadin I,Meyfroidt P and Lambin E 2016 International trade, andland use intensification and spatial reorganization explainCosta Rica’s forest transition Environ. Res. Lett. 11 035005

JakovacCC, Bongers F, Kuyper TW,Mesquita RC andPeña‐ClarosM2016 Land use as a filter for speciescomposition inAmazonian secondary forests J. VegetationSci. 27 1104–16

JakovacCC, Peña‐ClarosM,Kuyper TWandBongers F 2015 Lossof secondary‐forest resilience by land‐use intensification intheAmazon J. Ecol. 103 67–77

Johnstone J F, AllenCD, Franklin J F, Frelich L E,Harvey B J,Higuera P E,MackMC,Meentemeyer RK,MetzMR andPerryG L 2016Changing disturbance regimes, ecologicalmemory, and forest resilience Frontiers Ecol. Environ. 14369–78

JonesHP, Jones PC, Barbier E B, BlackburnRC, Benayas JMR,Holl KD,MccrackinM,Meli P,MontoyaD andMateosDM2018Restoration and repair of Earth’s damaged ecosystemsProc. R. Soc.B 285 20172577

Jones I L, Dewalt S J, LopezOR, Bunnefeld L, Pattison Z andDentDH2019Above-and belowground carbon stocks aredecoupled in secondary tropical forests and are positively

13

Environ. Res. Lett. 15 (2020) 043002 RLChazdon et al

https://doi.org/10.1016/j.landusepol.2015.06.002https://doi.org/10.1016/j.landusepol.2015.06.002https://doi.org/10.1016/j.landusepol.2015.06.002https://doi.org/10.1016/j.landusepol.2015.06.002https://doi.org/10.1890/1051-0761(1999)009[0998:SDAPFS]2.0.CO;2https://doi.org/10.1890/1051-0761(1999)009[0998:SDAPFS]2.0.CO;2https://doi.org/10.1890/1051-0761(1999)009[0998:SDAPFS]2.0.CO;2https://doi.org/10.1016/j.foreco.2008.03.021https://doi.org/10.1016/j.foreco.2008.03.021https://doi.org/10.1016/j.foreco.2008.03.021https://doi.org/10.1016/j.foreco.2017.07.015https://doi.org/10.1016/j.foreco.2017.07.015https://doi.org/10.1016/j.foreco.2017.07.015https://doi.org/10.1038/s41586-019-1306-0https://doi.org/10.1038/s41586-019-1306-0https://doi.org/10.1038/s41586-019-1306-0https://doi.org/10.1590/2179-8087.069017https://doi.org/10.1111/sjtg.12078https://doi.org/10.1111/sjtg.12078https://doi.org/10.1111/sjtg.12078https://doi.org/10.1371/journal.pone.0183210https://doi.org/10.1890/ES13-00182.1https://doi.org/10.1080/14728028.2019.1579673https://doi.org/10.1080/14728028.2019.1579673https://doi.org/10.1080/14728028.2019.1579673https://doi.org/10.1111/j.1526-4629.2000.tb00155.xhttps://doi.org/10.1111/j.1526-4629.2000.tb00155.xhttps://doi.org/10.1111/j.1526-4629.2000.tb00155.xhttps://doi.org/10.1525/bio.2009.59.10.8https://doi.org/10.1525/bio.2009.59.10.8https://doi.org/10.1525/bio.2009.59.10.8https://doi.org/10.1038/s41467-019-12539-1https://doi.org/10.1111/j.1442-8903.2011.00570.xhttps://doi.org/10.1111/j.1442-8903.2011.00570.xhttps://doi.org/10.1111/j.1442-8903.2011.00570.xhttps://doi.org/10.1016/j.tree.2015.08.001https://doi.org/10.1016/j.tree.2015.08.001https://doi.org/10.1016/j.tree.2015.08.001https://doi.org/10.1016/j.tree.2015.08.001https://doi.org/10.1146/annurev-environ-102016-060736https://doi.org/10.1146/annurev-environ-102016-060736https://doi.org/10.1146/annurev-environ-102016-060736https://doi.org/10.1017/S0266467411000010https://doi.org/10.1017/S0266467411000010https://doi.org/10.1017/S0266467411000010https://doi.org/10.1073/pnas.1710465114https://doi.org/10.1073/pnas.1710465114https://doi.org/10.1073/pnas.1710465114https://doi.org/10.1016/S0378-1127(99)00072-9https://doi.org/10.1016/S0378-1127(99)00072-9https://doi.org/10.1016/S0378-1127(99)00072-9https://doi.org/10.17528/cifor/005762https://doi.org/10.17528/cifor/005762https://doi.org/10.1126/science.aam5432https://doi.org/10.1126/science.aam5432https://doi.org/10.1126/science.aam5432https://doi.org/10.1016/j.foreco.2010.07.004https://doi.org/10.1016/j.foreco.2010.07.004https://doi.org/10.1016/j.foreco.2010.07.004https://doi.org/10.1111/avsc.12394https://doi.org/10.1111/avsc.12394https://doi.org/10.1111/avsc.12394https://doi.org/10.1007/s10457-016-0049-2https://doi.org/10.1007/s10457-016-0049-2https://doi.org/10.1007/s10457-016-0049-2https://doi.org/10.1038/nclimate2869https://doi.org/10.1038/nclimate2869https://doi.org/10.1038/nclimate2869https://doi.org/10.1038/nclimate2869https://doi.org/10.1002/2016GB005546https://doi.org/10.1002/2016GB005546https://doi.org/10.1002/2016GB005546https://doi.org/10.1016/j.biocon.2018.03.034https://doi.org/10.1016/j.biocon.2018.03.034https://doi.org/10.1016/j.biocon.2018.03.034https://doi.org/10.1038/s41559-019-1012-1https://doi.org/10.1038/s41559-019-1012-1https://doi.org/10.1038/s41559-019-1012-1https://doi.org/10.1007/s40725-017-0064-1https://doi.org/10.1007/s40725-017-0064-1https://doi.org/10.1007/s40725-017-0064-1https://doi.org/10.1007/s40725-017-0064-1https://doi.org/10.1088/1748-9326/11/3/035005https://doi.org/10.1111/jvs.12457https://doi.org/10.1111/jvs.12457https://doi.org/10.1111/jvs.12457https://doi.org/10.1111/1365-2745.12298https://doi.org/10.1111/1365-2745.12298https://doi.org/10.1111/1365-2745.12298https://doi.org/10.1002/fee.1311https://doi.org/10.1002/fee.1311https://doi.org/10.1002/fee.1311https://doi.org/10.1002/fee.1311https://doi.org/10.1098/rspb.2017.2577

related to forest age and soil nutrients respectively Sci. TotalEnviron. 697 133987

Kammesheidt L 2002Perspectives on secondary forestmanagementin tropical humid lowlandAmericaAmbio 31 243–50

Kowler L F, Ravikumar A, LarsonAM,Rodriguez-WardD,BurgaC andGonzales Tovar J 2016Analyzingmultilevelgovernance in PeruWorking Paper 203 (Bogor, Indonesia:CIFOR) (http://www.cifor.org/publications/pdf_files/WPapers/WP203Kowler.pdf)

KremenC andMerenlender AM2018 Landscapes that work forbiodiversity and people Science 362 eaaau6020

LacombeG, Ribolzi O,DeRouwA, Pierret A, LatsachakK,SilveraN, PhamDinhR,OrangeD, Janeau J and Soulileuth B2015Afforestation by natural regeneration or by treeplanting: examples of opposite hydrological impactsevidenced by long-term fieldmonitoring in the humidtropicsHydrol. Earth Syst. Sci. Discuss. 12 12615–48

Lambin E F andMeyfroidt P 2010 Land use transitions: Socio-ecological feedback versus socio-economic change LandUsePolicy 27 108–18

Lasanta T, Arnáez J, PascualN, Ruiz-Flaño P, ErreaMP andLana-RenaultN 2017 Space-time process and drivers of landabandonment in EuropeCatena 149 810–23

Latawiec A E,Dos Santos J S,Maioli V, Junqueira AB, Crouzeilles R,JakovacCC, Tubenchlak F and Strassburg BN2018 Forestlandscape restoration and land sparing–sharing: shifting thefocus towards nature’s contributions to people edSMansourian and J Parrotta Forest Landscape Restoration(London: Routledge) pp 116–34

Lembani R L, Knight J andAdamE2019Use of Landsatmulti-temporal imagery to assess secondary growthMiombowoodlands in Luanshya, Zambia Sothern Forests 81 129–40

LennoxGD et al 2018 Second rate or a second chance?Assessingbiomass and biodiversity recovery in regeneratingAmazonian forestsGlob. Change Biol. 24 5680–94

Levy-Tacher S, Ramírez-Marcial N, González-EspinosaMandRomán-Dañobeytia F 2012Rehabilitación ecológica de áreasagropecuarias degradadas en la selva lacandona: unaalternativafincada en el conocimiento ecológico tradicionalmaya La otra innovación para el ambiente y la sociedad en lafrontera sur deMéxico ed E Bello Baltazar et al (SanCristóbalde Las Casas, Chiapas: ECOSUR) pp 248–58

Levy-Tacher S I, Ramírez-Marcial N,Navarrete-Gutiérrez DA andRodríguez-Sánchez PV 2019AreMayan community forestreserves effective in fulfilling people’s needs and preservingtree species? J. Environ.Manage. 245 16–27

Li S and Li X 2017Global understanding of farmland abandonment:a review and prospects J. Geog. Sci. 27 1123–50

LiaoC andBrownDG2018Assessments of synergistic outcomesfrom sustainable intensification of agriculture need to includesmallho