The forest transformation: Planted tree cover and regional dynamics of tree gains and losses

https://doi.org/10.1016/j.gloenvcha.2019.101988Get rights and content

Highlights

  • We explored tree-cover change in tropical planted areas since 1990 for 11 countries.

  • Tree-cover gain and loss weakly distinguished planted areas from mosaic landscapes.

  • Losses were prominent in Southeast Asia (20% of planted areas).

  • Gains were more notable in South America (>14% of planted areas).

  • The forest-transformation concept is introduced to account for regional differences.

  • It complements the forest-transition concept and stresses forest political economy.

Abstract

Extensions of forest-transition theory to the tropics often depict sustained expansions of planted tree cover and corresponding long-term net gains in total tree cover. To explore the patterns and implications of continued tropical planted tree-cover expansion, we profiled sequences of tree-cover change over 1990–2010 according to Landsat imagery for recently observed (ca. 2014) planted tree-cover areas in 11 tropical countries. Alternative patterns of change emerged from these analyses. Termed the ‘reforestation treadmill’ and ‘forest transformation’ narratives, planted tree-cover change featured relatively ephemeral planted covers, modest net gains, and similar tree-cover change dynamics compared to nearby agricultural-forest mosaics. Planted areas were characterised not by unambiguous reforestation but rather combinations of tree-cover losses and gains, with losses typically being more prominent. Contemporary gains and losses during 5–10-year periods regularly distinguished planted areas from non-planted areas, with losses being 1.5–2.3 times more common than gains. Planted areas were only moderately distinguishable from non-planted areas overall with respect to tree-cover change dynamics. Relationships between tree-cover change and the export orientations of planted tree/tree-crop commodities were also examined. Greater export orientations did not significantly associate with tree-cover loss or larger planted patches, with partial exceptions for Southeast Asia. Regional disparities in planted tree-cover dynamics were apparent. In Southeast Asia, dominated by Indonesia, tree-cover declines in planted areas since 1990 were relatively pronounced (20% of planted areas), particularly with respect to progressive transitions from tree cover to cleared lands. Planted areas there were generally indistinguishable from nearby non-planted areas with respect to historical tree-cover change dynamics. In contrast, in South America, dominated by Brazil, tree-cover increases in planted areas since 1990 were more appreciable (at least 14% of planted areas), with most being progressive, stable, ‘net’ increases (10% of planted areas) and the remainder being dynamic increases entailing short-term losses since 1990 (4% of planted areas). Total tree-cover increases within South American planted areas were equal to or greater than total decreases since 1990. These patterns suggest a forest-transformation narrative in which major planted-area expansion occurs alongside minor net tree-cover change. This narrative appears particularly well suited to Southeast Asia, where planted areas are extensive and expansive but where net tree cover gains are tenuous, reflecting political-economic shifts in forest management and the devaluation of extensive, degraded natural forests.

Introduction

Planted tree cover is increasingly prominent in the tropics as natural forest cover continues to decline. According to FAO estimates, tropical planted tree cover expanded 87% since 1990 (Keenan et al., 2015) and accounts for 3.2% of total tropical forest area (Payn et al., 2015). In Southeast Asia, planted forests are reportedly more extensive than naturally regenerating forests (FAO, 2015). Such figures are doubtless under-estimates due in part to their exclusion of various agro-forestry and small-holder forestry systems (Schnell et al., 2015; Midgley et al., 2017). Planted tree-cover expansion is expected to continue due to looming demand shortfalls for forestry and agro-forestry products (Carlson et al., 2012; Barua and Lehtonen, 2013; Payn et al., 2015) but also its congruence with various large-scale reforestation schemes (Sloan, 2016; Lewis et al., 2019; Rudel et al., 2019).

Recent expansion of planted cover evokes imagery of new, enduring tree cover. Whether this is the case generally is unknown. To some degree, this outcome depends on the extent to which planted tree-cover expansion has aligned with agriculture, which is widely responsible for tropical deforestation (Gibbs et al., 2010). Planted-cover expansion may, for instance, entail land-use systems or economic processes of landscape change common to agriculture (Sloan, 2016), particularly as many tropical planted areas are dominated by agro-forestry (Petersen et al., 2016; Hurni et al., 2017) that has increasingly adopted intensive monocultural practices (Clough et al., 2016). This conceptualization of planted tree cover recalls Grainger's (1995) land-use transition, a precursor of the forest-transition, in which tree-cover changes in planted areas increasingly depart from the deforesting tendencies of associated agriculture to yield enduring gains in forest cover. A view of tropical planted tree cover as associated with agriculture expansion has emerged in several recent studies (Altamirano et al., 2013; Van Holt et al., 2016) and if generally true would qualify planted-cover expansion as a marginal contributor to stable net reforestation generally. This possibility has nonetheless remained under-appreciated in the forest-transition literature, which has focused on aggregate forest gains and paid relatively little attention to agro-forestry (Rudel et al., 2005; Sloan and Sayer, 2015; Sloan et al., 2019a). Regional transitions from natural to planted tropical tree cover over the last three decades remain unprecedented, but under-studied, aspects of forest change and apparent expansion.

Narratives of tropical forest gain, steeped in forest-transition theory (Rudel, 2009), may fail to reveal critical, dynamic aspects of tropical planted tree-cover expansion. The forest-transition narrative of enduring planted tree cover and its contribution to net forest gains was recently qualified by the reforestation ‘treadmill’ narrative (Rudel et al., 2016). Contrary to notions of monotonic gains of stable planted cover over long-cleared lands in response to forest scarcity, the treadmill narrative characterises planted covers as ‘commodity crops’ by which tree-cover gains are relatively dynamic, partially aligned with agricultural activities, and often entail forest conversion. According to this narrative, tropical tree-cover gains are increasingly recurrent but make limited contributions to stable net gains because planted tree covers, while expansive, are ephemeral and progressively replace natural forests (Sloan and Sayer, 2015). Buoyed by political-economic conditions and tropical climates promoting rapid tree reproduction for globalised markets, industrial-scale interests have been advanced as the primary agents of the treadmill, in keeping with trends in enterprise-driven deforestation (Rudel, 2007; Austin et al., 2017a; Curtis et al., 2018) and corporate ‘green grabs’ (Hall, 2011; Scheidel and Work, 2018; Sloan et al., 2019b). Over time, the treadmill of forest loss and serial tree reproduction would result in forest transformations, characterised by extensive new planted tree cover but modest, unstable net tree cover gains within planted areas alongside appreciable forest losses (Hansen et al., 2013; Tropek et al., 2014; Miranda et al., 2015; Rudel et al., 2016). The potential for such transformations is underscored by an increasing integration of commercial agro-forestry and forestry with multi-lateral reforestation programmes targeting tens of millions of hectares for climate-change mitigation and tropical forest restoration (Laestadius et al., 2015; Sloan, 2016; Lewis et al., 2019; Rudel et al., 2019).

The treadmill and forest-transformation narratives depart from the classical forest transition but the evidence for these narratives remains inconclusive. Most large-scale studies of tropical forest gain have conflated planted and natural forest covers (Hansen et al., 2013; Stibig et al., 2014; Zomer et al., 2014) or they employed crude measures of planted areas (Aide et al., 2013; Sloan and Sayer, 2015; Wilson et al., 2017). Provided satellite data on planted tree cover, most such studies are confined to direct transitions from natural to planted tree cover for a single brief period, specific commodity, and particular region (Gutiérrez-Vélez et al., 2011; Li and Fox, 2012; Gaveau et al., 2016b; Petersen et al., 2016; Van Holt et al., 2016; Austin et al., 2017b; Furumo and Aide, 2017; Hurni et al., 2017). Collectively, such studies tenuously indicate net tree-cover gains within planted areas. The variability and place-specific nature does, however, make it hazardous to generalize their findings to the globe or across regions. Further, their disregard for longer-term dynamics limits conceptual insights regarding how planted-cover expansion proceeds amidst broader, dynamic processes of forest change.

A generalised understanding of the role of planted tree cover requires long-term, pan-tropical observations of the pathways to planted tree cover. To this end, with a view to elaborate the treadmill and forest-transformation narratives, we explore three key aspects of tree-cover change over 1990–2010 within recent (2014) planted areas pan-tropically:

  • Land-change dynamism: To what degree are planted areas characterised by simple, monotonic versus dynamic sequences of tree-cover change?

  • Agricultural alignment: How similar are planted areas to surrounding agricultural-forest mosaics in terms of their tree-cover change dynamics?

  • Commercial nature: To what degree do export-oriented planted areas associate with unique tree-cover change dynamics, particularly those characterised by losses?

Regarding land-change dynamism, a majority of simple, monotonic sequences of increasing tree cover in the absence of losses would indicate stable planted covers and positive contributions towards net reforestation, as per forest-transition narratives. Conversely, dynamic tree-cover changes since 1990, marked over time by alternating tree-cover gains and losses, would indicate ephemeral planted tree covers and suggest possible forest conversion. Such dynamism would accord with treadmill and forest-transformation narratives and undermine the environmental goals of multi-lateral reforestation programmes, which depend on persistent new tree cover over cleared lands (de Jong, 2010; Körner, 2017; Lewis et al., 2019). A disregard of the dynamism of tree-cover change within planted areas (Gibbs, 2012; Meyfroidt et al., 2014; Sloan and Sayer, 2015) would conflate ephemeral gains with enduring gains, as well as temporary losses with longer-term deforestation. Such disregard would also discount ‘indirect deforestation’ driven by speculation over later plantation reforestation (World Bank 2006, Hoang et al., 2010). In this light, apparent net tree-cover increases due to planted-cover establishment are doubtless inflated (e.g., Altamirano et al., 2013; Gaveau et al., 2016b) and apparent forest-transition trajectories may be otherwise (Zhai et al., 2017). The treadmill narrative anticipates that dynamic sequences of tree-cover increase within planted areas predominate over simple, monotonic increases but only marginally exceed dynamic decreases (Hypothesis 1).

The agricultural alignment of planted areas, defined as the similarity of their tree-cover change dynamics compared to nearby agricultural-forest mosaic areas, bears on whether a forest transformation or forest transition is more likely. A high similarity between planted areas and surrounding agricultural-forest mosaics would indicate an ‘alignment’ of their land-change dynamics, if not their land-use systems, rendering planted areas less likely to make enduring contributions towards net reforestation generally. Similarities are likely where planted areas encompass or associate with agricultural activities (Williams, 1986; Farley, 2007; Cramb and Curry, 2012; Potter, 2012; Gaveau et al., 2016a; Sloan, 2016; Sloan et al., 2017); are highly industrialised, e.g., entailing short rotations or limited canopy cover; or where agricultural areas host recurrent forest regrowth, as due to long fallows, land abandonment, or forest degradation (Sloan, 2016; Reid et al., 2018). The treadmill narrative anticipates that planted areas exhibit similar tree-cover change dynamics compared to agricultural-forest mosaics, albeit with more prevalent tree-cover gains (Hypothesis 2).

Regarding the commercial nature of tree-cover change, the treadmill narrative expects that planted areas with greater export orientations are more likely to exhibit unique tree-cover change dynamics, centered on tree-cover losses, particularly amongst larger-scale planted patches (Hypothesis 3). Such a possibility has been implied by case studies (Gutiérrez-Vélez et al., 2011; Meyfroidt et al., 2014; Sloan, 2016; Scheidel and Work, 2018; Sloan et al., 2019b) and draws upon associations between deforestation and the prices of planted tree commodities in global markets (Williams, 1986, Gaveau et al., 2019). Conversion of degraded forests into planted tree cover endows commercial plantation enterprises with subsidies or efficiencies often crucial to their viability (Barr, 2001a; Gutiérrez-Vélez et al., 2011). Commercial enterprises in turn are relatively responsive to reforestation incentives (Sloan, 2016; Rudel et al., 2019) and able to negotiate access to State production forests (Resosudarmo et al., 2019). The pan-tropical prevalence of industrial-scale planted areas (Petersen et al., 2016) and, increasingly, of larger tropical forest clearings (Austin et al., 2017a), accords with the treadmill and forest-transformation narratives.

Focussing on these key aspects of tree-cover change in planted areas, we introduce the forest-transformation narrative as a complement to the forest transition, with emphasis on Southeast Asia. This narrative generalises land-change and political-economic processes by which significant planted-cover expansion has yielded only limited net tree-cover gains in planted areas and, by extension, limited contributions to reforestation generally. The narrative elaborates emergent perspectives of tropical reforestation as increasingly centrally-organised, accelerated, and centered on planted tree covers (Lewis et al., 2019; Rudel et al., 2019). It also reframes earlier notional ‘Asian forest transitions’ (Mather, 2007) and ‘strong State’ reforestation pathways (Lambin and Meyfroidt, 2010), as by underscoring the political-economic basis for limited long-term contributions of planted areas to general reforestation. The narrative likewise supports remote-sensing observations of elevated deforestation in Southeast Asian planted areas (Petersen et al., 2016).

While we discuss the juxtaposition of the forest transformation and forest transition, we do not test the forest-transition model per se. Its particular contexts, drivers, and timeframes are beyond our scope and have been explored elsewhere (Rudel, 2005; Sloan, 2015; Van Holt et al., 2016; Sloan et al., 2019a). Nor do we to seek to conclusively quantify the proportion of recent planted tree cover that replaced natural forests pan-tropically. Observations of strictly natural historical forest cover are beyond the scope of available pan-tropical satellite data (Sloan, 2012; Feng et al., 2016; Curtis et al., 2018) (Text S5). We focus therefore on the aforementioned aspects of historical tree-cover change within recent planted areas, with attention to indicative dynamics, to refine emergent narratives of tropical planted-cover expansion holding implications for broader forest change.

Section snippets

Land-change dynamics in planted areas

To explore the land-change dynamics in planted areas (Hypothesis 1, Section 3.1), sequences of tree-cover change over 1990–2010 according to successive Landsat classifications were summarised for 46.7 million hectares of planted areas across 11 tropical countries (Brazil, Cambodia, Colombia, Indonesia, Laos, Liberia, Malaysia, Panama, Peru, Thailand, Vietnam) (Table 1). Planted areas were delineated as of ca. 2014 – the sole year for which their pan-tropical extent is explicitly observed – such

Simple and dynamic tree-cover change for planted areas, 1990–2010

Tree-cover changes occurred over one-third of recent (ca. 2014) tropical planted areas during 1990–2010 (Fig. 1). Overall, sequences of tree-cover decrease were more prevalent than sequences of tree-cover increase within planted areas, with simple decreases being particularly prevalent. Over 1990–2010, simple and dynamic decreases in tree cover characterised 15% and 2.7% of planted areas, respectively, whereas simple and dynamic increases in tree cover characterised 7% and 3.6%, respectively (

A treadmill of planted-cover expansion

This study profiled tree-cover changes during 1990–2010 within recently-observed areas of planted tree cover. Our findings are generally consistent with the emergent treadmill and forest-transformation narratives of forest turnover and loss associated with planted tree-cover expansion, although they do not preclude forest-transition narratives of monotonic, enduring tree-cover gains in specific contexts.

In support of these emergent narratives, tropical planted areas experienced dynamic

Conclusion

Planted tree cover is expanding across the tropics but it does not generally appear to be associated with progressive, enduring increases in tree cover. Exceptions apply to eastern Brazil and Colombia, where tree-cover increases in planted areas were relatively prominent. Expectations of unambiguous reforestation within planted areas were qualified by characteristic combinations of contemporary losses and gains of tree cover, with losses being more prominent, as well as by an appreciable

Declaration of Competing Interest

The authors declare no conflict of interest.

Acknowledgements

We thank Dr Wei-Yin Loh (University of Wisconsin-Madison) for advising on the use of GUIDE analysis. We also thank Zhe Li (NOAA) for sharing MODIS classification for Southeast Asia, Kaspar Hurni (University of Bern) and Jefferson Fox (East-West Center) for securing these data, and Dr Do-Hyung Kim for arranging access to his Landsat GLS classification. This research was funded by the PARTNERS Research Coordination Network via National Science Foundation Grant DEB 1313788.

References (134)

  • Z. Li et al.

    Mapping rubber tree growth in mainland Southeast Asia using time-series MODIS 250 m NDVI and statistical data

    Appl. Geogr.

    (2012)
  • A.S. Mather et al.

    The course and drivers of the forest transtion: the case of France

    J. Rural Stud.

    (1999)
  • P. Meyfroidt et al.

    Middle-range theories of land system change

    Global Environ. Change

    (2018)
  • A. Miranda et al.

    Different times, same story: native forest loss and landscape homogenization in three physiographical areas of south-central of Chile

    Appl. Geogr.

    (2015)
  • T. Payn et al.

    Changes in planted forests and future global implications

    For. Ecol. Manage.

    (2015)
  • R. Pirard et al.

    Do timber plantations contribute to forest conservation?

    Environ. Sci. & Policy

    (2016)
  • T.K. Rudel

    Changing agents of deforestation: from state-initiated to enterprise driven processes, 1970-1990

    Land Use Policy

    (2007)
  • T.K. Rudel

    Tree farms: Driving forces and regional patterns in the global expansion of forest plantations

    Land Use Policy

    (2009)
  • S.A. Abood et al.

    Relative contributions of the logging, fibre, oil palm and mining industries to forest loss in Indonesia

    Conserv. Lett.

    (2015)
  • T.M. Aide et al.

    Deforestation and reforestation of Latin America and the Caribbean (2001–2010)

    Biotropica

    (2013)
  • K.G. Austin et al.

    Trends in size of tropical deforestation events signal increasing dominance of industrial-scale drivers

    Environ. Res. Lett.

    (2017)
  • S.R. Baptista et al.

    A re-emerging Atlantic forest? Urbanization, industralization and the forest-transition in Santa Catarina, southern Brazil

    Environ. Conserv.

    (2006)
  • C. Barr

    Discipline and Accumulate: State Practice and Elite Consolidation in Indonesia's Timber Sector, 1967-1998

    (1999)
  • C. Barr

    Banking on sustainability: structural adjustment and forestry reform in post-Suharto Indonesia

    (2001)
  • C. Barr

    The development of Indonesia's forest-based industries under Suharto's New Order Regime, 1966-1998, Banking on Sustainability: Structural Adjustment and Forestry Reform in Post-Suharto Indonesia CIFOR and WWF, Washington, DC.

    (2001)
  • C. Barr

    Will HPH reform lead to sustainable forest management? Questioning the assumptions of the “sustainable logging” paradigm, Banking on Sustainability: Structural Adjustment and Forestry Reform in Post-Suharto Indonesia CIFOR and WWF, Washington, DC

    (2001)
  • C. Barr

    Profits on paper: The political-economy of fiber and finance in Indonesia's pulp and paper industries, Banking on Sustainability: Structural Adjustment and Forestry Reform in Post-Suharto Indonesia CIFOR and WWF, Washington, DC

    (2001)
  • Barr, C., Dermawan, A., Purnomo, H., Komarudin, H. (2010) Financial governance and Indonesia's reforestation fund...
  • S.K. Barua et al.

    The great plantation expansion

    Int. Tropic. Timb. Org. Forest Update

    (2013)
  • J. Beaubien

    Visual interpretation of vegetation through digitally enhanced LANDSAT-MSS images

    Remote Sens. Rev.

    (1986)
  • L. Breiman et al.

    Classification and Regression Trees

    (1984)
  • R. Burgess et al.

    The political economy of deforestation in the tropics

  • K.M. Carlson et al.

    Committed carbon emissions, deforestation, and community land conversion from oil palm plantation expansion in West Kalimantan, Indonesia

    Proc. Natl. Acad. Sci.

    (2012)
  • R. Chazdon

    Landscape restoration, natural regeneration, and the forests of the future

    Ann. Missouri Botanic. Garden

    (2017)
  • R., L. Chazdon et al.

    Natural regeneration in the context of large-scale forest and landscape restoration in the tropics

    Biotropica

    (2016)
  • R.L. Chazdon et al.

    A policy-driven knowledge agenda for global forest and landscape restoration

    Conserv. Lett.

    (2017)
  • R.L. Chazdon et al.

    When is a forest a forest? Forest concepts and definitions in the era of forest and landscape restoration

    Ambio

    (2016)
  • Y. Clough et al.

    Land-use choices follow profitability at the expense of ecological functions in Indonesian smallholder landscapes

    Nat. Commun.

    (2016)
  • R. Cramb et al.

    Oil palm and rural livelihoods in the Asia–Pacific region: an overview

    Asia Pac. Viewp.

    (2012)
  • L.M. Curran et al.

    Lowland forest loss in protected areas of Indonesian Borneo

    Science

    (2004)
  • P.G. Curtis et al.

    Classifying drivers of global forest loss

    Science

    (2018)
  • P.L.B. da Rocha et al.

    What is the value of eucalyptus monocultures for the biodiversity of the Atlantic forest? A multitaxa study in southern Bahia, Brazil

    J. Forest. Res.

    (2013)
  • L.Y. da Silva Mariano Pereira et al.

    Woody species regeneration in Atlantic Forest restoration sites depends on surrounding landscape

    Natureza & Conservação

    (2013)
  • W. de Jong

    Forest rehabilitation and its implications for forest transition theory

    Biotropica

    (2010)
  • M.R. Dove

    Smallholder rubber and swidden agriculture in Borneo: a sustainable adaptation to the ecology and economy of the tropical forest

    Econ. Bot.

    (1993)
  • Special Audit of the Reforestation Fund: Final Report

    (1997)
  • Global Forest Resources Assessment, 2015: Desk Reference

    (2015)
  • Crops and Livestock Trade

    (2018)
  • Crops and Livestock Production

    (2018)
  • Forestry Production and Trade

    (2018)
  • Cited by (32)

    • Landscape changes over 30 years of intense economic activity in the upper Paraná River basin

      2022, Ecological Informatics
      Citation Excerpt :

      On the other hand, there is a gain in Forest areas in most sub-basins, especially Claro (6), Tijuco (8), and Upper Grande (15), all showing an increase above 10% of the sub-basin area. As reported in other studies, forest increases in the UPRB are mainly related to the expansion of commercial forest plantations and a lesser, but essential, contribution from the rise in secondary forests (Beuchle et al., 2015; Ferreira et al., 2015; Silva et al., 2018; Sloan et al., 2019). In the first case, increases in the Planted Forest class were mainly observed for sub-basin 35, as already mentioned, but also for sub-basins Sucuruí (16), Verde (20), and Upper Paranapanema, with gains of 1366 km2, 1789 km2, and 3606 km2, respectively.

    • Forest loss in Indonesian New Guinea (2001–2019): Trends, drivers and outlook

      2021, Biological Conservation
      Citation Excerpt :

      We delineated the boundaries of plantations at a scale of 1:50,000 using a visual interpretation method. This method has been used effectively previously (Gaveau et al., 2019; Sloan et al., 2019b). We performed this task in the GIS software ArcMap (10.4.1).

    View all citing articles on Scopus
    View full text