Restoration Planning of Degraded Tropical Forests for Biodiversity and Ecosystem Services

Restoration planning of degraded tropical forests for biodiversity and ecosystem services

Sugeng Budiharta

Bachelor of Forestry

Master of Science in Conservation Biology

A thesis submitted for the degree of Doctor of Philosophy at

The University of Queensland in January 2016

School of Biological Sciences

Abstract

Forest restoration has the potential to mitigate the impact of deforestation and forest degradation. Various global policies have been sought to put restoration into the mainstream agenda including under the Convention on Biological Diversity (CBD) and the program for Reducing Emissions from Deforestation and forest Degradation (REDD+). The Aichi Target of the CBD set a target for at least 15% of degraded ecosystems to be restored by 2020 for key goals including biodiversity conservation, carbon enhancement and the provision of livelihoods. A theoretical framework to underpin decision- making for landscape-scale restoration has been slow to emerge, resulting in a limited contribution from science towards achieving such policy targets. My thesis develops decision frameworks to guide the restoration of degraded tropical forests to enhance biodiversity and the delivery of ecosystem services. In this thesis, three critical questions on how to make better decisions for landscape-scale restoration are addressed by: (a) considering landscape heterogeneity in terms of degradation condition, restoration action and cost, and temporally-explicit restoration benefits; (b) leveraging restoration within competing land uses using emerging policy for offsetting; and (c) enhancing feasibility by accounting for the social and political dimensions related to restoration.

I use Kalimantan (Indonesian Borneo) as a case study area, as it represents a region that is globally important in terms of biodiversity and carbon storage. Kalimantan’s forests also provide essential livelihoods for local people. Yet, rapid deforestation and forest degradation threaten the forests in this region. Chapter 2 verifies that forest loss and degradation is the most significant threat to biodiversity in Kalimantan, impacting more than 80% of threatened animal species and 60% of threatened plant species. The future of Kalimantan’s wildlife depends on the survival of species in human-modified landscapes including in restored forest.

Quantifying carbon benefit in a policy, such as forest restoration under the REDD+, requires a standardised tool, which has not been available for data-poor regions including Kalimantan. In Chapter 3, I examine a process-based model, called 3-PG (physiological principles for predicting growth), to estimate the above-ground biomass (AGB) content of the major forest types occurring on the island of Borneo. Using readily available climate and soil data, the results indicate the 3-PG model accurately predicts AGB compared with field-measured data, revealing the potential application of this model for carbon sequestration analyses. The datasets along with a set of parameters used in this chapter are employed in the subsequent chapters.

Degraded tropical forests are characterised by a broad spectrum of forest condition states, mainly as a result of varying intensities of logging and fire. In Chapter 4, I develop a new framework to

optimally allocate restoration investments to forests of varying condition, with two contrasting objectives: carbon and biodiversity. The work takes into account a diverse suite of restoration techniques and their costs and quantifies time-dependent restoration benefits in terms of carbon sequestration and the improvement of habitat for threatened mammals. I find that the distribution of investments is highly dependent on the restoration objective, which inevitably involves trade-offs between objectives. Nonetheless, for greatest achievement of both objectives, I demonstrate that restoring highly degraded Bornean lowland forest should receive the greatest investment.

Environmental offsetting is an emerging opportunity to leverage restoration over economic activities. Employing a backcasting approach, Chapter 5 presents the first investigation of the potential application of restoration as a policy tool to offset carbon and biodiversity loss from agricultural development at a landscape scale. Using an oil-palm case study in Kalimantan, I find offsetting biodiversity loss from past oil-palm plantation developments requires 8.7% of Kalimantan’s landmass to be restored at a cost of US$7.6 billion. In contrast, compensating carbon emissions would require an area of less than 2% of the region at a cost of US$1.8 billion by restoring degraded peatlands, including the failed Ex-Mega Rice Project in Central Kalimantan. My findings raise questions on the overall capacity of the oil-palm industry, and the agriculture sector more broadly, to fully offset biodiversity loss.

Forest restoration has had variable success, with performance strongly moderated by specific socio- ecological and political contexts within a region. To enhance social feasibility and political permissibility in restoration, Chapter 6 demonstrates the first integration of a socio-ecological systems framework with systematic decision-making to develop a context-specific restoration plan for livelihoods provision. I compare areas that were prioritised for restoration identified solely on the basis of biophysical criteria with those that combine socio-political and biophysical criteria. It emerges that incorporating the socio-political context alters the identification of priority areas for restoration, with only half the priority areas remaining the same with, and without, the socio-political factors. While the social feasibility and political permissibility can be enhanced, accounting for these constraints is likely to incur substantial opportunity costs. My framework reveals significant deficiencies and inefficiencies associated with existing restoration policies for Kalimantan.

My thesis demonstrates that the best decisions for landscape-scale restoration are not simply made as the result of a binary answer: restore or not restore. Using carefully developed frameworks, I have shown how and where restoration should be carried out, with cost-effective implementation and with opportunities for landscape-scale projects. By putting the theoretical analyses undertaken in my thesis

into the context of a globally important case study, my research provides evidence for science-based restoration policy that is transparent and transferable to achieve ambitious global policy targets.

iii

Declaration by author

This thesis is composed of my original work, and contains no material previously published or written by another person except where due reference has been made in the text. I have clearly stated the contribution by others to jointly-authored works that I have included in my thesis.

I have clearly stated the contribution of others to my thesis as a whole, including statistical assistance, survey design, data analysis, significant technical procedures, professional editorial advice, and any other original research work used or reported in my thesis. The content of my thesis is the result of work I have carried out since the commencement of my research higher degree candidature and does not include a substantial part of work that has been submitted to qualify for the award of any other degree or diploma in any university or other tertiary institution. I have clearly stated which parts of my thesis, if any, have been submitted to qualify for another award.

I acknowledge that an electronic copy of my thesis must be lodged with the University Library and, subject to the General Award Rules of The University of Queensland, immediately made available for research and study in accordance with the Copyright Act 1968.

I acknowledge that copyright of all material contained in my thesis resides with the copyright holder(s) of that material. Where appropriate I have obtained copyright permission from the copyright holder to reproduce material in this thesis.

Publications during candidature

Budiharta, S., Meijaard, E. in press. State of Kalimantan’s biodiversity. In: Resosudarmo B.P., Imansyah, Napitupulu, L. (Eds), Development, Environment and the People of Kalimantan. Indonesian Regional Science Association (IRSA), Jakarta.

Budiharta, S., Meijaard, E., Erskine, P., Rondinini, C., Pacifici, M., Wilson, K.A. 2014. Restoring degraded tropical forests for carbon and biodiversity. Environmental Research Letters 9, 114020. [DOI: 10.1088/1748-9326/9/11/114020].

Budiharta, S., Slik, F., Raes, N., Meijaard, E., Erskine, P., Wilson, K.A. 2014. Estimating the above- ground biomass of Bornean forest. Biotropica 46, 507-511. [DOI: 10.1111/btp.12132].

Soejono, Budiharta, S. and Arisoesilaningsih, A. 2013. Proposing local trees diversity for rehabilitation of degraded lowland areas surrounding springs. Biodiversitas, 14: 37-42. [DOI: 10.13057/biodiv/d140106].

Wilson, K.A., Auerbach, N.A., Sam K., Magini, A.G., Moss, A.S.L., Langhans, S.D., Budiharta, S., Terzano, D., Meijaard, E. 2016. Conservation research is not happening where it is most needed. PLoS Biology, 14 (3), e1002413. [DOI: 10.1371/journal.pbio.1002413]

Budiharta, S., Meijaard, E., Wells, J.A., Abram, N.K, M., Wilson, K.A. in revision. Enhancing feasibility: incorporating a socio-ecological systems framework into restoration planning. Environmental Science and Policy.

Abram, K.A., Meijaard, E., Wilson, K.A., Davis, J.T., Wells, J.A., Ancrenaz, M., Budiharta, S., Durrant, A., Fakhruzzi, A., Runting, R.K., Gaveau, D., Mengersen, K. in review. Oil palm- community conflict mapping in Indonesia: A case for better community liaison in planning for development initiatives. Applied Geography.

Publications included in this thesis

My PhD thesis is prepared following the format of thesis by publication. Chapter 2, 3 and 4 are published or in press, Chapter 5 is in final preparation for submission and Chapter 6 is in revision. For the chapters prepared for publication (Chapters 2-6), I use the plural term “we” to acknowledge the contributions by other authors.

Contributor Statement of contribution Budiharta S (Candidate) Conceived study (50%) Reviewed literature (60%) Wrote and edited paper (60%) Meijaard E Conceived study (50%) Reviewed literature (40%) Wrote and edited paper (40%)

Budiharta, S., Slik, F., Raes, N., Meijaard, E., Erskine, P., Wilson, K.A. 2014. Estimating the above- ground biomass of Bornean forest. Biotropica 46, 507-511. – incorporated as Chapter 3

Contributor Statement of contribution Budiharta S (Candidate) Conceived and designed study (90%) Collated and synthesised data (50%) Ran model and analysed results (100%) Wrote and edited paper (70%) Slik F Provided data (40%) Wrote and edited paper (5%) Raes N Provided data (10%) Wrote and edited paper (5%) Meijaard E Wrote and edited paper (5%) Erskine P Wrote and edited paper (5%) Wilson KA Conceived and designed study (10%) Wrote and edited paper (10%)

Contributor Statement of contribution Budiharta S (Candidate) Conceived and designed study (85%) Collated and synthesised data (70%) Ran decision-support algorithm and analysed results (100%) Wrote and edited paper (70%) Meijaard E Conceived study (5%) Wrote and edited paper (5%) Erskine P Wrote and edited paper (5%) Rondinini C Provided data (30%) Wrote and edited paper (5%) Pacifici M Wrote and edited paper (5%) Wilson KA Conceived and designed study (10%) Wrote and edited paper (10%)

Budiharta, S., Meijaard, E., Wells, J.A., Abram, N.K., Wilson, K.A. in revision. Enhancing feasibility: incorporating a socio-ecological systems framework into restoration planning. Environmental Science and Policy. – incorporated as Chapter 6

Contributor Statement of contribution Budiharta S (Candidate) Conceived and designed study (90%) Collated and synthesised data (70%) Ran decision-support algorithm and analysed results (100%) Wrote and edited paper (75%) Meijaard E Provided data (30%) Wrote and edited paper (5%) Wells JA Wrote and edited paper (5%) Abram NK Wrote and edited paper (5%) Wilson KA Conceived and designed study (10%) Wrote and edited paper (10%)

vii

Contributions by others to the thesis

Associate Professor Kerrie Wilson provided advice on idea development and study design as well as edited the chapters prepared for publication and the overall thesis. Dr Erik Meijaard contributed to idea development and access to data. Data were provided by Associate Professor Ferry Slik, Dr Niels Raes, Dr David Gaveau, Dr Matthew Struebig, Dr Andreas Wilting, Dr Stephanie Kramer-Schadt, Jürgen Niedballa, Dr Nicola Abram, Ade Wahyu, Heri Setiawan, Dr Gusti Hardiansyah and Muhammad Yusuf.

Chapter 5

Budiharta, S., Meijaard, E., Gaveau, D., Struebig, M., Wilting, A., Kramer-Schadt, S., Niedballa, J., Raes, N., Possingham, H.P., Maron, M., Wilson, K.A. in preparation for Nature Climate Change. Paying back carbon and biodiversity debts from oil-palm plantations in Kalimantan.

The candidate conceived the idea, designed the study, ran the analysis and wrote the manuscript. KA Wilson advised on idea development and study design. E Meijaard, D Gaveau, M Struebig, A Wilting, S Kramer-Schadt, J Niedballa and N Raes provided the data. All authors discussed the results and edited the manuscript.

Statement of parts of the thesis submitted to qualify for the award of another degree

None.

viii

Acknowledgements

My deepest gratitude to God – Most Merciful. I now understand that doing a PhD is not only an academic process but also a spiritual journey. Thanks for helping me survive this journey!

I am indebted to my primary supervisor Kerrie Wilson. Kerrie, you gave me the freedom to define the destination of my research, and when I was lost you were always around to put me back on the right path. You inspired me with academic kindness, patience, effective and efficient work, and research networking which are valuable experiences that will shape my career and be something important I will share with others. Thank you for being my mentor, motivator and role model!

I owe Erik Meijaard. Erik, you fascinate me with many ideas about tropical conservation, some of which generate debates but that is how we create new knowledge. You opened so many doors to access data, I cannot imagine how my thesis would be without such data. Thanks for sparking my brain and fuelling my research!

I am honoured to be in the Centre for Biodiversity and Conservation Science (CBCS) led by Hugh Possingham. Hugh, I will always remember a note in your door: “If I know what I am doing, why am I doing research?” Thank you for the inspiration!

I appreciate the help of my co-supervisor Peter Erskine; my PhD readers Rod Fensham and Clive McAlpine; and my collaborators Niels Raes, Ferry Slik, Jessie Wells, Nicola Abram, David Gaveau, Matthew Struebig, Martine Maron, Andreas Wilting, Stephanie Kramer-Schadt and Jürgen Niedballa. It is my privilege to work with great people like you.

I thank my friends Maria Martinez-Harms, Nancy Auerbach, Veronica Gama, Hernan Caceres, Liz Law, Courtney Morgans, Angela Guerrero-Gonzales, Ruben Venegas Li, Hsien-Yung Lin and all the members of the CBCS. To my Indonesian friends in Brisbane, especially the members of Buaya Keroncong Brisbane, matur nuwun!

I acknowledge the Australian Awards Scholarships for providing financial support and the Indonesian Institute of Sciences for allowing me to pursue this PhD and accepting me when I finish.

My gratitude goes to my mother (Ibuk), my father (Bapak) and all my family. Ibuk, thank you so much for tirelessly mentioning my name in your prayers, day and night!

And for my love Sadrah Devi and my son Rabbani Marvel, thank you both for sharing the best moments in my life. Devi, I had never dreamed to be a scientist and do a PhD. But you have always encouraged me, given me strength and believed that I could do better in this life. Thanks for everything! ix

Keywords

Forest landscape restoration, decision science, restoration planning, spatial prioritisation, degraded tropical forest, biodiversity, carbon, livelihoods, trade-off, offsetting.

Australian and New Zealand Standard Research Classifications (ANZSRC)

ANZSRC code: 050205, Environmental Management, 40%

ANZSRC code: 050202, Conservation and Biodiversity, 40%

ANZSRC code: 050207, Environmental Rehabilitation, 20%

Fields of Research (FoR) Classification

FoR code: 0502, Environmental Science and Management, 100%

Table of Contents

Abstract ...... i

Declaration by author ...... iv

Publications during candidature ...... v

Publications included in this thesis ...... vi

Contributions by others to the thesis...... viii

Acknowledgements...... ix

Keywords ...... x

Table of Contents ...... xi

List of Figures ...... xv

List of Tables ...... xviii

List of Abbreviations used in the thesis ...... xix

1. Introduction ...... 1

1.1. Deforestation and forest degradation in the tropics ...... 2

1.2. Reversing deforestation through restoration ...... 4

1.3. The benefits of restoration...... 4

1.4. Knowledge gaps in landscape-scale restoration ...... 6

1.4.1. Addressing landscape heterogeneity ...... 6

1.4.2. Leveraging restoration within competing land uses ...... 7

1.4.3. Enhancing feasibility in landscape-scale restoration ...... 8

1.5. Decision-making for restoration...... 9

1.6. The context of the thesis and study area ...... 12

1.7. Structure of Thesis ...... 13

2. State of Kalimantan’s biodiversity ...... 16

2.1. Introduction ...... 17

2.2. Plant diversity ...... 18

2.3. Animal diversity ...... 20

2.4. The value of biodiversity to local livelihoods ...... 22

2.5. Threats to biodiversity in Kalimantan ...... 24

2.6. Conservation strategies ...... 27

2.7. Acknowledgements ...... 29

3. Estimating the above-ground biomass of Bornean forest ...... 30

3.1. Abstract ...... 31

3.2. Introduction ...... 31

3.3. Methods ...... 32

3.4. Results ...... 34

3.5. Discussion ...... 35

3.6. Acknowledgements ...... 37

4. Restoring degraded tropical forests for carbon and biodiversity ...... 38

4.1. Abstract ...... 39

4.2. Introduction ...... 39

4.3. Methods ...... 42

4.3.1. Study area ...... 42

4.3.2. Analysis framework ...... 43

4.3.3. Restoration zones ...... 43

4.3.4. Restoration actions and costs ...... 45

4.3.5. Restoration investment ...... 47

4.3.6. Estimating AGB accumulation ...... 47

4.3.7. Estimating the suitability of habitat for threatened mammals ...... 48

4.4. Results ...... 49

4.5. Discussion ...... 53

4.6. Conclusion ...... 57

4.7. Acknowledgements ...... 57

5. Paying back carbon and biodiversity debts from oil-palm plantations in Kalimantan... 58

5.1. Abstract ...... 59

5.2. Introduction ...... 59

xii

5.3. Methods ...... 61

5.3.1. Oil palm-driven land conversion data ...... 61

5.3.2. Quantifying the ecological impacts by oil-palm plantations ...... 61

5.3.2.1. Carbon emissions from oil-palm plantations ...... 61 5.3.2.2. The loss of native vegetation by oil-palm plantations ...... 63 5.3.2.3. The loss of mammal habitat by oil-palm plantations ...... 64 5.3.3. Potential areas for restoration ...... 64

5.3.4. Quantifying the benefits of restoration ...... 65

5.3.4.1. The benefit on carbon ...... 65 5.3.4.2. The benefit on the re-establishment of native vegetation ...... 66 5.3.4.3. The benefit on the re-creation of mammal habitat ...... 67 5.3.5. The costs of restoration ...... 68

5.3.6. Prioritising areas for restoration offsetting ...... 69

5.4. Results ...... 70

5.5. Discussion ...... 76

5.6. Conclusion ...... 78

6. Enhancing feasibility: incorporating a socio-ecological systems framework into restoration planning ...... 79

6.1. Abstract ...... 80

6.2. Introduction ...... 80

6.3. Methods ...... 82

6.3.1. Analytical framework ...... 82

6.3.2. Case study ...... 83

6.3.3. Diagnosing the social-ecological systems that represent an opportunity for restoration ...... 84

6.3.4. Identifying the context for restoration activities ...... 85

6.3.4.1. Ecological context ...... 85 6.3.4.2. Socio-economic context ...... 86 6.3.4.3. Political context ...... 88 6.3.5. Opportunity costs of restoration...... 88

6.3.6. Prioritisation analysis and planning scenarios ...... 89

xiii

6.4. Results ...... 90

6.5. Discussion ...... 94

6.6. Conclusion ...... 97

6.7. Acknowledgements ...... 97

7. General Discussion ...... 98

7.1. Identifying contextual ecosystem services for restoration ...... 99

7.2. Restoration planning in data-poor regions ...... 101

7.3. Research contributions: advancing decision science in landscape-scale restoration to inform policy ...... 102

7.3.1. Accounting for landscape heterogeneity in restoration planning ...... 102

7.3.2. Leveraging restoration through offsetting policy ...... 104

7.3.3. Utilising social-ecological systems framework to enhance feasibility in restoration ...... 107

7.4. Synthesis and future directions ...... 108

7.5. Concluding remarks ...... 111

References ...... 112

Appendices ...... 148

Appendices for Chapter 3 ...... 149

Appendices for Chapter 4 ...... 154

Appendices for Chapter 5 ...... 159

Appendices for Chapter 6 ...... 167

xiv

List of Figures

Figure 1.1. Examples of decision-making processes in restoration...... 10 Figure 1.2. Thesis structure...... 15

Figure 2.1. Variation in mammalian species richness and endemism across the Indo-Malayan Region (Boitani and Maiorano, unpublished data)...... 22 Figure 2.2. Processes that threaten endangered species in Kalimantan (IUCN, 2012)...... 25

Figure 3.1. Comparison between the 3-PG modelled (over 100 years) and field-measured data of AGB (Mg per hectare) and basal area (m2 per hectare) across Borneo. The straight line indicates 1:1 line...... 34 Figure 3.2. Predicted AGB (Mg per hectare) accumulation over 100 years for each forest type: (a) lowland forest; (b) montane forest; (c) heath forest; and (d) peat swamp forest. Solid line is the average AGB modelled for each forest type; dashed lines are the upper and lower confidence intervals...... 35

Figure 4.1. Map of forest conditions across forest types in East Kalimantan. The study area was restricted to the Forest Estate with native vegetation under the authority of the Ministry of Forestry and excluded areas of non-Forest Estate, industrial timber plantation (ITP), mangroves and water bodies...... 45

Figure 5.1. Net carbon gain from restoration on degraded areas (i.e. logged forest, burned forest, agroforest and non-forest) over 25 years. Carbon gain in mineral soils was assumed from the sequestration of above-ground biomass (AGB) estimated using the 3-PG model. Carbon gain in peatlands was assumed from AGB sequestration and avoided emissions from peat oxidation and burning...... 66 Figure 5.2. Net gain in term of re-establishment of native vegetation from restoration on degraded areas measured as intactness-adjusted area (IAA) of floristic ecoregion. Note that two floristic ecoregions (i.e. hill forest and montane forest of the upper Kapuas) are not mapped as no oil- palm plantations have been established...... 67 Figure 5.3. The cost of restoration in degraded areas accounting for the implementation and opportunity costs. Degraded areas were defined as non-forest, agroforest, burned forest and logged. The implementation cost covered expenses for planting and maintenance, and cost for hydrological restoration if it occurred on degraded peatlands. Opportunity cost was assumed as forgone revenues from oil-palm plantations, or logging or agroforestry assigned on the basis of biophysical suitability and existing land cover class...... 69

Figure 5.4. Oil palm-driven carbon emissions in Kalimantan between 1973 and 2013. Total net carbon emissions across land cover classes and soil types, assuming a 25-year oil-palm planting cycle. Negative carbon emissions indicate net carbon gain (carbon sequestered from oil-palm plantation exceeds carbon loss associated with converting non-forested areas). Error bars are lower and upper estimates of net carbon emissions...... 70 Figure 5.5. Average net carbon emissions per hectare from oil-palm plantation development for each soil type, peat depth, and land-cover type over a 25-year planting cycle. Error bars represent lower and upper estimates of carbon emissions...... 71 Figure 5.6. Resources required to offset the impacts of oil-palm plantation in Kalimantan. a, Extent of landscape selected for restoration offsetting. b, Total offsetting costs accounting for opportunity and implementation costs. Each offsetting scenario aims to compensate for the loss of: floristic ecoregion (scenario F); mammal habitat (scenario M); floristic ecoregion and mammal habitat (scenario F&M); carbon (scenario C); carbon and floristic ecoregion (scenario C&F); carbon and mammal habitat (scenario C&M). Error bars represent the range of results accounting for lower and higher estimates of total carbon emissions...... 74 Figure 5.7. Priority areas for offsetting the impacts of oil-palm development via restoration in Kalimantan. Each figure represents different scenarios for compensating the loss of: (a) floristic ecoregion; (b) mammal habitat; (c) floristic ecoregion and mammal habitat; (d) carbon; (e) carbon and floristic ecoregion; (f) carbon and mammal habitat. Label refers to province: West Kalimantan (WK), Central Kalimantan (CK), South Kalimantan (SK), East Kalimantan (EK) and North Kalimantan (NK)...... 75 Figure 5.8. Overlapping priority areas between offsetting the impacts of oil-palm development on carbon emissions and compensating biodiversity loss (i.e. mammal’s habitat and floristic ecoregion combined) in Kalimantan...... 76

Figure 6.1. Analytical framework for operationalising social-ecological systems concepts in restoration planning. Black boxes depict the social-ecological systems framework adapted from Ostrom (2009) including direct links (solid arrows) and feedbacks (dashed arrows). Blue boxes depict steps in systematic decision-making adapted from Gregory et al. (2012) and Ban et al. (2013). Grey shading shows the links between elements of the two frameworks. Italic text illustrates applications to the contextual setting in the study area (Paser District, East Kalimantan)...... 82

xvi

Figure 6.2. The land uses of Paser District as of 2011. Protected areas (covering 209,355 hectares) consist of two categories: Conservation Forest (e.g. national park and nature reserve) and Protection Forest (for protecting hydrological services). Areas for logging have either active concessions (302,239 hectares) or are without any concession licences (70,253 hectares). Other land uses include industrial timber plantations (18,323 hectares), small-scale agriculture (32,771 hectares), degraded lands/severely logged forests (260,570 hectares), oil-palm plantations (145,312 hectares), mined areas (8,150 hectares) and settled areas (12,937 hectares)...... 83 Figure 6.3. Maps of priority areas for agroforestry restoration under four social-ecological scenarios: (a) Scenario 1 (ecological variables only); (b) Scenario 2 (ecological and social variables); (c) Scenario 3 (ecological and social variables with existing political constraints); (d) Scenario 4 (ecological and social variables with constraints associated with land use removed). The coloured areas show the potential areas for agroforestry restoration from the ecological context (i.e. areas suitable for at least one fruit/crop type, and where the land cover is compatible)...... 92 Figure 6.4. Outcomes for each scenario in regards to biophysical suitability, social feasibility and opportunity cost for the highest priority areas for agroforestry restoration (areas within the top 20%). For each class of values (biophysical or social), the aggregate value is the mean proportion of the landscape’s summed potential value that is encompassed by the selected restoration areas. The proportion of total landscape value was calculated separately for each variable in each class, and then the mean taken across all variables within the class...... 93 Figure 6.5. Cost-effectiveness across four scenarios formulated as total cost required to retain total values of (a) crop suitability and (b) social variables, when a given proportion of the available landscape is selected (X axis)...... 93

xvii

List of Tables

Table 2.1. Species counts for the main Indonesian islands (excludes introduced species) ...... 21

Table 4.1. Restoration zones in East Kalimantan based on a forest condition matrix across five forest types. The five condition classes are critically degraded/deforested (CDF), highly degraded (HDF), moderately degraded (MDF), lightly degraded (LDF) and primary forest (PF). The highest value of AGB observed for each forest type was assumed to represent intact/pristine forest, with the condition classes assigned in relation to this baseline such that CDF = < 20%; HDF = 21-40%; MDF = 41-60%; LDF = 61-80%; and PF = >81%. AGB stock is in Mg per hectare...... 44 Table 4.2. Restoration actions and costs (in US$ per hectare) across restoration zones. Restoration costs consist of the implementation costs of planting (Table A4.1) and the revenue foregone from timber extraction...... 47 Table 4.3. Summary of degradation levels across forest types: total extent (in 000 of hectares) and the proportion of the overall extent of this forest type...... 49 Table 4.4. Recommended areas for restoration under each scenario using the lower investment level (US$500 million)...... 51 Table 4.5. Recommended areas for restoration with investment using the higher restoration budget (US$1.25 billion)...... 52

Table 5.1. Extent of native vegetation in Kalimantan replaced by oil-palm plantations. Intactness- adjusted area (IAA) represents ecological integrity of extant native vegetation. Native vegetation is represented by floristic ecoregion using clustering analyses (Raes, 2009). For the purpose of this paper, some nomenclatures of the ecoregions were modified from the original dataset (Raes, 2009)...... 72

xviii

List of Abbreviations used in the thesis

AGB Above-ground biomass CBD Convention on Biological Diversity Dbh Diameter at breast height EMRP Ex-Mega Rice Project (Central Kalimantan) ERC Ecosystem Restoration Concession IAA Intactness-adjusted area ITP Industrial timber plantation IUCN International Union for Conservation of Nature MaxEnt Maximum Entropy MODIS Moderate Resolution Imaging Spectroradiometer MoF Ministry of Forestry (Indonesia) NPV Net Present Value NTFP Non-timber forest product PADDD Protected Area Downgrading, Downsizing and Degazettement REDD+ Reducing Emissions from Deforestation and forest Degradation RIL Reduced-Impact Logging RSPO Roundtable on Sustainable Palm Oil SES Social-ecological systems UNFCCC United Nations Framework Convention on Climate Change

xix

Chapter 1. Introduction

Chapter 1

Introduction

No publication associated with this chapter

Chapter 1. Introduction

1. Introduction

1.1. Deforestation and forest degradation in the tropics

Tropical forest has been suffering extensive loss and degradation since the 1950s (DeFries et al., 2002; Food and Agriculture Organisation, 2010; Hansen et al., 2013; Woodwell et al., 1983). Although the overall rate of deforestation in the tropics has declined substantially from approximately 12 million hectares per year in the 1980s to 5.5 million hectares per year in the 2000s (Achard et al., 2002; Asner et al., 2009; DeFries et al., 2002; Food and Agriculture Organisation, 2010), the reasons for this decline are unclear. Some argue it is due to improved knowledge on the management and protection of tropical forests (Blaser et al., 2011) whereas others see it simply as a consequence of a reduced forest extent (Shearman et al., 2012). While forest is converted to other land uses, including more than 80 million hectares for agriculture (Gibbs et al., 2010), approximately 500–600 million hectares (or 30–40% of the total forest area in the tropics) has been left in a degraded state (Blaser et al., 2011; Brown and Lugo, 1990; ITTO, 2002). Selective logging alone contributes to the degradation of 400 million hectares of tropical forests (Asner et al., 2009; Blaser et al., 2011).

Extensive tropical deforestation and forest degradation have had both direct and indirect impact on biodiversity and forest ecosystem services at different spatial scales. While tropical regions cover only 7% of the total land mass on earth, they contain approximately half the global biodiversity (Wilson, 1988). For example, tropical forests comprise 40,000–53,000 tree species with most of them having an extremely narrow distribution and rare populations, often being represented by only one or two individuals per hectare (Roos et al., 2004; Sheil et al., 2010; Slik et al., 2015). To put this into context, temperate Europe harbours only 124 tree species (Slik et al., 2015). Consequently, if tropical deforestation continues at a rate of 10 million hectares annually, between 2–7% of species across all taxa will become extinct by 2025, equating to the loss of 20–75 species per day (Reid, 1992).

Tropical forest is among the richest ecosystems for carbon storage with 195 PgC stored as above- ground biomass (AGB), equating to 54% of global AGB (Liu et al., 2015). The comparatively high value of AGB in tropical forests is primarily contributed by the high density of large trees with a diameter greater than 70 cm (Slik et al., 2013). While they account for only 11% of the global peatland area, tropical peatlands contain 82–92 PgC, mostly stored as soil carbon (Page et al., 2011). As such, deforestation and degradation of tropical forests and peatlands has become the second largest source of anthropogenic greenhouse gas emissions, releasing 1–2 PgC per year and contributing to 20% of total global emissions (Gullison et al., 2007; Houghton, 2005; Liu et al., 2015; van der Werf et al., 2009).

Chapter 1. Introduction

At a local scale, approximately 300 million people directly depend on tropical forests for various provisioning services such as timber and non-timber forest products (Agrawal et al., 2013; ITTO, 2002). Tropical forests generate an annual income of US$440 (US$217–1,354) per household for forest-frontier communities, contributing between 5.5–63% of total household incomes (Angelsen et al., 2014). As access to markets and health facilities is generally limited, forest is a primary resource system for forest-dependent people providing a variety of wild foods, energy sources (e.g. fuel wood) and materials for medicinal purposes (Agrawal et al., 2013; Wunder et al., 2014). The loss of these benefits due to deforestation and forest degradation potentially disrupts regular cash income sources of forest-dependent communities and increases vulnerability in relation to food security (Adger, 2006; Agrawal et al., 2013; Kasperson et al., 2005; Vira et al., 2015). In some cases, the decline of the provisioning services provided by forest ecosystems contributes to deepening poverty and social conflict (Millennium Ecosystem Assessment, 2005).

A range of policy options has been advocated to halt the continuing loss and degradation of tropical forests. In the context of biodiversity conservation, the primary strategy is the expansion of protected area networks (Watson et al., 2014). While ideal in terms of philosophical goals, the performance of forest protected areas varied across tropical regions with many rated as ineffective due to downgrading, downsizing and degazettement (PADDD), illegal logging and land encroachment (Curran et al., 2004; Gaveau et al., 2012; Laurance et al., 2012; Levang et al., 2012). A second approach is through improved governance and better management of production forests through Reduced-Impact Logging (RIL) and forest certification (Ebeling and Yasué, 2009; Putz et al., 2012). Yet, only 17 million hectares (9.2%) of tropical production forest under selective logging concessions is certified as sustainable forest (Blaser et al., 2011; Putz and Pinard, 1993). Certification schemes are also promoted for sustainable agriculture to mitigate the impacts of forest clearing on biodiversity, such as the Roundtable on Sustainable Palm Oil (RSPO), but it is criticized for being limited achievement in mitigating the impacts as the main focus is only on rare species or ecosystems (Edwards and Laurance, 2012). The third approach is through agricultural intensification to improve land productivity, which in turn reduces the pressure to clear more forests (Green et al., 2005; Tilman et al., 2011). However, it is still not clear whether intensification reduces the agricultural footprint overall (Law et al., 2015; Law and Wilson, 2015). The fourth strategy is through forest restoration (Lamb et al., 2005; Meyfroidt and Lambin, 2011). Despite its practical significance, restoration has received less academic attention compared with other strategies (Possingham et al., 2015; Suding, 2011).

Chapter 1. Introduction

1.2. Reversing deforestation through restoration

The urgency for restoration has emerged in recent decades with the rapidly ongoing global biodiversity crisis, deterioration of ecosystem services and climate change (Bullock et al., 2011; Suding, 2011). Various global policy instruments have been sought to hasten restoration. The Millennium Ecosystem Assessment proposed restoration as one of the main policy interventions to recover the delivery of ecosystem services in degraded areas when protection management alone would be ineffective (Millennium Ecosystem Assessment, 2005). More recently, payment for ecosystem services in the form of REDD+ (Reducing Emissions from Deforestation and forest Degradation) included forest restoration as a strategy for enhancing carbon stocks (Alexander et al., 2011; UNFCCC, 2007). The Convention on Biological Diversity (CBD) through the Aichi Target 14 stated that by 2020 essential services, including the provision of livelihoods, should be restored taking into account women, indigenous and local communities and the poor and vulnerable (Convention on Biological Diversity, 2011). The CBD also set the target (Aichi Target 15) for at least 15% of degraded ecosystems to be restored by 2020 for ecosystem resilience, biodiversity conservation and carbon enhancement.

Restoration is not a new practice in tropical regions with some countries such as the Philippines commencing reforestation efforts early in the last century, not long after the increasing trend of deforestation and forest degradation in the region (de Jong, 2010; Le et al., 2014). According to forest transition theory, the history of tropical forest restoration occurred as the consequence of three major drivers (Lamb, 2010; Meyfroidt and Lambin, 2011; Rudel et al., 2005): (i) the scarcity of forest products, mainly woods and foods, owing to the preceding deforestation; (ii) ecological crisis due to increasing catastrophic events such as floods, fires, land denudation and soil erosion; and (iii) change in social lifestyle from an agriculture-based society into urban society as a result of urbanisation, leading to agricultural abandonment. In relation to the characteristics of the drivers, forest restoration appears in various forms including naturally regenerating forests following land abandonment, enrichment planting in secondary forests to provide timber and non-timber products, timber plantations including either monoculture or mixed species comprising native or exotic species, ecological planting to enhance ecological resilience, and agroforestry (Lamb, 2010; Lamb et al., 2005; Meyfroidt and Lambin, 2011).

1.3. The benefits of restoration

Although restoration will not fully recover all disturbed ecosystems to an intact state within the short term (Curran et al., 2013), it could mitigate further losses of biodiversity and ecosystem services

Chapter 1. Introduction

(Millennium Ecosystem Assessment, 2005). Averaged across various global biomes, restoration could enhance biodiversity and various ecosystem services by 44% and 25% respectively compared with a degraded state and be able to achieve 86% and 80% of those services in intact systems (Benayas et al., 2009). The greatest increase would be achieved in tropical terrestrial ecosystems rather than any other biomes, suggesting a window of opportunity to restore degraded ecosystems in this region (Benayas et al., 2009).

Forest restoration in the tropics could enhance biodiversity values significantly. For example, naturally regenerating forest in abandoned pasture lands in the tropical Andes contains comparable bird species richness, including 82% of threatened birds, to that of primary forest after 15–30 years of abandonment (Gilroy et al., 2014). Artificial treatments, such as enrichment planting in severely logged forest, could accelerate the biodiversity outcomes indicated by the rapid recovery of bird populations compared with natural regrowth (Edwards et al., 2009). While of less benefit for overall biodiversity, monoculture timber plantations (e.g. Acacia mangium) could still provide marginal habitat for generalist species such as orangutans (Meijaard et al., 2010a).

Restoration of tropical forest has the potential to mitigate climate change (Venter et al., 2012). Tropical rainforest has the highest net primary productivity among terrestrial ecosystems on earth because of the large amount of available water and solar energy throughout the year (Monteith, 1972; Rosenzweig, 1968; Running et al., 2000). As such, restoring degraded forests in these regions could rapidly sequester carbon at an average rate of 3.33 MgC of AGB per hectare per year during its first 20 years after disturbance (Brown and Lugo, 1990; Silver et al., 2000). Within 30 years, the accumulated AGB carbon of restored forest approaches that of primary forest (Gilroy et al., 2014; Saner et al., 2012). A higher rate of carbon sequestration could be attained by conducting silvicultural techniques, such as through strip planting (Hardiansyah, 2011).

Restoration developed with specific strategies could also enhance ecosystem provisioning services to support livelihoods while improving soil condition and water retention (Cao et al., 2009; Le et al., 2014; Poffenberger, 2006; Soejono et al., 2013). An example from the Philippines shows that restoration using an agroforestry approach, such as in the form of fruit tree orchards and timber trees intercropped with annual crops, increases food security up to 14 times (Le et al., 2014). A well-known large-scale restoration program in Shinyanga, Tanzania demonstrates that restored lands provide an important source of fodder production for livestock during the dry season, hence securing protein sources for local communities (Monela et al., 2005). Reforestation projects also create jobs and generate incomes for villagers, which in turn could be used to purchase foods, providing a safety net

Chapter 1. Introduction when crops fail or local food supply is limited (Le et al., 2014; Monela et al., 2005; Nawir et al., 2007b).

1.4. Knowledge gaps in landscape-scale restoration

While restoration delivers eminent benefits and despite various policy tools and global targets being adopted, the knowledge underpinning the science of restoration ecology has lagged behind (Menz et al., 2013; Suding, 2011). This knowledge gap is argued as an obstacle to the effective implementation of large-scale restoration programs (Calmon et al., 2011; Chazdon, 2008; Chazdon et al., 2015; Hobbs and Norton, 1996). Existing studies have largely been focused on site-scale analysis (e.g. Ansell et al., 2011; Cole et al., 2011; Heriansyah et al., 2013; Omeja et al., 2011; Shoo et al., 2013; Zahawi and Holl, 2009), while research at a landscape scale is still lacking. A deficit of research undertaken at a landscape scale has resulted in a lack of knowledge of the spatial and temporal variability of restoration costs and benefits. Various opportunities for investing in restoration emerge, yet these have not been backed with robust science for the best uses of available resources (Chazdon et al., 2015). The interdependency among social and ecological variables in determining restoration outcomes has also not received sufficient attention (Dudley et al., 2005; Suding, 2011).

1.4.1. Addressing landscape heterogeneity

As the basic rule of thumb in landscape approaches, there is always a trade-off between social, economic and ecological objectives (Sayer et al., 2015; Sayer et al., 2013). Restoration implemented at a landscape scale delivers multiple benefits, which are rarely distributed evenly spatially and temporally. For instance, degraded areas in which restoration would sequester the highest above- ground biomass are not necessarily those that would significantly reduce groundwater recharge as well as mitigate soil erosion (Crossman and Bryan, 2009). Also, the rates of the above-ground carbon sequestration differ across time periods and follow a logistic rather than linear curve, depending on successional stage and restoration treatment (Brown and Lugo, 1990; Silver et al., 2000; Zahawi and Holl, 2009). The non-overlapping nature of the spatial and temporal distribution of restoration benefits implies that restoration should aim to maximise synergies and/or minimise trade-offs among benefits (Brown, 2005). Limited studies have integrated both the spatial and temporal distribution of restoration benefits at a landscape level.

Heterogeneity also applies on restoration strategy implemented in regard to landscape condition. This is particularly important in the tropics as degraded landscapes in this region are characterised by a broad spectrum of condition states ranging from totally deforested to moderately and lightly degraded (Langner et al., 2013; Sasaki et al., 2011; Sasaki and Putz, 2009). Landscape of varying condition

Chapter 1. Introduction implies that a single restoration strategy is unlikely to suit all areas, hence a diverse suite of strategies is required (Mori, 2001; Sasaki et al., 2011). When a set of restoration strategies can be applied, there are a number of potential combinations across the landscape complicating the selection of where to restore, the combination of strategies to use and the amount of each combination likely to achieve the greatest restoration benefit for the least cost. This remains an important unresolved problem in restoration ecology, especially in tropical regions.

Still lacking is knowledge related to the economics of restoration, an integration between the fields of restoration ecology and environmental economics (Blignaut et al., 2014; Wilson et al., 2012). Restoration is a high cost and labour intensive activity (De Groot et al., 2013; Wuethrich, 2007). With limited funding available, restoration should aim to achieve the greatest return on investment (i.e. efficiency) (Bullock et al., 2011; De Groot et al., 2013; Wilson et al., 2012). To define efficiency, there is a minimum of two components that should be accounted for: restoration benefit and costs associated with strategy implemented. However, the conceptual knowledge in quantifying restoration costs across heterogeneous landscapes contains some flaws such as the assumption that costs are uniform across landscape conditions (Evans et al., 2015; Kotiaho and Moilanen, 2015). Failure to conceptualise efficiency spatially and temporally will risk missing the best use of restoration investments, potentially channelling the money to inappropriate areas (Kotiaho and Moilanen, 2015). Therefore, a critical research question remains on how to develop a framework to define efficiency in restoration by accounting for costs and benefits over large and heterogeneous landscapes, and over long time periods.

1.4.2. Leveraging restoration within competing land uses

Restoration does not occur in a vacuum. In the context of landscape management, restoration competes with other land uses (Lamb et al., 2005; Sayer et al., 2013). Restoring degraded areas for less-marketable ecosystem goods and services implies forgoing alternative opportunities that might have been more profitable such as managing the areas for agriculture and mineral extraction (Calmon et al., 2011; Crossman et al., 2011; Kujala et al., 2015; Wilson et al., 2012). This challenge is more prevalent in tropical developing countries such as Indonesia where agriculture sectors have been significant contributors to national economies in recent decades, primarily due to the rise of oil palm (Koh et al., 2011; Meijaard and Sheil, 2013). Oil-palm plantations in Kalimantan (Indonesian Borneo) expanded at a rate of 232,800 hectares per year between 2000 and 2010 and the expansion will triple in the next five years (Carlson et al., 2013). These figures dwarf reforestation efforts in the region with less than 50,000 hectares of degraded land being reforested annually (Ministry of Forestry,

Chapter 1. Introduction

2014). In leveraging forest restoration, significant challenges lie in finding alternative ways to embed suitable land management in production landscapes.

As restoration competes poorly with lucrative land uses, several novel strategies have been introduced. These include restoration offsetting to compensate the impact of development activities on biodiversity and ecosystem services (BBOP, 2012; Blignaut et al., 2014; Gardner et al., 2013; Maron et al., 2012). Analysis on restoration offsetting in agricultural landscapes is rare compared with, for example, mining and urban development, which are generally less land-consumptive yet more profitable (Madsen et al., 2010). Furthermore, the functionality of restoration offsetting at a landscape scale has not been evaluated (Maron et al., 2012). For example, an offsetting project has been developed through Malua BioBank in Sabah, Malaysia to sell biodiversity credits in voluntary markets with potential buyers from logging, oil-palm and tourism industries (Madsen et al., 2010). Yet, this project is carried out at a site scale, potentially obscuring its performance (e.g. leakage, sub- optimal location of offsetting sites) in compensating environmental damage at a landscape scale. No studies have been attempted to assess the opportunities and challenges of restoration offsetting by agricultural sectors at the scale of whole landscapes.

1.4.3. Enhancing feasibility in landscape-scale restoration

Restoration is not only about biophysical elements such as biodiversity and carbon. Restoration also relates to people who are either taking benefits from restoration, or potentially being perversely affected by it. The Global Environment Facility of the World Bank developed an ambitious initiative to restore one million hectares of degraded riparian zones in the Brazilian Atlantic rainforest (Wuethrich, 2007). Yet, of 98 areas planned for reforestation, almost all sites are considered as failed due to limited support from local communities who were expected to be the main implementing actors (Wuethrich, 2007). The initiative faced initial opposition from local farmers who believed that restoration would be at the expense of their agricultural lands (Wuethrich, 2007). Once the goal was refined to restoring deteriorated natural water springs for irrigation, the initiative started to gain community support. Therefore, identifying appropriate strategies that align with the goal as well as the social context is essential for effective restoration implementation (Chazdon, 2008; Galabuzi et al., 2014).

Many restoration programs are planned and developed using a uniform strategy across regions by taking successful reference from other regions (Nawir et al., 2007a). This approach potentially results in a mismatch between predefined restoration goals and the socio-political situation on the ground (Guerrero et al., 2013). A systematic and repeatable theoretical framework to capture social and political dimensions to identify contextual restoration approaches does not exist. Even when the likely

Chapter 1. Introduction best approach is able to be identified, its social feasibility could vary spatially across regions with variables influenced by factors such as willingness to be involved and perceived socio-economic benefits generated from restoration (Curran et al., 2012; Le et al., 2014). Existing policies could also significantly influence the effectiveness of restoration, for example through laws and government regulations that either constrain or enable restoration and which have mixed spatial implications (Aronson et al., 2011). There is limited knowledge on how social feasibility and political permissibility should be considered in landscape-scale restoration to enhance the likelihood of success.

1.5. Decision-making for restoration

The complexity of restoration at a landscape scale implies that the answers to the questions of where and how much to restore and how to carry out the restoration are problematic (Mansourian and Vallauri, 2014; Stanturf et al., 2014; Suding, 2011). Decision science (or decision analysis) could help to structuralise the problems and find the best solutions when developing plans and making decisions for restoration. The systematic approaches employed in decision science ensure every step in planning and decision-making is transparent, measurable and repeatable (Gregory et al., 2012). There are generally five broad steps in decision-making processes (Martinez-Harms et al., 2015): (i) defining the problem and goal; (ii) identifying the objective(s) and performance measure(s); (iii) developing alternatives to achieve the objective; (iv) prioritising alternative(s) and assessing trade- offs; and (v) making the decision(s). The divergence of decision-making processes in restoration can be illustrated via the two case studies below (see Figure 1.1), which compare restoration for biodiversity conservation in Sumatra, Indonesia with restoration for carbon sequestration in Sabah, Malaysia.

The goal is a general expression of qualitative vision to provide direction in tackling a problem and are often value-laden (Bullock et al., 2011; Gregory et al., 2012; Pressey and Bottrill, 2009; Sayer, 2005). The goal prescribed at the initial stage of planning highly depends on the intended endpoint of restoration (Lamb et al., 2012). In the Sumatran case study a matrix of lowland forest, which is an important habitat for the Sumatran tiger (Panthera tigris sumatrae), elephants (Elephas maximus) and 305 bird species, has been in a degraded condition due to selective logging (Birdlife International, 2014; PT. Restorasi Ekosistem Indonesia, 2015). From the perspective of conservationists, this degraded forest is regarded as a problem necessitating restoration while agribusiness entities view it as an opportunity for oil-palm plantation development, showing contrasting values between both groups (Smit et al., 2013). A restoration project in the form of an Ecosystem Restoration Concession (ERC), called Harapan Rainforest (The Forests of Hope), has been established to improve the habitat

Chapter 1. Introduction quality of the forest and to protect it from oil-palm expansion for the preservation persistence of wildlife (Birdlife International, 2014). In the Sabah example, an area of logged-over mixed dipterocarp forest would be re-logged and converted to oil-palm plantation under a business-as-usual scenario, potentially releasing a large amount of carbon emissions (FACE the Future, 2011). Instead of being cleared, the INFAPRO project has been developed to restore this forest to avoid emissions and sequester carbon, which could be traded in the carbon market (FACE the Future, 2011). These examples demonstrate two different objectives: biodiversity conservation and carbon sequestration. Restoration could also be developed for multiple goals, for example by combining the goal for biodiversity and carbon simultaneously.

Decision-making processes Sumatra restoration Sabah restoration project project Goal Protecting and improving Avoiding emissions and the habitat of threatened sequestering carbon mammals and birds

Objective Restoring 98,000 ha of Restoring 25,000 ha of degraded forest cost- logged forest to sequester effectively 4.1 million tCO2-e over 30 years

Strategies Natural regeneration, Climber cutting and enrichment planting, enrichment planting selective thinning and direct seeding

Prioritisation Under development Under development

Consensus-based decision- Focus group discussions Not available making involving stakeholders

Figure 1.1. Examples of decision-making processes in restoration.

The objective is an articulation of the goal, which commonly uses quantitative measures (Gregory et al., 2012; Pressey and Bottrill, 2009). In the Sumatran example, the overarching objective is to cost- effectively restore and protect 98,000 hectares of degraded forest within its concession area (Birdlife International, 2014; PT. Restorasi Ekosistem Indonesia, 2015). For the Sabah project, the objective is to restore 25,000 hectares of logged forest within 30 years to achieve a net emissions reduction of

4.1 million tCO2-e (FACE the Future, 2011). While the objective is sometimes determined deductively through setting targets explicitly at the outset of a project, it could also be defined through inductive processes (i.e. quantitative measures are set after analytical observation).

Chapter 1. Introduction

Once the restoration objective is identified, plausible alternative strategies to achieve the objective are developed (Gregory et al., 2012; Lamb et al., 2005; Mansourian et al., 2005). In the Sumatran case study, a mix of strategies is being tested including natural regeneration, selective thinning, enrichment planting of native tree species and direct seeding (Harrison and Swinfield, 2015). A relatively similar approach is implemented in the Sabah project, which uses climber cutting and enrichment planting as the main strategies (FACE the Future, 2011). With more specific objectives, the restoration strategy requires careful development, for instance, through planting of trees that generate non-timber forest products (e.g. fruits, honey) to provide local communities alternative livelihoods to reduce pressure from timber cutting (Cao et al., 2009).

When implementing a restoration strategy, each site will have a varying contribution towards achieving the objective. The selection of areas to be restored requires careful consideration of which objectives could be achieved effectively with the available resources (Menz et al., 2013; Suding, 2011; Wilson et al., 2012). Prioritisation will be more complicated if two or more objectives with a broad range of alternative strategies are incorporated as there is likely to be a trade-off between the objectives. There are three common approaches to assessing the performance of alternative options (Wilson et al., 2012): (i) cost-effectiveness analysis by evaluating the performance measures of objectives gained via a unit of cost; (ii) cost-utility analysis by aggregating objectives into a single value that factors in the associated costs; and (iii) cost-benefit analysis by translating objectives and costs into a monetary value. Numerous decision-support tools have been developed to identify priority areas for restoration and these employ a wide variety of prioritisation algorithms of varying levels of complexity (Crossman and Bryan, 2006, 2009; McBride et al., 2010; Thomson et al., 2009; Wilson et al., 2011).

In the final stage, a decision must be made about which sites to restore and when this restoration will be undertaken (Gregory et al., 2012). This activity often involves collaborative and participatory approaches involving stakeholders and thus internalising values, norms and interests (Gregory et al., 2012). For example, several focus group discussions were held in Harapan Rainforest to explore the aspirations of the stakeholders including government and indigenous communities in order to resolve potential conflicts arising as a result of the restoration project (Silalahi and Erwin, 2015). During such learning processes, the existing plans, policies, economic instruments and institutional arrangements facilitating restoration could be refined or new ones might be created (Gregory et al., 2012; Martinez- Harms et al., 2015; Wilson et al., 2012). Multiple rounds of such learning processes are sometimes required to find consensus among stakeholders in order to yield sound decisions that have wide acceptance (Boedhihartono and Sayer, 2012; Gregory et al., 2012).

Chapter 1. Introduction

1.6. The context of the thesis and study area

This thesis develops and applies novel decision-support frameworks to guide the restoration of degraded tropical forests. The research in this thesis fills important knowledge gaps and provides fundamental guidance for the development of restoration policy in the context of the study region and tropical forests more broadly. By doing so, my thesis provides evidence for science-based restoration policy that is applicable and transferable.

In particular, this thesis focuses on three elements of ecosystem services delivered by restoration: habitat of biodiversity, carbon sequestration and provisioning services for livelihoods. The selection of biodiversity in my thesis is relevant to global policy contexts as epitomised in the Aichi Target 15 of the CBD (Convention on Biological Diversity, 2011). This thesis also accounts for carbon to provide insights on REDD+ policy development as it expands to include a new component (represented by the “+”) of the enhancement of forest carbon stocks (UNFCCC, 2007). The livelihoods provision is selected to align with the Aichi Target 14 of the CBD (Convention on Biological Diversity, 2011).

In national and regional contexts of the study region, my thesis has significance for informing policies related to conservation, land use planning and forestry in Indonesia. The Indonesian Government developed a new conservation strategy through Ecosystem Restoration Concessions (ERC) to restore degraded forests (Ministry of Forestry, 2008). ERCs are funded by private investments and are given rights to sell ecosystem services, including carbon and biodiversity. Although an ambitious target was set to develop 2.5 million hectares of state forest for ERCs by 2014, only 400,000 hectares for ERC licences have been granted so far (Boer, 2012; Ministry of Forestry, 2014). The framework developed in this thesis provides insights to achieve policy targets such as this.

Indonesia, as a member of the Roundtable on Sustainable Palm Oil (RSPO), also seeks to implement environmentally friendly oil-palm development by exploring alternative mitigation strategies including restoration offsetting to compensate for the impact on carbon and biodiversity (RSPO, 2013, 2014). Yet, there is lack of guidance for implementing this policy including the identification of potential areas for restoration offsetting and estimated cost. My thesis helps to identify opportunities and constraints associated with offsetting by the oil-palm industry.

In the forestry sector, the Indonesian Government has been transforming its policies in forest management by shifting from a timber-based approach managed by large corporations towards a multiple-uses approach, to provide broader livelihood options for local communities (President of the Republic of Indonesia, 1999). The Indonesian Ministry of Forestry aimed to designate two million

Chapter 1. Introduction hectares of state forest as community forest by 2014 but only 438,000 hectares had been allocated so far (Sardjono et al., 2013; Ministry of Forestry, 2014). This thesis provides timely guidance for the allocation of community forest in the context of community-based reforestation.

My thesis uses Kalimantan (Indonesian Borneo) as a case study area as it represents a region that is globally important in terms of biodiversity and carbon storage and also has a strong dependency by local people on the ecosystem services provided by forests. Borneo ranks first in terms of plant species richness globally, harbouring more than 14,000 species of plants, of which almost 4,100 are endemic (Kier et al., 2005; Roos et al., 2004). More than 1,600 vertebrate species occur on the island and 25% of them are endemic (Runting et al., 2015). Bornean forests store high amounts of above-ground carbon, averaging 230 MgC per hectare, which is 60% higher than Amazonian forests (Slik et al., 2010). Forest is also the primary resource system for indigenous communities living in the interior of the island, providing a broad range of products including medicinal plants, fruits and vegetables, timber and many more (Abram et al., 2014). Despite these important values, forests in Kalimantan have shrunk by 30% in the past four decades, mainly due to logging, oil-palm plantation and forest fires (Gaveau et al., 2014b).

1.7. Structure of Thesis

This thesis is built from individual papers prepared for publication (Chapters 2-6). To provide context for the thesis, I develop a literature review on the state of Kalimantan’s biodiversity (Chapter 2). In Chapter 3, I assess the utility of a modelling tool to estimate above-ground biomass carbon in Bornean forest. The analytical chapters in this thesis (Chapters 4-6) are aimed at developing decision science framework for landscape-scale restoration that are informed by the principles of conservation planning and grounded in social-ecological theory. In these three chapters, I employ all the steps in decision-making processes, as described above, with the exception of step five (i.e. making decisions by consensus). The general overview of this thesis is shown in Figure 1.2.

Analysis of ecosystem services requires the identification of contextual services pertinent to a region. Through a workshop arranged in Banjarmasin, Kalimantan, I identified that an environmental crisis is occurring in the region and this is resulting in the loss of biodiversity. In Chapter 2, I review the state of biodiversity and the drivers for its loss. I also summarise the value of biodiversity to rural livelihoods, highlighting that biodiversity loss will impact indigenous communities. This chapter provides strong justification that the focus of forest restoration for biodiversity and the provision of livelihoods is relevant in the context of Kalimantan.

Chapter 1. Introduction

Another contextual ecosystem service in Kalimantan to address in this thesis is carbon storage. Yet, there is no practical tool that can be utilised to estimate carbon sequestration during forest restoration. In Chapter 3, I appraise a process-based model, called 3-PG (physiological principles for predicting growth), to estimate above-ground biomass for the major forest types in Borneo. I use publicly available climate and soil data as inputs, as well as parameter values of physiological traits for the Bornean forest. I also test the performance of the model by comparing model outputs with field- measured data.

Restoration in heterogeneous landscapes, such as in degraded tropical forests, delivers benefits that vary spatially and temporally. In chapter 4, I develop a novel framework to cost-effectively allocate restoration investments across forests of varying conditions and forest types in East Kalimantan, accounting for time-dependent restoration outcomes. I develop plausible restoration techniques for each degradation condition, calculate costs and estimate the benefits in term of carbon sequestration and habitat improvement for threatened mammals through time. I contrast scenarios representing the objectives of carbon sequestration and biodiversity to investigate the trade-off between restoration benefits.

In Chapter 5, I investigate the functionality of restoration offsetting as a policy tool to compensate the loss of carbon and biodiversity by oil-palm plantations in Kalimantan. I use a backcasting approach to calculate total carbon emissions, habitat loss of 81 mammals and the loss of native vegetation. I estimate gains from restoration in term of the carbon sequestered, habitat improvement and the re-establishment of native vegetation, accounting for the difference between the initial state and the restored state. I prioritise areas for offsetting to identify cost-effective restoration sites by incorporating costs and gains.

The effectiveness of large-scale restoration programs is strongly influenced by specific socio- ecological and political contexts in a region. In Chapter 6, I integrate theoretic decision approaches with a socio-ecological systems (SES) framework to develop a context-specific restoration plan for livelihoods provision. I employ diagnostic methods to identify a restoration strategy compatible with the social characteristics of the study area. I develop a comprehensive approach using the SES framework to identify variables influencing the ecological suitability, social feasibility and political permissibility of such restoration strategies. I compare the areas prioritised for restoration using my comprehensive approach with those identified on the basis of biophysical criteria alone.

Chapter 1. Introduction

Figure 1.2. Thesis structure.

Chapter 2. Kalimantan’s biodiversity

Chapter 2

State of Kalimantan’s biodiversity

Published as peer-reviewed book chapter:

Chapter 2. Kalimantan’s biodiversity

2. State of Kalimantan’s biodiversity

2.1. Introduction

As one of the global biodiversity hotspots (Whitten et al., 2004) and the third-largest island in the world, Borneo continues to be somewhat of an enigma in the conservation literature and practice. Despite years of research, significant uncertainty remains about the relative contribution of different threats to the species diversity of this Southeast Asian island (Corlett, 2007; Koh et al., 2010; Meijaard et al., 2012). An example of this is the recent discussion in the Indonesian media about how much forest is actually left on the island. In 2011, the Indonesian president made a commitment to maintain at least 45% of the forests of Kalimantan, the Indonesian part of Borneo (Presidential Regulation 3/2012 on Kalimantan Spatial Planning). Some environmental groups responded by saying that only 30% of Kalimantan actually remained covered in forests, and that this was therefore a meaningless commitment (Satriastanti, 2012). On the other hand, the Indonesian Ministry of Forestry reported that 54.9% of Kalimantan was still covered in forest (Ministry of Forestry, 2011). This figure is close to an estimate in a recent peer-reviewed analysis that indicated that 54.4% of forest cover remained (Gaveau et al., 2013; Gaveau et al., 2014b). Some confusion may arise from the unclear definition of what constitutes a forest (Sasaki and Putz, 2009), but more generally there appears to be a lack of scientific and reliable data about how much forest remains, what its ecological status is and how it is being used—for timber or non-timber production (Production Forest), biodiversity and ecosystem conservation (Conservation Forest) or hydrological and soil protection (Protection Forest).

This lack of scientific understanding also extends to the species diversity of Borneo. Despite at least two centuries of taxonomic research on the island, Borneo’s species and their conservation status remain relatively poorly known. Many species records, both plant and animal, are based on just a few observations. In Borneo, 15–35% of the flora may not have been collected (Beaman and Burley, 2003). Estimates suggest that our knowledge of some higher plant families is far from complete— perhaps only 28% complete for the Fabaceae family of legumes in Southeast Asia, for example (Giam et al., 2010). We are seldom sure whether a taxon is rare and localised or simply neglected (Abeli et al., 2009; Cardoso et al., 2011). J.D. Holloway’s 18-volume work describing Borneo’s macro moths is 70% complete; hundreds of new species have been identified so far and many more are likely to be added. In short, we know neither exactly what species diversity exists on Borneo, nor how it is affected by the wide range of threats, such as overharvesting and habitat loss. This makes effective conservation planning difficult. Still, there are broader patterns of biodiversity that we can analyse, and, if recognised early enough and incorporated into overall land use planning, these biodiversity values and associated forest ecosystem services can be protected and further contribute to the

Chapter 2. Kalimantan’s biodiversity wellbeing of the people of Borneo (Abram et al., 2014). Here we summarise the state of knowledge about Borneo’s plant and animal biodiversity, the value of biodiversity to local livelihoods and threats to its survival, and discuss some possible solutions to the ongoing biodiversity crisis.

2.2. Plant diversity

There is no doubt that the island of Borneo is among the richest biodiversity regions in the world. In the early twentieth century, Merrill (1921) estimated that Borneo had 9,000 vascular plant species, and several decades later Ashton (1989) mentioned between 10,000 and 15,000 species. With improved data, updated taxonomy and new methods, these numbers were updated in recent studies by Roos et al. (2004), who estimated 14,423 plant species for Borneo. Based on a global comparative analysis, Kier et al. (2005) concluded that Borneo ranked first in term of plant species richness among terrestrial ecoregions, outperforming the well-known Amazonian plant hotspot of South America (Mittermeier et al., 2005). The island’s exceptional richness is generated by an overall high level of biodiversity on all scales, including site (alpha) diversity, habitat (beta) diversity and landscape (gamma) diversity.

Not only does Borneo have outstanding biodiversity richness, but it is also recognised for its high level of endemism, with 4,089 of its plant species, or some 28% of the total, found nowhere else (Roos et al., 2004). Among 3,000 tree species on the island, 30% are considered endemic, while 40% of Borneo’s 290 palm species are recorded there alone (MacKinnon et al., 1996; Soepadmo and Wong, 1995; Wong, 1998). Borneo is also a centre for orchid richness and endemism with an estimated 1,500–3,000 species, of which more than half are endemic (Chan et al., 1994; Lamb, 1991; Wood and Cribb, 1994)

Among 109 families of tree in Borneo, the Dipterocarpaceae is the most prominent family, not just because of its economic importance as the most widely harvested timber species, but also due to its ecological dominance. Of the 386 described dipterocarp species in the world, 291 (75%) are recorded from Borneo, with 156 being endemic (Soepadmo and Wong, 1995). In term of abundance, the Dipterocarpaceae dominates tree composition, with 21% of inventoried trees belonging to this family, followed by Euphorbiaceae (12.2%), Myrtaceae (5.2%), Sapotaceae (5.0%) and Lauraceae (4.6%) (Slik et al. 2003). At the genus level, Shorea (meranti-merantian) stands out as the most common genus, accounting for 12.3% of trees, followed by Syzygium (5.0%), Diospyros (3.4%), Madhuca (3.2%) and Dipterocarpus (3.1%) (Slik et al., 2003).

Despite the overall high level of flora diversity in a regional context, the patterns of species richness and endemicity in Borneo vary among landscapes. In general, the northeastern part of the island has

Chapter 2. Kalimantan’s biodiversity a larger number of plant species than the southwestern part. The higher biodiversity richness in the northeast is probably driven by the mid-domain effect, meaning that diversity is concentrated in the interior and mountainous areas of the island (Slik et al., 2003). There are some areas that have long been regarded as centres of plant diversity, such as Mount Kinabalu and the Crocker Mountain Range in the Malaysian state of Sabah, and the Meratus Mountains in southeast Borneo (MacKinnon et al., 1996; Slik et al., 2003; Wong, 1998). Contemporary analysis using species distribution models of herbarium records shows that the high mountain peaks of East Kalimantan also have a high level of species richness (Raes et al., 2009). In addition, endemic plants are found in abundance in areas that are biophysically distinct, in terms of altitude, edaphic condition or annual precipitation, for example. This includes the Crocker Mountain Range, the northern parts of the Muller Mountains in central Borneo, the lowland areas east of the Meratus Mountains and the eastern Sangkulirang Peninsula (Raes et al., 2009).

In addition to the regional variation in diversity pointed out above, tree species diversity tends to vary with altitude, with the lowland areas being especially rich in species. This is exemplified by the lowland forests on the mineral soils of northeastern East Kalimantan, central West Kalimantan and northern South Kalimantan, which are all areas of high tree species diversity. Unfortunately, in terms of biodiversity conservation at least, these lowland areas are also important logging concession areas (Slik et al., 2009), where unsustainable timber extraction and resulting forest loss has led to rapid loss of species diversity (Paoli et al., 2010). Not all lowland areas harbour high levels of biodiversity. The peat swamp forests of Central Kalimantan and the heath ecosystem of southern East Kalimantan generally have low species diversity, not just for plants but also for a range of other species groups (Paoli et al., 2010). Nevertheless, the species that are recorded in these areas are often unique and found nowhere else, including Shorea venulosa, S. coriacea and S. materialis, which occur only in heath forest, and S. albida, S. balangeran, S. macrantha, S. platycarpa and S. teysmanniana, which are found in peat swamp forest (Raes et al., 2009; Soerianegara and Lemmens, 1994)

At the site scale, numerous floristic inventories have found that natural ecosystems in Borneo contain extremely high level of alpha diversity. An example from a lowland forest near the Malinau River, East Kalimantan, showed that as many as 759 trees with a diameter at breast height (dbh) of 10 centimetres or more (belonging to 205 species, 110 genera and 47 families) occurred within a 1- hectare sample plot, making this site one of the richest in Indonesia (Sheil et al., 2010). Among them, 77 species were represented by only one individual, 43 species by two individuals, and 28 species by three individuals. Miyamoto et al. (2003) surveyed trees with a dbh of at least 5 centimetres in a heath forest in Kapuas, Central Kalimantan, and recorded 2,016 individuals per hectare belonging to 144 species.

Chapter 2. Kalimantan’s biodiversity

2.3. Animal diversity

Borneo is the most species-rich island in western Indonesia in terms of its absolute vertebrate diversity, although in relative terms (that is, number of species per square kilometre), it is not as rich as the other islands (Table 2.1). Still, it is clear that regionally in Asia and Oceania, the Southeast Asian tropics and particularly the Sundaic islands of Sumatra, Java and Borneo stand out for their species diversity, as exemplified by the diversity of mammal species (Figure 2.1). As with plants, the actual number of animal species is under constant revision, and new species are added on a regular basis. In 2012 alone, researchers described several new species and a new genus of tapeworm from Borneo (Eyring et al., 2012; Schaeffner and Beveridge, 2012a, b, c), as well as four new species of fish (Kottelat, 2012; Kottelat and Hui, 2011; Kottelat and Tan, 2012; Ng and Kottelat, 2012), a new species of mud beetle (Fikacek, 2012) and a new species of frog (Hamidy et al., 2012). Several more new frog species were described for Borneo in 2011 (Matsui, 2011; Shimada et al., 2011), a new species of mammal in 2013 (Sargis et al., 2013) and a new subspecies of mammal in 2011 (Wilting et al., 2011), and a new species of snake in 2008 (Das et al., 2008). Obviously, the data in Table 2.1 should therefore be considered an approximation, with all numbers likely to increase in the near future (assuming the description of new species outpaces the rate at which species become extinct in the wild).

Apart from obvious altitudinal and habitat limits on the distribution of animal species, there are several different zoogeographical divisions on the island of Borneo (MacKinnon et al., 1996). They appear to be determined mainly by geographical barriers such as rivers and mountains. For example, the Barito River separates two species of gibbon (Hylobates albibarbis and H. muelleri), and in the area east of the Barito and south of the Mahakam, there appear to be no orangutans (Pongo pygmaeus). The northern part of the island stands out for its unique fauna, with many endemics being restricted to the high mountains in that part of Borneo, although some of the north Bornean lowland forests also harbour endemic animal species.

Chapter 2. Kalimantan’s biodiversity

Table 2.1. Species counts for the main Indonesian islands (excludes introduced species)

Borneo Sumatra Java Sulawesi No. of No. of No. of No. of No. of No. of No. of No. of species species species species species species species species per square per square per square per square kilometre kilometre kilometre kilometre Freshwater fishes 394 (0.00053) 272 (0.00057) 132 (0.00094) 68 (0.00039) Reptiles 290 (0.00039) 239 (0.00050) 171 (0.00122) 133 (0.00076) Terrestrial mammals 288 (0.00039) 196 (0.00041) 183 (0.00131) 127 (0.00073) Resident birds 420 (0.00057) 465 (0.00098) 340 (0.00243) 240 (0.00137)

Source: Kottelat et al. (1996); MacKinnon et al. (1996); Coates and Bishop (1997); Ahmad and Khairul-Adha (2007); Froese and Pauly (2012)

Animal distribution patterns also appear to be influenced by past and present hunting, at least for those species with financial, nutritional or other value to people, or species that are considered a nuisance because of their impact on crops. Species such as the Sumatran rhinoceros (Dicerorhinus sumatrensis), banteng (Bos javanicus) and crocodiles (Crocodylus spp.) were still widespread in Borneo in the 1930s (Nederlandsch-Indische Vereeniging tot Natuurbescherming, 1939), but unsustainable hunting exacerbated by habitat loss and degradation has reduced the range of these species to a fragment of their former size (Meijaard and Sozer, 1996; Rabinowitz, 1995; Timmins et al., 2012). The influence of hunting in shaping species distributions has also been noted for orangutans (Pongo pygmaeus) (Meijaard et al., 2011; Meijaard et al., 2010b) and other primates (Nijman, 2004), and is likely to be a major factor in determining densities and patterns of absence or presence for many Bornean animal species.

In conclusion, despite many years of scientific research, the fauna of Borneo remains relatively poorly known, especially compared with better-studied tropical areas such as northeast Australia or the Amazon. For only a handful of species are there relatively accurate descriptions of range and some understanding of population trends and threats; these species include orangutans (Wich et al., 2012), most primates (Meijaard and Nijman, 2003), some bat species (Struebig et al., 2010) and some of the small carnivore species. For the remaining many thousand species, we rely on occasional specimens from a few locations to make very rough estimates of their distribution. This lack of basic knowledge makes it very different to strategically plan the conservation of Borneo’s many faunal species, although maximising permanent forest cover is likely to maintain populations of most species.

Chapter 2. Kalimantan’s biodiversity

Figure 2.1. Variation in mammalian species richness and endemism across the Indo-Malayan Region (Boitani and Maiorano, unpublished data).

2.4. The value of biodiversity to local livelihoods

Local people in Kalimantan, especially those living in the interior of the island (often generally referred to as “Dayak”), depend to a large extend on goods and ecosystem services provided by forests (Abram et al., 2014; Meijaard et al., 2013). To fulfil their daily needs for food, fibre and fuel, they have developed a range of lifestyles that depend more or less on forests, varying from a fully forest- based hunter–gatherer lifestyle to semi-intensive agriculture such as shifting cultivation, agroforestry and forest gardens. Many observers, however, from professional foresters to agriculturalists, consider these systems to be rather primitive and unproductive ways of managing lands (Michon et al., 2005). More recently, the green revolution has changed how farmers manage their lands, resulting in a shift from semi-intensive multi-species systems to highly productive monocultures such as rubber and oil- palm plantations and timber estates for pulp production. Within the context of debating optimal land use strategies for Kalimantan’s people, we review the roles of biodiversity and conservation for local livelihoods.

One of the most important elements of biological diversity in Kalimantan is the use of a high variety of medicinal forest plants. An interview-based survey of 1,837 reliable respondents in rural Kalimantan showed that 34% of them used forests as an important source of traditional medicine,

Chapter 2. Kalimantan’s biodiversity with such uses increasing towards the island’s interior (Abram et al., 2014). Medicinal plants were the fourth most commonly mentioned use of forests, after timber (67% of respondents), rattan (52%) and hunting (45%) (Meijaard et al., 2013). Studies found over 250 medicinal plant species from 165 genera and 75 families that were used by a local healer in West Kalimantan (Caniago and Stephen, 1998), and 203 species of plant that were used by the Kenyah people of the Apo Kayan plateau, East Kalimantan (Leaman, 1996). Plants with medicinal purposes are used to treat various ailments, ranging from common illnesses such as fever, headaches and digestive problems to less common ones such as kidney disease and heart problems. Some medicinal plants that are widely used include Gendarussa vulgaris (gandarusa) for the treatment of kidney disease, Alstonia scholaris (pulai) for skin ailments, and Eurycoma longifolia (pasak bumi) as an aphrodisiac and anti-malarial drug. Medicinal plants are mostly found in primary and late successional forests, and some illnesses can be treated by specific plants occurring only in these forest types, such as Mapania cuspidata (serapat), which is used to treat heart disorders.

Forest fruit, either grown in the wild or domesticated, is another commodity of significant importance to communities, both to meet dietary needs and as a source of cash income. In a study conducted in 1991, Siregar (2006) documented at least 130 edible fruit species in Kalimantan, with 91 of them occurring in agroforestry landscapes. The major groups of local fruit species were in the genus Artocarpus with 15 species, Mangifera with 13 species, Garcinia and Baccaurea with 12 species each, and Durio with seven species. Siregar mentioned that 45 of the edible fruit species were sold in local markets, with some having high economic value. These included Artocarpus integer (cempedak), A. heterophyllus (nangka), Mangifera indica (mangga), M. odorata (kuweni), Garcinia mangostana (manggis), Baccaurea motleyana (menteng), Durio zibethinus (durian), D. kutejensis (lai), D. oxleyanus (kerantungan), D. griffithii (Chazdon et al.), D. dulcis (lahung, red durian) and Lansium domesticum (langsat).

Other sources of cash income generated from non-timber forest products (NTFPs) include rattan and honey. Rattan is commonly gathered from wild clumps in secondary forests or harvested from cultivated stems in forest gardens (simpukng). Some rattan species frequently collected for trade are Calamus caesius (rotan sega), C. optimus (rotan taman), C. manan (rotan manau) and C. trachycoleus (rotan jahap) (Meijaard et al., 2014). Several large trees are protected for their importance as habitat of wild bees that produce honey. Called ‘‘tanyut’’ (honey trees) by the Dayaks, these plants include Koompassia malaccensis (kempas), K. excelsa (tualang), Shorea laevis (bengkirai) and Dryobalanops lanceolata (kapur) (Mulyoutami et al., 2009). Some NTFPs are produced from exudates, such as eaglewood (gaharu) yielded by Aquilaria spp, camphor (kapur) extracted from

Chapter 2. Kalimantan’s biodiversity

Dryobalanops aromatica, gutta percha produced by Palaquium spp. and Payena leeri, and damar resin extracted from Shorea spp. and Agathis spp. (Katz, 1997).

Many argue that traditional land use systems such as agroforestry are low in direct economic value, despite their potential benefit for biodiversity and the maintenance of land use cultures. This could be a debatable conclusion. For example, Saragih (2011) calculated the monetary value of forest for forest-edge people in Paser District, East Kalimantan, who generally implement a combination of swidden rice farming, forest gardening, and hunting–gathering in forests. In this area, approximately 19,000 hectares of primary and secondary forest supported the livelihoods of 123 households with 577 inhabitants. Saragih found that the forest contributed almost US$1,000 annually to household income, with one-third of this income generated from NTFPs such as wild honey, rattan and fruits. With an average GDP per capita (excluding income from oil and gas) in East Kalimantan of only $1,980 in 2011 (Indonesian Bureau of Statistics, 2012), the income generated by NTFPs clearly made a significant contribution to the livelihoods of rural people.

One could probably argue that the traditional ways of managing forests are not the best form of land management, considering the low economic return. However, there are limited alternative livelihoods for forest-edge people, owing to limited knowledge and lack of access to capital, so subsistence use of forest is the only available option (Saragih, 2011). Also, this analysis calculates only the direct value of forest and does not account for non-use values such as water regulation, flood prevention and carbon retention.

2.5. Threats to biodiversity in Kalimantan

The Sundaland ecoregion, where Borneo island is located, has been identified as a biodiversity hotspot because its rich and unique biodiversity is under a high level of threat from humans (Myers et al., 2000). Sodhi et al. (2004) reviewed key threats to biodiversity in Southeast Asia, including deforestation, forest fire, hunting for bushmeat, and wildlife trade. At a national scale, the major threatening processes for Indonesian flora are habitat loss caused by logging and infrastructure development, small population size, restricted range and overexploitation (Budiharta et al., 2011).

We evaluated threats to biodiversity in Kalimantan using data extracted from the Red List of Threatened Species compiled by the International Union for Conservation of Nature (IUCN, 2012). We found 251 threatened species in Kalimantan, in the categories ‘least concern’, ‘near threatened’, ‘vulnerable’, ‘endangered’, ‘critically endangered’ or ‘extinct in the wild’. Data on the threats or threatening processes were available for 176 of those species (67 animals and 109 plants). Six broad

Chapter 2. Kalimantan’s biodiversity threatening processes were identified, namely habitat loss (including habitat degradation), overexploitation, pollution, restricted range, small population size and predation (Figure 2.2).

100

Animal 80 Plant

40 Percentage

Threats

Figure 2.2. Processes that threaten endangered species in Kalimantan (IUCN, 2012).

Habitat loss was the major threatening process, affecting 82.1% of listed animal species and 60.5% of listed plant species (Figure 2.2). Kalimantan lost 10 million hectares of forest cover between 1985 and 1997 (World Resources Institute, 2002) and 3 million hectares between 1996 and 2002, much of it within protected areas (Fuller et al., 2004). Around 1.4 million hectares was cleared between 2000 and 2010 (Gaveau et al., 2013). Forest loss is especially deleterious because it occurs mostly in areas known as centres of biological diversity and endemism, as mapped by Raes et al. (2009). While logging was the key driver of habitat loss in the past, forest conversion to monoculture plantations and mining areas now seems to be the primary cause of deforestation. Carlson et al. (2013) calculated that between 1990 and 2010, approximately 90% of oil-palm plantations in Kalimantan were established in previously forested areas, with expansion rates exceeding 210% annually (equal to 232,800 hectares per year). If this trend continued, they predicted that by 2020 oil palm plantations would occupy almost one-third of Kalimantan’s lowland areas outside protected zones. The loss or reduction of forest cover impacts severely on sensitive interior species such as Arctitis binturong (binturong), Elephas maximus (Asian pygmy elephant), Pongo pygmaeus (orangutan) and Neofelis diardi (Sunda clouded leopard) (Corlett, 2007; Meijaard et al., 2005; Meijaard et al., 2012)

Chapter 2. Kalimantan’s biodiversity

Overexploitation also contributed to biodiversity degradation in Kalimantan, threatening 55.2% of animal and 24.8% of plant species. Animals are hunted for bushmeat, such as Ratufa affinis (cream- coloured giant squirrel), Sus barbatus (bearded pig), Cervus unicolor (sambar deer), Pongo pygmaeus (orangutan) and Chelonia mydas (green turtle); for their skins, including Pardofelis marmorata (marbled cat) and Pardofelis badia (bay cat); for the pet trade, such as Hylobates muelleri (Bornean gibbon) and Trichopodus leerii (pearl gourami); and for medicinal uses, such as Presbytis hosei (grey leaf monkey) and P. frontata (white-faced langur), which are killed for the ‘bezoar stones’ in their bladders. As mentioned, overexploitation was also responsible for the local extinction of Dicerorhinus sumatrensis (Sumatran rhinoceros) in Kalimantan over 40 years ago (Michon et al., 2005), although the recent rediscovery of the species gives a glimmer of hope that it can still be saved from extinction (Meijaard and Nijman, 2014). More recently, demand, primarily from mainland Asia, for species such as Manis javanica (pangolin), Gekko gecko (the once common tokek) and a range of hornbill species has decimated populations throughout Kalimantan. Similarly, overharvesting is the primary threatening process facing highly commercial plants such as Aquilaria spp. (eaglewood), Shorea spp. (meranti) and Dipterocarpus spp. (keruing). The increased demand for gaharu in the 1970s drove an ‘eaglewood rush’ in Kalimantan, leading to depleted populations by the 1990s, as indicated by much longer collecting periods (up to a month) than in the past (just a week) (Katz, 1997). Such population shortages have driven up prices, leading to even more rampant collection practices, especially by outsiders, who tend to ignore all sustainability standards.

The impact of pollution and predation is apparent only for a number of animal species, although it is likely to affect plant life and other species as well. Pollution threatens freshwater animals living in Kalimantan’s major rivers, including Acrochordonichthys chamaeleon (lakut) and Pseudomystus myersi (ikan pisang), which suffer from mercury poisoning caused by illegal gold mining in the Kapuas River, and Chendol lubricus, which is affected by soil sedimentation and pesticides associated with oil-palm plantation development along the Mahakam River (Jenkins et al., 2009a, b, c).

Intrinsically biological factors cause two other threatening processes: restricted species range, which affects 17.9% of animal and 32.1% of plant species; and small population size, affecting 1.5% of animal and 8.3% of plant species. The relatively high proportion for restricted range is attributable to the high level of endemism in Kalimantan. A well-known example of an endemic animal is Pongo pygmaeus (orangutan), which is found only in Borneo. Crocodylus siamensis (Siamese crocodile), recorded only in Mesangat Lake in the Mahakam River system, is threatened by its very small population size, with only 30 individuals estimated to remain. Plant species threatened by a restricted range include Mangifera casturi (kasturi), which is found only in cultivated areas in South

Chapter 2. Kalimantan’s biodiversity

Kalimantan, some species of Nepenthes (pitcher plants/kantung semar), for example, N. boschiana, N. clipeata, N. ephippiata, and endemic species of Knema spp. and Horsfieldia spp. We further note that the threat posed by small population size in Kalimantan may be underestimated because information is available only for a small number of species. The IUCN guidelines for population size based on the number of mature individuals use an equation for which the following variables need to be known: population density, range area and the proportion of individuals that are mature. Such information is available only for a few species—for most, we simply have no idea of population size.

2.6. Conservation strategies

It is not a trivial task to integrate conservation into regional development, considering the contrasting perspectives of proponents of economic growth and supporters of resource conservation. Conservationists often view economic growth in developing countries as the primary driver of the depletion of natural capital and ecosystem services, while many others view conservation as a hindrance to development. The emerging concept of a green economy in which conservation takes place in multi-functional landscapes provides the opportunity to harmonise these two points of view. Underpinning this concept is the idea that new approaches to biodiversity are possible beyond the strict protection of biodiversity in protected areas—that landscapes can be managed for productive purposes while minimising the impact on biodiversity.

Certainly, a major strategy for biodiversity conservation should be to maintain areas identified as centres of biological diversity and culturally important forests, and incorporate them into protected area networks. Protected areas in Kalimantan now cover 11.1 million hectares (21% of the total land area), in the form of Conservation Forest (nature reserves, wildlife sanctuaries, national parks) and Protection Forest (Ministry of Forestry, 2011). There is evidence, however, that some protected areas (for example, Gunung Palung National Park and Kutai National Park) are not well managed, with illegal logging and mining, land encroachment and forest fires reducing the effective area of protection by more than half (Curran et al., 2004; Fuller et al., 2004). Other nature reserves and wildlife reserves (such as Muara Kendawangan and Muara Kaman) are in even worse shape, with very little natural forest cover remaining. In 2007, the three countries that share responsibility for the island, Indonesia, Brunei Darussalam and Malaysia, signed a declaration launching Heart of Borneo, an ambitious conservation program to protect and sustainably manage vast areas of intact, connected forest in mountainous areas in the central part of Borneo. If it is to achieve its goals, this initiative will need to address the underlying causes of deforestation and forest degradation, learning from the experience of continuing forest loss in many protected areas in lowland Kalimantan.

Chapter 2. Kalimantan’s biodiversity

One way to sidestep the ineffectiveness of the protected area system and the high opportunity cost of maintaining intact forest is conservation in multi-functional landscapes. Biodiversity conservation in well-managed logging concessions and multi-species agroforestry are examples of such an approach. Currently, 24.8 million hectares of land in Kalimantan (46.6% of the land area) is designated as production forest, with 10.6 million hectares under active logging concessions (Gaveau et al., 2013). This vast forest area could contribute significantly to the persistence of wildlife if managed properly through Reduced-Impact Logging (RIL), the retention of key species (for example, Ficus spp. (fig), fruit trees and trees with hollows) and the setting aside and management of high conservation value forest (Meijaard et al., 2005). Any regime for selective logging in production forest should also include the preservation of forest corridors between protected areas (Proctor et al., 2011). A recent analysis indicates that about 78% of the range of the Bornean orangutan falls outside protected areas, most notably in logging concessions (29% of the range) (Wich et al., 2012). With most orangutans found outside protected areas and only a minimal chance that much more habitat will be formally protected, the inevitable conclusion is that the conservation of orangutans will have to take place not only in protected areas, but also in forests that are used for other purposes (Meijaard et al., 2012).

Considering the high level of threat caused by overexploitation, another important strategy is to control wildlife hunting and gathering. This does not necessarily mean banning the practices altogether, but rather the careful management of populations of interest through improved harvesting practices (for example, minimum allowable sizes) and regulated trade (Budiharta et al., 2011). Although Government Regulation 8/1999 on Wild Flora and Fauna Exploitation is supposed to regulate the use of wildlife, abuse of its provisions is widespread, as indicated by reports of a number of cases of illegal trade in endangered fauna such as Buceros vigil (helmeted hornbill), Manis javanica (Sunda pangolin) and Helarctos malayanus (Malayan sun bear) (Fachrizal and Pahlevi, 2013). A range of measures is needed to increase the effectiveness of the relevant laws, including more accountable government, more effective law enforcement and a strengthened judiciary, as well as increased public awareness. To reduce the incentive to hunt wild animals for bushmeat, local communities should be taught how to raise livestock, which would provide an alternative source of animal-based protein. This would not only benefit biodiversity, but would also help to improve local economies.

Prudent land use planning occupies a key role in the integration of conservation and development goals. There will always be trade-offs between the costs and benefits of particular land use decisions; allocating intact or lightly degraded forest to mining development inevitably affects biodiversity, for example, and providing oil-palm licences in areas that contain orangutans may trigger conflict between the animals and plantation workers. The use of a science-based optimisation approach should

Chapter 2. Kalimantan’s biodiversity make it easier to maximise the benefits, and minimise the costs, of land use decisions. For instance, the allocation of logging concessions could be tied to measures to minimise the impact on biodiversity (Wilson et al., 2010), and oil-palm plantations and logging concessions could be developed in ways that contribute to the conservation of forest carbon and minimise carbon emissions (Venter et al., 2013). One prerequisite is to collate reliable and relevant data, especially in a spatial context, that can be used to explore potential scenarios.

Well-managed forests, and even plantations in which some natural forest is retained, can provide habitat for Kalimantan’s orangutans and other threatened species, although much more research is needed to determine the exact factors that contribute to the survival of such species in multi-functional landscapes. For such landscapes to provide viable habitat for biodiversity as well as people, a significant shift in perspective is needed among conservation groups, governments, forest managers and local communities. We need to stop seeing conservation in black-and-white terms of unprotected versus protected areas, or natural versus unnatural landscapes. We need to acknowledge that the future of Kalimantan’s wildlife depends on the survival of species in human-made landscapes, not just in pristine habitats.

2.7. Acknowledgements

We thank two anonymous reviewers and Elizabeth Thomson for editing the manuscript. We also acknowledge the inputs from Budy Resosudarmo, Lydia Napitupulu and Muhammad Handry Imansyah. The book is initiated by Indonesian Regional Science Association (IRSA) with supports from United States Agency for International Development (USAID), Support for Economic Analysis Development in Indonesia (SEADI), Indonesia Project of Australian National University (ANU) and Bank Kalsel.

Chapter 3. Estimating AGB Bornean forest

Chapter 3

Estimating the above-ground biomass of Bornean forest

Published as:

Budiharta, S., Slik, F., Raes, N., Meijaard, E., Erskine, P., Wilson, K.A. 2014. Estimating the above- ground biomass of Bornean forest. Biotropica 46, 507-511.

Chapter 3. Estimating AGB Bornean forest

3. Estimating the above-ground biomass of Bornean forest

3.1. Abstract

We apply a process-based model, called 3-PG (physiological principles for predicting growth), to estimate above-ground biomass for the primary forests of Borneo. Using publicly available soil and climate data, and parameterised with physiological traits of Bornean forest, the modelled above- ground biomass and basal area showed statistically significant relationships with field-measured data from 85 sites across four major forest types. Our results highlight the possibility to expand the application of 3-PG to forests of varying condition, which would facilitate inclusion of modelled forest biomass data for developing a Tier 3 carbon inventory system for Borneo.

3.2. Introduction

The Kyoto Protocol under the United Nations Framework Convention on Climate Change (UNFCCC) requires every signatory nation to standardise their carbon accounting systems (United Nations, 1998). All 44 Annex 1 countries have submitted a National Inventory Report which contains a toolbox of methodologies for calculating sources and sinks of carbon (UNFCCC, 2013). Non-Annex 1 countries are requested to submit a voluntary National Communication biennially, which outlines progress with the development of a carbon inventory system. Since 1999, only four non-Annex 1 countries (of 154 countries) have updated the third National Communication and only Mexico has completed the fourth and fifth reports (UNFCCC, 2013).

Previous attempts to measure carbon dynamics associated with the tropical forests of Southeast Asia have mostly focused on regional quantification of stocks and losses generalised across forest types (e.g. Berry et al., 2010; Carlson et al., 2013; Gibbs et al., 2007; Harris et al., 2012; Saatchi et al., 2011) which could potentially fail to capture carbon gains at a local scale and thereby lead to the overestimation of net carbon emissions. These approaches would be classified as Tier 1 and Tier 2 due to the lower resolution of information associated with the stock-change methods employed (IPCC, 2006). Tier 3 approaches require detailed forest inventory data across a variety of forest types, complemented with region-specific process-based model(s), which have been field validated (IPCC, 2006). To our knowledge, there have been no Tier 3 approaches trialled in Southeast Asia.

This study is the first application of a process-based model, 3-PG (physiological principles for predicting growth), to predict above-ground biomass (AGB) dynamics of the primary forests of Borneo. The 3-PG model estimates stand development based on physiological processes that are simplifications of plant-environment interactions (Landsberg and Waring, 1997). This model has

Chapter 3. Estimating AGB Bornean forest been used for single-species plantations in temperate and subtropical regions (Landsberg and Sands, 2011) and has resulted in accurate estimates of the growth of highly diverse forests in the Australian wet tropics and Amazon (Nightingale et al., 2008; White et al., 2006). We aim to apply the 3-PG model to obtain baseline estimates of the upper limits of AGB accumulation for the tropical forests of Borneo. We evaluate the performance of the model, and appraise the utility of the approach for developing a Tier 3 carbon accounting system for Borneo.

3.3. Methods

The 3-PG is a process-based model, which calculates forest productivity from absorbed photosynthetically active radiation and canopy quantum efficiency, constrained by atmospheric vapour pressure deficit, soil characteristics, and temperature (Landsberg and Waring, 1997). The model consists of five biological sub-models for estimating biomass production, biomass allocation, stem stocking and mortality, soil and water balance, and stand management (Appendix S3.1; Nightingale, 2005). A detailed description of the model is presented by Landsberg and Sands (2011). The inputs into the 3-PG model include soil and climate data, and field-measured data are required for model validation. Estimates of variables pertaining to forest growth (including biomass and stand basal area) are predicted on a monthly and annual basis for up to 120 years.

Empirical data on the biomass and basal area of primary forest was obtained from 85 sites across Borneo (Slik et al., 2010 and additional sources: Appendix S3.2), along with the geographical coordinates and elevation for each site. The site data were differentiated into four major forest types: lowland forest (n=62), montane forest (n=7), heath forest (n=9), and peat swamp forest (n=7) (Wikramanayake et al. 2002, Raes 2009). Slik et al. (2010) found no spatial autocorrelation among the same sampled sites.

Data on soil properties were generated from the Harmonized World Soil Database, including soil texture class, maximum and minimum plant available soil water, and fertility ratings (FAO/IIASA/ISRIC/ISSCAS/JRC, 2012). Soil texture was categorised into four soil classes: clay; clay loam; sandy loam; and sand. Plant available soil water was obtained from the available water storage capacity, and classified into seven classes, ranging from 0 to 150 mm/m. The fertility rating was assigned based on the percentage of organic carbon, which was standardised between zero and one. We assume that the modus organic carbon value (equating to 1%) associated with the mineral soils of tropical regions corresponds to a soil fertility rating of 0.2 (Nightingale et al., 2008). Considering the high organic content but low available nutrients of peat soil, a uniform fertility rating of 0.19 was assigned, which is similar to the average rating attributed to the similarly infertile soil of heath forest (Cannon and Leighton, 2004). 32

Chapter 3. Estimating AGB Bornean forest

Four climate input variables, including solar radiation, temperature, precipitation and vapour pressure deficit were obtained from a number of sources. Daily solar radiation data (MJ/m/d) was acquired from the monthly averages of earth surface insolation incidence from 22 years of satellite observation (NASA, 2013). Monthly maximum, mean and minimum temperature (°C), as well as monthly precipitation (mm/month) were derived from the WorldClim database (Hijmans et al., 2005). A comprehensive vapour pressure deficit dataset does not exist for Borneo, therefore the inputs for this variable (mBar) were generated from the WorldClim database and computed as 0.62 times the difference between the saturated vapour pressure at the maximum and minimum temperatures (Waring, 2013).

We estimated stand development starting from the commencement of forest succession, assuming that there are tree seedlings present and vegetation regenerates passively (White et al., 2000; White et al., 2006). The 3-PG model was originally developed for monoculture plantations with parameters describing allometric relationships, canopy conductance and canopy structure assigned in the context of a single species. As our sample sites comprise a diversity of species, we assigned parameters specific to tropical forest based on studies conducted on the island of Borneo or elsewhere in the tropics and used default values where more detailed data could not be located (Table A3.1). The outputs are highly sensitive to the canopy quantum efficiency parameter (Nightingale et al., 2008) and therefore a range of values were employed (Table A3.1), representing the three groups of plant functional traits (i.e. fast growing pioneer, fast growing dipterocarp and slow growing dipterocarp) and three canopy layers (i.e. understorey, main canopy and emergent) that are representative of Bornean forest (Eschenbach et al., 1998; Huth and Ditzer, 2000). We determined the optimum temperature for growth for each sampled site using climate data following Nightingale et al. (2008) with formulation adapted from Waring (2013) (Equation 3.1).

푇표푝푡 = [(푇푚푎푥 – 푇푚푖푛) × 0.6 ] + 푇푚푖푛 (Equation 3.1)

We iteratively ran the model and subsequently matched the modelled outputs to field-measured data to optimise the value of canopy quantum efficiency for each site (Landsberg et al., 2003; Nightingale et al., 2008). In doing this, we would need at least two variables to validate the parameter for canopy quantum efficiency for each site in which we used AGB and basal area. The input for canopy quantum efficiency was then the value within the published range for Bornean forest (i.e. 0.023-0.043) that give the optimum output for the two variables. We employed regression analysis to assess the relationships between the modelled and field-measured AGB using R Statistical Software (R v. 3.0.1, R Development Core Team, 2013).

Chapter 3. Estimating AGB Bornean forest

3.4. Results

The 3-PG model accurately predicted AGB compared with field-measured data of 85 sites (R2 = 0.958; P < 0.001) with standard error (SE) of 25.9 Mg dry mass per hectare (Figure 3.1) and also provided strong predictions of basal area (R2 = 0.774; P < 0.001; SE = 3.37 m2 per hectare). Higher standard error was observed at sampled sites with the lowest and highest basal area, which is likely the result of inaccurate allometric parameterisation (Figure 3.1).

Figure 3.1. Comparison between the 3-PG modelled (over 100 years) and field-measured data of AGB (Mg per hectare) and basal area (m2 per hectare) across Borneo. The straight line indicates 1:1 line.

AGB is not uniformly distributed across forest types, with lowland forest having the highest stock on average, reaching 477 Mg per hectare, and heath forest predicted to have approximately 70% of this stock (Table A3.2). In general, 3-PG was able to estimate the AGB of the four major forest types (R2 > 0.8), with the model tending toward overestimation (Table A3.2 and Figure A3.1).

The modelled forest growth follows a logistic curve with the rate of predicted AGB accumulation greatest in the first 20-30 years and reaching a steady state at around 60-70 years (Figure 3.2). On average, by 20 years, Bornean forest could accumulate 156 Mg per hectare with the greatest variation of growth exhibited by montane forest, which is driven by the high climate variability (i.e. monthly temperature and precipitation) across the sampled sites for this forest type.

Chapter 3. Estimating AGB Bornean forest

a b

c d

Figure 3.2. Predicted AGB (Mg per hectare) accumulation over 100 years for each forest type: (a) lowland forest; (b) montane forest; (c) heath forest; and (d) peat swamp forest. Solid line is the average AGB modelled for each forest type; dashed lines are the upper and lower confidence intervals.

3.5. Discussion

Our study is the first application of 3-PG to multiple forest types on the island of Borneo, representing further expansion of 3-PG to multi-species and multi-age vegetation tropical forests (Nightingale et al., 2008; White et al., 2006). Despite its simplicity, the 3-PG model accurately predicted forest growth, particularly AGB (Figure 3.1 and Figure A3.1). The modelled AGB over 100 years of simulation approximates the upper limits of AGB accumulation. Forests continue to accumulate AGB beyond this period although the net accumulation is minimal due to stand mortality and decelerating primary productivity after reaching a mature successional state (Brown and Lugo, 1990; Guariguata and Ostertag, 2001). It is therefore likely that AGB is overestimated in our predictions. 35

Chapter 3. Estimating AGB Bornean forest

Using a statistical model, Slik et al. (2010) found that annual precipitation, soil fertility and soil drainage determine the distribution of AGB across Borneo. As would be anticipated, we find that sites occurring on fertile soil (e.g. in lowland and montane forests) accumulate higher AGB than those on poor soil (e.g. in heath and peat swamp forests). Aside from environmental factors, vegetation intrinsic traits also affect AGB accumulation. The canopy quantum efficiency resembles the rate of photosynthesis of a plant as a response to absorbed light by the canopy (Eschenbach et al., 1998). The range for this parameter used in this study (i.e. 0.023-0.043 mol C/mol photons) is higher than that used for application of 3-PG to the forests of tropical Australia (i.e. 0.013-0.0175 mol C/mol photons), resulting in higher average AGB (i.e. 382 Mg per hectare for Borneo compared with 257 Mg per hectare for the tropical forests in Australia) (Nightingale et al., 2008). Using 3-PG model, the accumulation of AGB in the Brazilian Amazon after 20 years has been predicted to be approximately 102 Mg per hectare (White et al., 2006). The higher value of modelled AGB of Bornean forest in this study (i.e. 156 Mg per hectare after 20 years), reflects the overall higher biomass carrying capacity of Bornean forests (Slik et al., 2010).

The primary advantage of 3-PG compared with other tropical forest growth models is the simplification of physiological processes into mathematical equations without necessarily requiring a large amount of environmental data and input parameters (Landsberg and Waring, 1997; Nightingale et al., 2004). This is particularly relevant to data-poor regions such as Borneo where long-term and comprehensive climate data are generally lacking. We show that even using publicly available climate and soil data, 3-PG performs well in predicting AGB in highly diverse Bornean forest.

A technical limitation of the 3-PG model is a lack of transparent and objective means to assign the soil fertility rating (Landsberg and Sands, 2011). In many applications of 3-PG, the fertility rating is determined arbitrarily through expert judgement (Nightingale et al., 2008; White et al., 2000; White et al., 2006). In this study, we attempted to reduce this subjectivity by using organic carbon as a surrogate, although this approach does not account for the influence of the availability of phosphorus, potassium, and magnesium or soil acidity on the fertility of Bornean soils (Paoli et al., 2008). The limited amount of available inventory data for primary forests to validate the model, in addition to the variation in sample sizes across forest types, impacts the reliability of the predictions we present. Furthermore, the lack of inventory data precludes the systematic analysis of forest biomass across a variety of disturbance histories (e.g. logged and burnt forest) and degradation conditions, which is required to deliver a comprehensive Tier 3 carbon accounting systems (IPCC, 2006).

Chapter 3. Estimating AGB Bornean forest

We have demonstrated the application of a simple, yet robust, process-based model to estimate forest biomass, revealing the potential to accurately predict the upper limits of biomass accumulation for the major forest types occurring on the island of Borneo. Our research is the first step towards the integration of a process-based model into a Tier 3 carbon accounting system for the three nations of Borneo, which would support the implementation of climate policies and associated REDD+ projects. In the future, a standardised forest inventory methodology for Borneo with sampling distributed across forest types and within stands of varying forest condition would improve the accuracy and utility of the biomass predictions presented here.

3.6. Acknowledgements

SB was supported by an Australian Awards Scholarship and Australian Research Council Centre of Excellence for Environmental Decisions, NR by the Netherlands Research Council NWO-ALW with grant 819.01.014 and KAW by an Australian Research Council Future Fellowship. We thank Luke Shoo for graphical preparation.

Chapter 4. Restoring heterogeneous degraded tropical forests

Chapter 4

Restoring degraded tropical forests for carbon and biodiversity

Published as:

Budiharta, S., Meijaard, E., Erskine, P., Rondinini, C., Pacifici, M., Wilson, K.A. 2014. Restoring degraded tropical forests for carbon and biodiversity. Environmental Research Letters 9, 114020.

Chapter 4. Restoring heterogeneous degraded tropical forests

4. Restoring degraded tropical forests for carbon and biodiversity

4.1. Abstract

The extensive deforestation and degradation of tropical forests is a significant contributor to the loss of biodiversity and to global warming. Restoration can potentially mitigate the impact of deforestation, yet knowledge on how to efficiently allocate funding for restoration is still in its infancy. We have systematically prioritised investments in restoration in the tropical landscape of East Kalimantan, Indonesia, and through this application demonstrated the capacity to account for a diverse suite of restoration techniques and forests of varying condition. We developed a map of forest degradation for the region, characterised on the basis of above-ground biomass and differentiated by broad forest types. We estimated the costs of restoration as well as the benefits in terms of carbon sequestration and improving the suitability of habitat for threatened mammals through time. When the objective was solely to enhance carbon stocks, the restoration of highly degraded lowland forest was the most cost-effective activity. If the objective was to improve the habitat of threatened species, multiple forest types needed to be restored, reducing the accumulated carbon by up to 24%. Our analysis framework provides a transparent method for prioritising where and how restoration should occur in heterogeneous landscapes in order to maximise the benefits for carbon and biodiversity.

4.2. Introduction

Deforestation and forest degradation are among the major drivers of biodiversity loss and carbon emissions in the tropics (Harris et al., 2012; Sodhi et al., 2004). Despite the rate of tropical deforestation declining in the 1990s (Achard et al., 2002; DeFries et al., 2002), this trend has since reversed, with an increasing rate of annual forest loss in some countries, notably Indonesia (Hansen et al., 2013). While forest degradation has a broad definition, it is often associated with a reduction in forest biomass (Sasaki et al., 2011; Sasaki and Putz, 2009; Thompson et al., 2013). Forest degradation is extensive, totalling 500-600 million hectares (or 30–40%) of the forest area in the tropics (Blaser et al., 2011). In Indonesia alone, 82.9 million hectares (equating to 63.1%) of the Forest Estate that is under the authority of the Indonesian Ministry of Forestry is either degraded or deforested (Ministry of Forestry, 2013a). Approximately 20 million hectares of this degraded forest is considered idle without a concession and hence targeted for illegal logging, land grabbing or gazetted for the development of oil-palm plantations (Boer, 2012; Carlson et al., 2013; Gaveau et al., 2013; Levang et al., 2012).

Indonesia, and other tropical countries, can obtain funding to restore forests through payments for ecosystem services under a variety of programs including REDD+ projects (Reducing Emissions 39

Chapter 4. Restoring heterogeneous degraded tropical forests from Deforestation and forest Degradation) (UNFCCC, 2007). Globally, the REDD+ program has generated carbon investments of up to US$2 billion since 2009 (Forest Trends, 2014; Peters-Stanley et al., 2013), in addition to a US$1 billion climate agreement between Norway and Indonesia (Solheim and Natalegawa, 2010). Forest restoration is potentially more advantageous than other mechanisms (e.g. a focus on avoided deforestation alone) as it can engage and employ local people to undertake and maintain plantings (Alexander et al., 2011). There is also a long history of experience with forest rehabilitation initiatives in Indonesia conducted by government and non-governmental organisations, local communities and the private sector, which could inform the implementation of restoration programs (Barr and Sayer, 2012; Nawir et al., 2007a).

The available funding and institutional resources however are unlikely to be sufficient to restore all degraded forest and priority must be given to particular areas (Boer, 2012; Brockhaus et al., 2012; Nawir et al., 2007a). The selection of areas to restore is complicated by variation in the capacity of forest types to sequester and store carbon due to inherent ecological characteristics and influenced also by climatic and edaphic gradients (Budiharta et al., 2014b; Slik et al., 2010) and external pressures (e.g. logging, fire and smallholder agriculture) (de Jong et al., 2001; Gaveau et al., 2013; Langner et al., 2007). These interacting processes result in a broad spectrum of forest condition states, ranging from completely deforested to old-growth forest with minimal disturbance (Sasaki et al., 2011; Thompson et al., 2013). The condition of forests also influences the suitability of habitat for wildlife with some highly sensitive species requiring pristine forest and other species being more adaptable and able to persist in degraded forest (Ansell et al., 2011; Meijaard et al., 2005).

Forests of varying condition necessitate the consideration of a diverse suite of restoration techniques (Mori, 2001; Sasaki et al., 2011), which incur different costs associated with implementation and lost opportunities including the costs associated with foregone profits of alternative land use or management (Lamb et al., 2005; Wilson et al., 2012). For example, restoring a completely deforested site, such as an Imperata cylindrica grassland, would involve intensive planting with a high implementation cost (Otsamo et al., 1996) but the opportunity cost for the logging industry would be low due to a limited volume of harvestable timber remaining. In contrast, some degraded sites with considerable extant vegetation cover would require only enrichment planting, such as gap planting and strip planting, requiring a lower implementation cost (Korpelainen et al., 1995; Maswar et al., 2001; Soekotjo, 2009). Restoration in this later case would likely forgo the potential revenues from harvesting remaining timber and thereby incur a higher opportunity cost (Ruslandi et al., 2011; van Gardingen et al., 2003).

Chapter 4. Restoring heterogeneous degraded tropical forests

Each restoration technique makes a different contribution to forest growth, with the amount of accumulated carbon in the form of living biomass being related to planting density and design, often in a non-linear way (Korpelainen et al., 1995; Maswar et al., 2001). The sequestration rates following planting also change through time. For instance, forest growth from strip plantings in a logged dipterocarp lowland forest in Borneo follows a sigmoid curve with rapid carbon sequestration during the first 30 years (Hardiansyah, 2011). Incorporating the time-dependent relationship of carbon sequestration and the costs associated with restoration allows the cost efficiency of each restoration action through time to be determined (Wilson et al., 2011). The cost efficiency of restoration is particularly relevant due to the finite amount of funding available for restoration programs (Barr et al., 2010; Nawir et al., 2007a; Peters-Stanley et al., 2013).

Only a handful studies concerning the systematic prioritisation of restoration on degraded tropical forests address the heterogeneity of these landscapes (e.g. Goldstein et al., 2008; Marjokorpi and Otsamo, 2006) but have not explicitly accounted for carbon accumulation as a restoration objective. While more recent studies investigate restoration within the framework of carbon sequestration, they do not utilise systematic decision theoretic approaches (e.g. Gilroy et al., 2014; Sasaki et al., 2011). Analyses focused on prioritising restoration for carbon sequestration have been in the context of cleared agricultural lands in non-tropical regions (e.g. Crossman et al., 2011; Renwick et al., 2014) and have focused on only a single restoration technique (i.e. high density planting). The trade-off between carbon sequestration and biodiversity goals also warrants investigation given emerging initiatives to restore the habitat of threatened species (e.g. the Indonesian Orangutan Habitat Restoration Project) and degraded ecosystems (e.g. Katingan Peat Forest Restoration Project) (Rahmawati, 2013; Starling Resources, 2014). It is therefore timely to investigate restoration prioritisation approaches that can facilitate the identification of solutions for restoration that aim to satisfy the dual objectives of carbon sequestration and biodiversity conservation.

Here, we develop and apply a new framework for prioritising restoration that accounts not only for deforested areas but also forest areas of varying condition. The framework is applied to East Kalimantan, Indonesia, to inform REDD+ policy implementation and conservation initiatives in the region. It provides the first assessment of the extent of degraded forests in East Kalimantan and their distribution across forest types. Alternative objectives for restoration are explored and trade-offs between carbon sequestration and improving the suitability of habitat for threatened mammals investigated. Finally, we provide recommendations of where and how restoration should be implemented in order to maximise the accumulation of carbon and improve the habitat for threatened species.

Chapter 4. Restoring heterogeneous degraded tropical forests

4.3. Methods

4.3.1. Study area

East Kalimantan is an Indonesian province of the island of Borneo (Figure 4.1). East Kalimantan, and Borneo in general, is of global importance for both carbon storage and biodiversity. The above- ground biomass (AGB) carrying capacity per unit area of Bornean forests is 60% higher than Amazonian forests (Slik et al., 2010). Borneo is also the major evolutionary source of biodiversity in Southeast Asia, resulting in extremely high species richness (de Bruyn et al., 2014). Despite their vital role in the global carbon cycle and biodiversity conservation, the combination of logging, oil- palm development, mineral extraction, and forest fires threatens biodiversity in the region and is the primary source of carbon dioxide emissions for the region (Budiharta and Meijaard, in press; Carlson et al., 2013; Pearson et al., 2014).

Jurisdictionally, land use in East Kalimantan is categorised into two classes: Forest Estate (Kawasan Hutan/KH) and non-Forest Estate (Area Penggunaan Lain/APL). Forest Estate, either forested or deforested, is defined as land use designated by the government to be permanent forest and under authority of the Ministry of Forestry (i.e. Production Forest, including logging concession and industrial timber plantation; Protection Forest; and Conservation Forest such as national park). Non- Forest Estate is land outside the Forest Estate and includes both forested lands (e.g. private forest/forest garden) and non-forested lands (e.g. settled areas, road network and agricultural lands).

In this study, we were interested in forest landscape restoration using native species and therefore focused only on the Forest Estate that had native vegetation cover. We excluded the non-Forest Estate and Forest Estate allocated for industrial timber plantation (ITP), which primarily used a single exotic species (e.g. Acacia mangium). Within this part of the Forest Estate, delineated based on Provincial Spatial Planning and Forest Land Use by Consensus (Ministry of Forestry, 2011, 2012a), we created a planning unit layer by dividing the study area into pixels with a resolution of 400 hectares. This resolution represented a compromise to reduce computational time, capture environmental variability, and informative for policy and management implementation given the overall extent of the study region. Our study area covered 12.3 million hectares (63.2%) of the total land area of East Kalimantan (Figure 4.1). All spatial analyses were conducted using ArcMap version 10.2 (ESRI, 2013).

Chapter 4. Restoring heterogeneous degraded tropical forests

4.3.2. Analysis framework

We used a decision-support tool, Robust Offsetting (RobOff) (Pouzols and Moilanen, 2013). RobOff is a resource allocation algorithm useful for prioritising “how much” of an alternative management option should be allocated to “which” environment. The objective is to maximise the benefit to features (e.g. species) given a particular level of investment considering the different responses of features to management action across environments and through time (Pouzols and Moilanen, 2013). The algorithm can account for multiple environments and multiple management actions.

RobOff requires the environments being analysed to be characterised. These environments are an aggregation of planning units that share similar characteristics, such as land uses, land cover, forest types, and management activities. For our analysis, we defined environments as a combination of forest types and degradation levels and herein referred to them as zones. Information on the characteristics of each zone was required including the proposed restoration action(s) and cost(s) of implementing these, the available area for undertaking each action, the occurrence of features (as a binary presence or absence), and the response of each feature to the action (measured on an annual basis). We focused on carbon and biodiversity as our features of interest and employed AGB accumulation (in tonnes) and the suitability of habitat for threatened mammals as surrogates for these features. Among the taxonomic groups that occur in the region, mammals have the best available data and are the most prominent in conservation policy (Soehartono et al., 2008).

Within the RobOff framework, we explored alternative scenarios representing four objectives: (a) prioritising carbon benefit by maximising AGB accumulation, (b) prioritising biodiversity benefit by maximising the improvement of habitat for threatened mammals, (c) prioritising both AGB accumulation and improvement of habitat for threatened mammals, and (d) prioritising forest types that were highly degraded. We ran the algorithm over a 30-year planning horizon and employed a 6.5% discount rate in calculating costs (Bank Indonesia, 2013).

4.3.3. Restoration zones

We developed a stratification matrix of forest conditions for East Kalimantan based on the level of degradation and applied this across broad forest types following Gibbs et al. (2007). In East Kalimantan, logging and fire have been the main drivers of forest degradation with obvious impacts on the reduction of AGB (Gaveau et al., 2014b; Langner et al., 2007; Pearson et al., 2014). We therefore used the extant stock of AGB as a proxy for the level of degradation, assuming that the highest value of AGB observed for a forest type is representative of an intact/pristine condition. We stratified degradation level into five broad classes (i.e. intact/primary forest, lightly degraded,

Chapter 4. Restoring heterogeneous degraded tropical forests moderately degraded, highly degraded and critically degraded/deforested) using the range of extant AGB for Bornean forest obtained from the available literature (Mori, 2001; Sasaki et al., 2011; Table 4.1). We used AGB map developed by Saatchi et al. (2011) to stratify the degradation level. The Saatchi AGB map has a 1-km spatial resolution, and among available carbon maps, it has the best agreement with the plot inventory data for Bornean rainforest collated by Slik et al. (2010). We used ecoregion data (Olson et al., 2001; Wikramanayake et al., 2002) to deliver six forest types across East Kalimantan. Mangrove forest was excluded from our analysis as it represents a transitional ecosystem between terrestrial and marine biomes, complicating the specification of soil parameters in the AGB modelling (see below). We identified 25 restoration zones representing a combination of five forest types and five degradation levels (Table 4.1; Figure 4.1).

Forest type Primary Lightly Moderately Highly Critically forest degraded degraded degraded degraded Lowland forest >411 308-410 206-307 103-205 <102

Montane forest >374 280-373 187-279 94-186 <93

Heath forest >351 263-350 176-262 88-175 <87

Peat swamp forest >345 260-345 173-259 87-172 <86

Freshwater swamp forest >213 160-212 107-159 54-106 <53

Chapter 4. Restoring heterogeneous degraded tropical forests

4.3.4. Restoration actions and costs

We determined the plausible restoration action for each restoration zone based on the published literature and government regulations (Hardiansyah, 2011; Korpelainen et al., 1995; Lamb et al., 2005; Maswar et al., 2001; Ministry of Forestry, 2012d, 2013b; Mori, 2001; Sasaki et al., 2011; Wibisono et al., 2005) (Tables 4.2 and A4.1). Specifically, the focus was on active restoration, assuming plantings of native tree seedlings of species predominately from the Dipterocarpaceae family at three degradation levels: critically, highly and moderately degraded forest (Table 4.2). We did not account for passive restoration (i.e. natural regrowth) on primary or lightly degraded forest.

Cost was divided into two categories: implementation and opportunity cost. Implementation costs consisted of expenditure for planting and maintenance. We used the standard cost of planting

Chapter 4. Restoring heterogeneous degraded tropical forests prescribed by the Indonesian government (Ministry of Forestry, 2012d) as a baseline with additional maintenance costs (e.g. fire control and patrolling activities) up to the fourth year as suggested by Hardiansyah (2011). These data were employed for lowland forest with modifiers based on literature review and personal communications with restoration practitioners to obtain implementation costs for the remaining forest types (Table A4.1). For example, preparing planting sites on peat swamp forests would require the construction of artificial mounds to minimise seedling mortality due to prolonged flooding (Wibisono et al., 2005). This activity would incur an additional cost of US$77 per hectare when restoring critically degraded peat swamp forest. We also explored the sensitivity of our results to the potential additional cost required for hydrological restoration in peat swamp forest using data from the Ex-Mega Rice Peatland Rehabilitation Project in Central Kalimantan (Kalimantan Forest Carbon Partnership, 2009). As the costs vary depending on the type of dam constructed and material used, we used an average cost of US$278 per hectare.

We defined the opportunity cost of forest restoration as the revenue foregone from timber extraction as a consequence of restoration. The average net present value (NPV) of timber harvesting in intact forest in Kalimantan (i.e. US$2,268 per hectare) was employed as a baseline for this cost type (Ruslandi et al., 2011). For moderately degraded forest, harvestable timber is reduced by approximately 40% (van Gardingen et al., 2003), resulting in an NPV of US$975.24 per hectare. For highly degraded forest, most of the trees are below the diameter cutting limit and large-scale commercial logging operations are no longer profitable (Sasaki et al., 2011). A conservative value of US$300 per hectare was therefore assigned as the opportunity cost for timber harvesting in these areas. For critically degraded forest, we did not assign an opportunity cost for timber harvesting as most remaining trees are pioneer species (Macaranga spp.) and commercial dipterocarp species are rare and small in size (Mori, 2001).

Chapter 4. Restoring heterogeneous degraded tropical forests

Table 4.2. Restoration actions and costs (in US$ per hectare) across restoration zones. Restoration costs consist of the implementation costs of planting (Table A4.1) and the revenue foregone from timber extraction.

Restoration zones Actions Cost (US$ per hectare) Lowland CDF Intensive square planting 1,275 Lowland HDF Strip planting 943 Lowland MDF Gap planting 1,395 Montane CDF Intensive square planting 1,450 Montane HDF Strip planting 1,024 Montane MDF Gap planting 1,435 Heath CDF Intensive square planting 1,964 Heath HDF Strip planting 1,231 Heath MDF Gap planting 1,539 Peat swamp CDF Intensive square planting 1,518 Peat swamp HDF Strip planting 1,047 Peat swamp MDF Gap planting 1,447 Freshwater swamp CDF Intensive square planting 1,463 Freshwater swamp HDF Strip planting 1,025 Freshwater swamp MDF Gap planting 1,435

4.3.5. Restoration investment

We employed a theoretical budget of US$500 million. We increased the budget by 2.5 times to explore the influence of budget availability on the zones prioritised for restoration and on the trade- offs observed between carbon and biodiversity.

4.3.6. Estimating AGB accumulation

The response of AGB when restoring each zone was estimated using a process-based model called 3- PG (physiological principles for predicting growth). Despite the initial purpose to predict the growth of monoculture plantations in temperate regions (Landsberg and Sands, 2011), this model has also performed well for estimating AGB of highly diverse forests in the tropics (Budiharta et al., 2014b; White et al., 2006). The model requires soil and climate variables as inputs, along with a set of physiological parameters.

We obtained soil texture classes, fertility ratings and maximum and minimum plant available soil water estimates from the Harmonized World Soil Database (FAO/IIASA/ISRIC/ISSCAS/JRC,

Chapter 4. Restoring heterogeneous degraded tropical forests

2012). Climate input variables, including monthly temperature (i.e. maximum, mean and minimum), monthly precipitation and vapour pressure deficit, were obtained from the WorldClim database (Hijmans et al., 2005) and solar radiation estimated from the monthly averages of surface insolation from 22 years of satellite observation provided by the POWER Project (The Prediction of Worldwide Energy Resource) (NASA, 2013). We employed physiological parameters developed specifically for Bornean rainforest and differentiated by forest type (Budiharta et al., 2014b). We estimated AGB accumulation over 30 years for each degradation level for each forest type (i.e. each restoration zone). In each zone, we assumed that AGB is accumulated from two sources: natural regrowth of existing vegetation and restoration planting in a specific design depending on the level of degradation (Table 4.2). Therefore, we separately modelled the AGB accumulated from both processes, then combined the results to estimate the total AGB accumulated for each restoration zone.

4.3.7. Estimating the suitability of habitat for threatened mammals

Our analysis focused on threatened mammal species within the IUCN classes Vulnerable and above (27 Vulnerable species and 10 Endangered species) (IUCN, 2013). To reduce uncertainty in species distributions, we intersected the habitat suitability models of terrestrial mammals (Rondinini et al., 2011) with their known geographical extent (IUCN, 2013). The resultant map was then overlaid with the restoration zones to determine the possible occurrence of each mammal in each zone. In total, 35 threatened mammal species occurred within the zones being considered for restoration (Table A4.2).

We assumed that the relationship between the suitability of habitat for each species and the extent of forest cover is determined by the sensitivity of each species to forest degradation (Wilson et al., 2010) and the habitat suitability would change overtime as forest restored. The sensitivity level was assessed by expert elicitation complemented with a literature review, and classified into three classes: low, medium and high sensitivity (Wilson et al., 2010). We used the enhancement of AGB stored during restoration to predict the trajectory of forest cover. In doing this, we conducted linear regression analysis of the relationship between forest cover derived from MODIS (Moderate Resolution Imaging Spectroradiometer) (Townshend et al., 2011) and AGB data from Saatchi et al. (2011). For each forest type, we randomly sampled 100 points to extract the values of percent forest cover and AGB from Townshend’s MODIS and Saatchi’s AGB maps respectively with a minimum distance of 10 km to minimise spatial autocorrelation. We used R Statistical Software (R v. 3.0.1, R Development Core Team, 2013) for the regression analysis and the relationships were significant at the level of P < 0.001 (Table A4.3). We then employed the relationship per forest type to predict the extent of forest cover using stored AGB as predictor and the resultant forest cover was used to derive the habitat suitability of threatened mammals.

Chapter 4. Restoring heterogeneous degraded tropical forests

Habitat suitability was estimated for each species within each restoration zone over 30 years using the following arbitrary rules: (i) for highly sensitive mammals nearly complete (90%) forest cover was required; (ii) for moderately sensitive species the majority (60%) of forest cover was required; and (iii) for low sensitivity species then only a small proportion (30%) of forest cover was required. We explored the sensitivity of our results by decreasing and increasing the baseline threshold of forest cover suitable for each mammal sensitivity class by 10%.

4.4. Results

We estimated that 6.1 million hectares (49.6%) of Forest Estate land (excluding ITP) in East Kalimantan was in critically, highly or moderately degraded condition, requiring active restoration (Table 4.3). Degradation was not uniformly distributed across forest types with peat swamp forest having the largest proportion of degraded areas (93.7%) followed by freshwater swamp forest (92.9%), while montane forest had the lowest (46.3%). The total budget needed to restore all degraded forests identified would be approximately US$8.37 billion if we considered both the implementation and opportunity costs, and US$3.24 billion if we only accounted for the implementation cost.

Table 4.3. Summary of degradation levels across forest types: total extent (in 000 of hectares) and the proportion of the overall extent of this forest type.

Forest types Degradation Lowland Montane Heath Peat swamp Freshwater Total classes swamp

Intact and lightly degraded 2,994 (53%) 3,029 (53.7%) 171 (25%) 17 (6.3%) 4 (7.1%) 6,217 Moderately degraded 2,178 (38.6%) 2,467 (43.7%) 258 (37.8%) 65 (23.6%) 13 (21.9%) 4,984 Highly degraded 428 (7.6%) 145 (2.6%) 206 (30.2%) 70 (25.3%) 36 (57.6%) 887 Critically degraded 44 (0.8%) 1 (0.0%) 48 (7%) 124 (44.8%) 8 (13.4%) 227

Total 5,646 5,644 683 278 63 12,316

Regardless of the objective and budget, it was recommended that highly degraded lowland forest be extensively restored (Tables 4.4 and 4.5). Restoring this zone incurred the lowest restoration costs, contributed to the restoration of the habitat of 29 of 35 threatened mammals, and accumulated the greatest amount of AGB. Heath forest had the smallest area recommended for restoration, with the exception of when the goal was to recover forest types that had been extensively degraded (Scenario 4). The implementation cost of planting heath forest was the highest of all forest types, while the

Chapter 4. Restoring heterogeneous degraded tropical forests accumulation of AGB was relatively low due to infertile soil (Table A4.1). Regardless of forest type, moderately degraded forest was least favoured for restoration (Tables 4.4), due to the presence of large commercial trees, which incurred a very high opportunity cost.

If the lower restoration budget (US$500 million) was assumed and the objective was to prioritise AGB accumulation (Scenario 1) the entire budget was allocated predominately to restoring highly degraded lowland and montane forest. This scenario achieved the best carbon outcome with 226 million tonnes of AGB stored over 30 years (Table 4.4). The restoration of highly degraded lowland forest contributed 83% of the AGB predicted to be accumulated and consumed 81% of the total budget. However, this carbon outcome would be at the expense of biodiversity goals with the habitat of the Bornean elephant (Elephas maximus borneensis) and aquatic mammals, including oriental small-clawed otter (Aonyx cinerea), hairy-nosed otter (Lutra sumatrana) and smooth-coated otter (Lutrogale perspicillata), not receiving any restoration investment (Table A4.2).

The highest biodiversity benefit was achieved under Scenario 2 (prioritising the restoration of threatened mammal habitat) with the zones selected for restoration encompassing the distributions of all threatened mammals (Tables 4.4 and A4.2). This comes at a cost in terms of the total AGB accumulated which was reduced by 24.4% compared with that achieved under Scenario 1. The overall extent of the area restored was also reduced by 16%. The restoration of the habitat of Bornean elephant was accommodated by selecting 3,582 hectares of moderately degraded lowland forest. Within our restoration zones, Bornean elephants were restricted to a small area near the border with Sabah that overlaps with moderately degraded forest in lowland and montane forest. A large amount of the budget was also allocated to restoring the habitat of aquatic mammals with the selection of 124,894 hectares of critically degraded peat swamp forest as this zone harbours all aquatic threatened mammals.

Prioritising both AGB accumulation and the restoration of threatened mammal habitat (Scenario 3) would deliver a more balanced solution. Under this scenario, total AGB stored was only 4.5% lower than Scenario 1 (prioritising AGB accumulation alone) with only the Bornean elephant not receiving any investment (Tables 4.4 and A4.2). Prioritising heavily degraded forest types (Scenario 4) results in the investment fund being distributed across a greater diversity of restoration zones, regardless of the higher cost of restoration (Table 4.4). As a consequence, this scenario performs worst in terms of carbon sequestration and biodiversity outcomes, with only 126 million tonnes of AGB stored and no restoration afforded to the habitat of Bornean elephant (Table A4.2). Under this scenario, all available restoration zones within freshwater swamp were selected for restoration, while for peat swamp and heath forest, only 52% and 17.3% of the available areas were prioritised respectively. Only 98% of

Chapter 4. Restoring heterogeneous degraded tropical forests the available budget was used in Scenario 4, indicating that the prioritisation of heavily degraded forest types obtained the maximum possible outcome for this objective and continued allocation of funding would be counterproductive considering the very high costs incurred.

Table 4.4. Recommended areas for restoration under each scenario using the lower investment level (US$500 million).

Restoration zones Available Scenario 1 Scenario 2 Scenario 3 Scenario 4 area (ha) (ha) (ha) (ha) (ha)

Lowland CDF 44,127 0 0 0 44,127 Lowland HDF 428,645 428,645 164,238 428,645 52,042 Lowland MDF 2,178,922 0 3,582 0 0 Montane CDF 1,776 0 0 1,776 1,776 Montane HDF 145,760 93,045 145,760 10,508 0 Montane MDF 2,467,541 0 0 0 0 Heath CDF 48,057 0 0 0 48,057

Heath HDF 206,454 0 0 0 40,868 Heath MDF 258,400 0 0 0 0 Peat swamp CDF 124,894 0 124,894 0 0 Peat swamp HDF 70,584 0 0 70,584 70,584 Peat swamp MDF 65,744 0 0 0 65,744 Freshwater swamp CDF 8,478 0 0 0 8,478 Freshwater swamp HDF 36,401 0 0 4,142 36,401 Freshwater swamp MDF 13,802 0 0 2,271 13,802

Total area (ha) 6,099,585 521,690 438,474 517,926 381,879 Total budget (US$ million) 499.86 498.94 499.25 491.25 Total AGB (million tonnes) 226.16 170.93 215.99 126.70 Number of mammals occurring in prioritised restoration zones 31 35 34 34

When the budget was increased (by 2.5 times), the trade-offs between objectives was reduced (Table 4.5). If the restoration of threatened mammal habitat was the focus (Scenario 2) then all mammals would obtain some investment and accumulation of AGB would be reduced by 6.7% compared with Scenario 1 (prioritising AGB accumulation). Conversely, under Scenario 1, no investment would be

Chapter 4. Restoring heterogeneous degraded tropical forests allocated to restore the habitat of three aquatic mammals (Tables 4.5 and A4.2). Optimising both carbon and biodiversity objectives (Scenario 3) results in the habitat of all threatened mammals receiving investment with AGB accumulation was only 1.2% lower than under Scenario 1.

Table 4.5. Recommended areas for restoration with investment using the higher restoration budget (US$1.25 billion).

Restoration zones Available Scenario 1 Scenario 2 Scenario 3 Scenario 4 area (ha) (ha) (ha) (ha) (ha)

Lowland CDF 44,127 44,127 0 0 44,127 Lowland HDF 428,645 428,645 428,645 428,645 428,645 Lowland MDF 2,178,922 454,704 387,873 429,635 0 Montane CDF 1,776 1,776 1,776 1,013 1,776 Montane HDF 145,760 145,760 81,518 145,760 145,760 Montane MDF 2,467,541 0 0 0 0 Heath CDF 48,057 0 535 0 48,057

Heath HDF 206,454 0 0 0 0 Heath MDF 258,400 0 0 0 80,185 Peat swamp CDF 124,894 0 124,894 0 34,392 Peat swamp HDF 70,584 0 0 70,584 70,584 Peat swamp MDF 65,744 0 0 0 65,744 Freshwater swamp CDF 8,478 0 8,478 8,478 8,478 Freshwater swamp HDF 36,401 0 10,318 3,162 36,401 Freshwater swamp MDF 13,802 0 0 3,773 13,802

Total area (ha) 6,099,585 1,075,012 1,044,037 1,091,050 977,951 Total budget (US$ million) 1,247.41 1,245.61 1,249.95 1,121.35 Total AGB (million tonnes) 462.28 431.44 456.60 373.38 Number of mammals occurring in prioritised restoration zones 32 35 35 35

Our results were insensitive to the threshold of forest cover suitable for each mammal sensitivity class. Either by decreasing and increasing the baseline threshold by 10%, all available areas of highly degraded lowland forest were consistently prioritised when the objective was carbon accumulation (Scenario 1) and dual objectives (Scenario 3). In addition, when focusing on threatened mammal

Chapter 4. Restoring heterogeneous degraded tropical forests habitat (Scenario 2), a comparably large portion of investment was constantly allocated to highly degraded lowland and montane forests, and critically degraded peat swamp forest. Adding a cost associated with hydrological restoration in peat swamp forest did not change the overall results, even for Scenario 2 in which all areas of critically degraded peat swamp forest were prioritised for restoration. This was likely because of the unique conservation value of this zone due to the occurrence of all aquatic mammals.

4.5. Discussion

We present a novel framework for systematically prioritising restoration areas that explicitly accounts for the contribution of a variety of restoration actions and for the heterogeneous nature of landscapes. We characterised a broad spectrum of landscape conditions, assigned plausible restoration actions, estimated costs and benefits for multiple actions, and adapted a newly-developed decision-support tool to prioritise the allocation of restoration investments.

By applying our framework in East Kalimantan, we find that highly degraded lowland forest (with 21-40% AGB remaining) is favoured for restoration, particularly if the objective is to enhance carbon stocks. This is due to the comparatively low costs associated when restoring this zone and the rapid accumulation of AGB. Regardless of objective, a large portion of the budget is always allocated to restoring highly degraded lowland forest, highlighting the importance of this zone for both carbon sequestration and restoring the habitat of threatened mammals. Murdiyarso et al. (2011) criticised the Indonesian government following the REDD+ agreement with Norway, which only protects primary forests and peat lands from new permits for logging concessions and conversion and excludes 46.7 million hectares of logged and secondary forests. Our finding corroborates Murdiyarso and colleagues’ proposal for the inclusion of degraded forests into the moratorium, particularly for over 400,000 hectares of highly degraded lowland forest in East Kalimantan. Apart from government funded REDD+ initiatives, our results suggest that these areas are the first priority for allocating Ecosystem Restoration Concession (ERC) (Boer, 2012). This new policy has attracted private investments to restore degraded forests either for carbon credit (e.g. Rimba Raya InfiniteEarth) or biodiversity conservation (e.g. Harapan Rainforest) or both (Birdlife International, 2014; The Guardian, 2013).

Moderately degraded forests (with 41–60% AGB remaining) are recommended for restoration if the budget is not limiting and the objective is to restore the habitat of threatened species (Scenario 2). The considerable volume of remaining timber in moderately degraded forests incurs a high opportunity cost, suggesting that forests under this degradation category are better managed for selective logging, provided they also occur on forests sanctioned for timber production (Hutan 53

Chapter 4. Restoring heterogeneous degraded tropical forests

Produksi). Across Kalimantan, well managed logging concessions have maintained forest cover and delivered significant contributions to biodiversity conservation (Gaveau et al., 2013; Meijaard et al., 2005; Wilson et al., 2010). Restoring critically degraded forests (with 0–20% AGB remaining) is recommended only if a larger budget is available (Table 4.5). An emerging option for restoration of these areas is to generate socio-economic benefits for local people, for example by the establishment of agroforestry under the legal framework of community forest (Hutan Kemasyarakatan) (Ministry of Forestry, 2007), which can be funded by the ongoing national forest rehabilitation program (i.e. Gerakan Nasional Rehabilitasi Lahan dan Hutan/GERHAN). In the worst conservation scenario, if the Forest Estate is converted to agriculture (e.g. oil-palm plantation), targeting critically degraded forests for such conversion is a wise option, as also suggested by Koh and Ghazoul (2010).

We discovered trade-offs between the objectives of carbon and biodiversity but these were highly dependent upon the budget available. When the budget was limited, AGB accumulation under Scenario 2 (prioritising the restoration of threatened mammal habitat) would be 24.4% lower than Scenario 1 (prioritising AGB accumulation only). However, the AGB scenario would fail to restore the habitat of four of 35 mammals. Under the higher budget, AGB accumulation under Scenario 2 would only be 6.7% lower than Scenario 1, with the habitat of an additional mammal receiving investment. Our finding reveals that with small adjustments, we could achieve a compromise solution for the dual objectives of carbon and biodiversity (Scenario 3). Employing the lower budget, the accumulation of AGB under Scenario 3 is only 4.5% lower than Scenario 1 with one mammal not receiving habitat restoration. This more balanced solution was also found by Venter et al. (2009a) for the REDD+ activity of avoided deforestation, with the number of species protected doubled for a reduction of just 4–8% in emissions avoided. In our case study, the trade-offs are influenced by species with either limited geographic range (such as the Bornean elephant) or unique habitat requirements (such as oriental small-clawed otters, hairy-nosed otters and smooth-coated otters).

Although it might reduce the carbon benefit and increase opportunity costs, restoration for biodiversity objectives is likely to attract socio-political support and financial contributions from the public (Dinerstein et al., 2013). For example, in 2012, the Indonesian Orangutan Habitat Restoration Program generated revenues amounting to US$6.4 million for managing three restoration and reintroduction projects in Kalimantan, increasing from US$4.0 million in 2011 (BOS Foundation, 2013). Similar initiatives could be expanded to restore the habitat of the Bornean elephant, as this species has a limited distribution in East Kalimantan and current protected areas do not afford adequate protection (Wilson et al., 2010). Protection of this species will however incur high opportunity costs as its distribution overlaps with moderately degraded forests that are currently allocated for logging concessions (Gaveau et al., 2013).

Chapter 4. Restoring heterogeneous degraded tropical forests

Our restoration framework does not account for passive restoration alone as this option is unlikely to be effective in the medium and long term. In highly degraded dipterocarp forest in Borneo, local recruitment of dipterocarp seedlings is limited because of seed predation and a lack of reproductively mature adults due to past harvesting (Ådjers et al., 1995; Curran and Webb, 2000). Recruitment through seed dispersal is low because dipterocarp seeds are heavy, poorly dispersed and recalcitrant, which results in low germination success (Ådjers et al., 1995; Kettle, 2010). Further, the resilience of pioneer species (e.g. Macaranga stands) constrains the remaining dipterocarp seedlings from obtaining access to nutrients and sunlight, necessitating active management, such as strip planting, to enhance restoration success (Aoyagi et al., 2013; Cannon et al., 1994; Hardiansyah, 2011). Apart from ecological constraints, Zahawi et al. (2014) warns that passively restored sites are often regarded as abandoned land with open access, leading to perverse actions (e.g. land encroachment, fallow practices), which put restoration goals in jeopardy. This situation could incur what they termed a “hidden cost of passive restoration” and is likely to be more apparent in developing countries where land tenure is often unclear.

To the best of our knowledge, comprehensive data representing forest degradation is not available for Borneo, or more specifically East Kalimantan, with the exception of spatial data of degraded lands (Lahan Kritis) released by the Ministry of Forestry (Ministry of Forestry, 2011). This dataset broadly delineates the state of certain areas for supporting hydrology and soil protection (e.g. vulnerability to flooding and soil erosion) rather than detailed vegetation conditions. While we mapped forest conditions based on the remaining AGB within a forest type, maximum AGB for intact forest could vary among sites (Ziegler et al., 2012) and therefore it would be preferred for the degradation threshold to be differentiated at a local scale. For selectively logged forests, logging techniques have varying intensity and diverse impacts on soil, with some areas suffering extensive topsoil erosion and compaction, impeding vegetation recovery either through natural regrowth or enrichment planting (Pinard et al., 2000). In peat swamp forest, degradation is often associated with modifications to drainage systems, and hydrological restoration through re-wetting and canal blocking can be required prior to vegetation restoration (Kalimantan Forest Carbon Partnership, 2009). The coarse resolution of the spatial data that we employ (particularly for forest types), along with the simplified, yet robust approach to stratify AGB warrants further field validation. In the future, integration of fine resolution estimates of extant biomass and vegetation cover as well as soil and hydrological condition would be beneficial to develop a more accurate assessment of forest degradation.

The suitability of habitat for threatened species was estimated through expert elicitation although our results are insensitive to the changes in the sensitivity thresholds. Wildlife exhibit specific responses to the dynamics of habitat resources following restoration, and the habitat suitability of particular

Chapter 4. Restoring heterogeneous degraded tropical forests species is often associated with elements in addition to canopy cover, such as tree hollows and the occurrence of vines and shrubs (Ansell et al., 2011; Edwards et al., 2009). Unfortunately, this kind of detailed information is unknown for most of mammals in East Kalimantan, except for well-known species such as the Bornean orangutan. Further empirical analysis of the dynamic contribution of restored forest to habitat suitability is required to fill this knowledge gap. Spatial configuration also affects the habitat suitability of species in regard to the shape and connectivity of habitat patches with some mammals (e.g. Bornean elephant) requiring extensive and well-connected habitat (Alfred et al., 2012). This problem is not addressed by RobOff as it does not produce spatially-explicit results. Subsequent analysis using spatial decision-support tools, for example Marxan (Ball et al., 2009) and Zonation (Moilanen et al., 2009a), might be required to translate the results of RobOff spatially. However, this analysis would need information on the level of spatial cohesion and connectivity required for each species, which is scarce for most mammals in Borneo. Ideally, the impact of climate change would be accounted for to estimate the suitability of habitat during the entire planning horizon. This analysis would require fine-resolution downscaling of environmental variables validated with long-term meso-and-fine scale climate data, which is presently unavailable for East Kalimantan.

We assumed that forests prioritised for restoration would be protected in perpetuity due to their contribution to carbon and biodiversity objectives whereas non-prioritised areas would be managed for other purposes (e.g. logging concession, community forest). As such, we assumed the opportunity cost of restored forest is incurred in perpetuity, but it is possible for profit to be generated from restored forests also. While our study area is within the Forest Estate, which is designated for forestry and conservation purposes, it could be relevant to also explore the opportunity costs of alternative land use options given the highly dynamic land use changes in the region aimed at boosting the economy (Koh and Ghazoul, 2010). In Kalimantan, oil-palm plantations alone are predicted to triple in extent by 2020 (Carlson et al., 2013). However, this modification should also account for the socio- political costs of changing land use (e.g. lobbying officials, amending regulations, potential legal disputes, and potential conflicts with communities), which is difficult to estimate, due to low levels of governance and high levels of corruption in Indonesia (Faisal, 2013).

While our analysis pioneers an approach for prudent and transparent allocation of restoration investments in heterogeneous landscapes from ecological and economical perspectives, accounting for socio-political variables would also be vital for restoration planning and decision-making. Social perception of local communities in valuing the forest varies spatially across the landscapes, depending on how they utilise the goods and services provided (Abram et al., 2014). As such, there will be communities that benefit and are disadvantaged by restoration, and the outcomes for communities may trade-off with ecological outcomes. To avoid perverse social impacts, some restoration projects

Chapter 4. Restoring heterogeneous degraded tropical forests could involve local people by recruiting them as labour (Birdlife International, 2014), create incentive systems to ensure that the restored forests are protected and maintained (Limin et al., 2008), and provide alternative livelihoods through better management of non-timber forest products (e.g. honey bees, rattan and medicinal plants) generated by the restored forests (Birdlife International, 2014; Rahmawati, 2013). In Indonesia, substantial improvement in governance would also be required, as these were the major causes of the failure of previous forest rehabilitation programs that have been co-ordinated by the government (Barr and Sayer, 2012). This includes resolving problems of tenure to minimise conflicts, developing transparent funding mechanisms to avoid corruption and fraud, and creating monitoring systems to measure restoration success (Barr and Sayer, 2012; Nawir et al., 2007a). In planning for implementation, multi-year funding should be accounted for to ensure the maintenance of sites after planting, including fire prevention and management (Nawir et al., 2007a).

4.6. Conclusion

Our analysis addresses the ecological heterogeneity of degraded forest in the tropics and the availability of a diverse suite of restoration approaches. Accounting for this heterogeneity is important as there is a trade-off between multiple objectives, implying that exploration of alternative solutions would be valuable in planning and decision-making for restoration. Our findings have policy implications for East Kalimantan, and Indonesia in general, in the context of REDD+, other ongoing and emerging restoration initiatives, and related land use planning. While the success of restoration programs is also determined by socio-political variables, our framework is a first step towards prudent planning for restoration to maximise the outcomes for carbon and biodiversity.

4.7. Acknowledgements

SB was financially supported by an Australian Awards Scholarship and Australian Research Council Centre of Excellence for Environmental Decisions and KAW by an Australian Research Council Future Fellowship. We thank Ramona Maggini, Atte Moilanen and Federico Pouzols for discussions on RobOff.

Chapter 5. Restoration offsetting in oil-palm landscape

Chapter 5

Paying back carbon and biodiversity debts from oil- palm plantations in Kalimantan.

To be submitted to Nature Climate Change as:

Budiharta, S., Meijaard, E., Gaveau, D., Struebig, M., Wilting, A., Kramer-Schadt, S., Niedballa, J., Raes, N., Possingham, H.P., Maron, M., Wilson, K.A. Paying back carbon and biodiversity debts from oil-palm plantations in Kalimantan.

Chapter 5. Restoration offsetting in oil-palm landscape

5. Paying back carbon and biodiversity debts from oil-palm plantations in Kalimantan.

5.1. Abstract

Oil palm plays a significant role in the Indonesian economy and global vegetable-oil production. It is an important contributor to deforestation in the country with concomitant impacts on carbon emissions and biodiversity. Offsetting is an emerging conservation strategy being proposed by the Roundtable on Sustainable Palm Oil to help mitigate these impacts. The potential for offsets to effectively compensate for oil palm impacts of different types at a landscape-scale is untested. Here, we estimate that to offset biodiversity impact attributable to 4.65 million hectares of industrial oil- palm development in Kalimantan would require 4.661 million hectares of degraded areas to be restored (equating to 8.74% of Kalimantan) at a cost of US$7.647 billion. Offsetting the carbon emissions from past oil-palm development is most cost-effectively achieved by restoring 0.789–1.057 million hectares of degraded peatlands, including the failed Ex-Mega Rice Project in Central Kalimantan, at a cost of US$1.33–1.847 billion. We show that the priority areas for offsetting biodiversity loss poorly overlap with those for compensating carbon emissions, limiting the potential for win-win solutions that simultaneously offset carbon and biodiversity impacts at a minimal cost and area extent. Action will therefore be constrained by the oil-palm industry’s economic capacity and the extent of land needed for Kalimantan’s development.

5.2. Introduction

The oil palm (Elaeis guineensis) industry contributes 2.64% (US$23 billion) of the Gross Domestic Product for Indonesia annually (Indonesian Bureau of Statistics, 2015a, b). The country champions palm oil supply worldwide generating 54% of global production (Food and Agriculture Organisation, 2015). While the contribution of the oil-palm industry to economic development is prominent, the costs associated with its social and environmental impacts remain debatable (Meijaard and Wilson, in review). In recognition that these costs exist, there has been political and market pressure on the oil-palm industry to become more environmentally responsible (Smedley, 2015; Smit et al., 2013). In response, oil-palm stakeholders have attempted to develop better governance systems through certification schemes, including the Roundtable on Sustainable Palm Oil (www.rspo.org). In 2014 the RSPO introduced a Remediation and Compensation Procedure allowing for past forest-clearing activities with impacts on biodiversity and ecosystem services to be offset, mainly through on-site and off-site restoration (RSPO, 2014).

Chapter 5. Restoration offsetting in oil-palm landscape

Environmental offsetting is a policy tool to mitigate the impacts of development activities (Kiesecker et al., 2009a; Maron et al., 2015). Offsets aim to counterbalance environmental damage through the generation of an equivalent benefit, such as through protection and/or restoration, elsewhere (i.e. no- net loss principle). Such policies have become increasingly common in the context of mining and urban development over the last decade, but with little application in less-profitable industries such as industrial agriculture (Madsen et al., 2010; Maron et al., 2015). Considering that agricultural activities have converted 80 million hectares of tropical forests alone (Gibbs et al., 2010), the emerging interest in offsetting by the agricultural sector presents potential opportunities for conservation. Yet, the potential for restoration to offset past impacts, and the capacity of agricultural industries such as plantation oil palm to pay for it is unclear.

Here, using a backcasting approach, we investigate the potential of restoration to offset the past impacts on biodiversity and carbon storage from oil-palm development in Kalimantan (Indonesian Borneo). Borneo is a global biodiversity and regional evolutionary hotspot with 534 threatened species (de Bruyn et al., 2014; Runting et al., 2015). The region also has high carbon storage capacity in the form of forest biomass and peat soil carbon (Page et al., 2011; Slik et al., 2010), but this capacity is highly threatened by deforestation, fire, and peat decomposition (Carlson et al., 2013; Gaveau et al., 2014b; Page et al., 2002). Oil palm is a major driver of these processes, with the industrial oil- palm estate estimated to have caused up to 3.4 million hectares (15%) of forest cover loss in Kalimantan (Gaveau et al., in review).

To calculate carbon emissions, and the loss of native vegetation and mammal habitat, we employed recently developed medium-resolution data on land conversion for industrial-scale oil palm in Kalimantan between 1973 and 2013 (Gaveau et al., in review). We then mapped the potential of degraded areas for restoration and quantified the benefits of restoration in term of carbon sequestration and avoided emissions, re-establishment of native vegetation of key floristic ecoregions, and re-creation of mammal habitat. We calculated the costs of restoration accounting for planting and maintenance, hydrological rehabilitation on peatlands, and opportunity costs as revenues forgone from oil-palm plantation, logging or agroforestry. We used a spatial decision-support tool (Moilanen et al., 2009b) to identify the most cost-effective areas for restoration with the target that restoration benefits gained were at least equal to the impacts from oil-palm development. We highlight the implications of our results to inform industry approaches for redressing past environmental impacts and for the future development of oil palm in Kalimantan.

Chapter 5. Restoration offsetting in oil-palm landscape

5.3. Methods

5.3.1. Oil palm-driven land conversion data

The oil palm-driven conversion map (Gaveau et al., in review) was generated from 357 LANDSAT images from 1973–2013 to determine the existing land cover prior to oil-palm plantation establishment. The use of finer resolution satellite imagery (i.e. 30x30m pixel) and a 5-year-interval analysis to detect intermediate land cover class is assumed to result in higher accuracy compared with other available data (e.g. 250m resolution of MODIS data (Abood et al., 2015; Koh et al., 2011), a 10-year-interval analysis (Abood et al., 2015; Carlson et al., 2013; Koh et al., 2011). The oil-palm map (Gaveau et al., in review) classified existing land cover into six broad categories: intact forest, logged forest, scrub/regrowth, other non-forested, “non-forest since 1973” and uncertain (e.g. cloud coverage/no data).

The class of “non-forest since 1973” in Gaveau et al. (in review) was defined as areas outside old- growth forest in 1973 due to the presence of human activities such as agroforest, fallow land and paddy field. As this class could have a broad range of land cover definitions and carbon stocks, we cross-checked the oil-palm map (Gaveau et al., in review) with the land cover map published by the Indonesian Ministry of Forestry (MoF; Ministry of Forestry, 2012c) to differentiate areas classified as “non-forest since 1973” into two broad categories: agroforest and non-forest. We then merged the non-forest areas into the class of other non-forested, omitted the class of non-forest since 1973 and delineated agroforest as a new class. As the MoF map is only available for the period of 2000 onward, oil-palm plantations established prior to 2000 and under non-forest since 1973 were merged into the uncertain category.

5.3.2. Quantifying the ecological impacts by oil-palm plantations

5.3.2.1. Carbon emissions from oil-palm plantations

We spatially calculated carbon dynamics from oil-palm plantation establishment using a loss-gain method (IPCC, 2006; Murdiyarso et al., 2010). We spatially stratified parameters used in the models (i.e. existing land cover class, mineral or peat soils and peat depth) to allow for better accuracy and to reduce uncertainty (Paoli et al., 2011). For oil-palm plantations occurring on mineral soils, carbon loss was estimated as the loss of above-ground biomass (AGB) of existing vegetation during land clearing while gain was calculated as AGB stored in oil-palm plantations (Equation 5.1). We used a 0.5 conversion factor as a fraction of carbon in dry biomass (Brown and Lugo, 1982).

Chapter 5. Restoration offsetting in oil-palm landscape

ΔCmineral = CAGB(i) - CAGB(OP) (Equation 5.1)

where ΔCmineral is net carbon emissions in mineral soils, CAGB (i) is the AGB carbon stock under land cover class i, and (CAGB (OP)) is the AGB carbon of oil-palm plantations. We did not account the for changes in soil carbon from the conversion of forest into oil palm plantation in mineral soils as there are large uncertainties associated with the quantify of this change (Falloon and Smith, 2003).

For the parameter of input of land cover class for intact forest, we used the mean value of AGB carbon (238 + 58.5 MgC per hectare) obtained from 62 sites of old growth forest on mineral soils across Borneo (Budiharta et al., 2014b; Slik et al., 2010). The estimates of AGB carbon for logged forest (130.0 + 74.67 MgC per hectare) were obtained from a pilot study that measured typical logged forests prior to conversion into oil-palm plantations (Dewi et al., 2009). Scrub was defined as degraded forest following forest fires (Gaveau et al., in review). Therein, for this land cover class, we employed the average AGB carbon of burned forest in East Kalimantan (Van der Laan et al., 2014) with value of 57.0 + 39.61 MgC per hectare. For agroforest, we extracted a value range of AGB carbon of agroforests and fallow lands across Kalimantan (Ziegler et al., 2012), resulting in 41 + 16 MgC per hectare. We assumed non-forested land to be severely degraded land dominated by grasses (e.g. Imperata cylindrica) and pioneer ferns and shrubs (e.g. Macaranga spp.), and assigned an input value of 10 + 8 MgC per hectare (Dewi et al., 2009; Otsamo, 1998; Ziegler et al., 2012).

The AGB carbon of oil-palm plantations (CAGB (OP)) was defined as the time-averaged AGB carbon over a 25-year planting cycle based on field data from Central Kalimantan (Dewi et al., 2009) with a value of 39 + 7.4 MgC per hectare, assuming 19% variability of the mean value (Morel et al., 2011). This value applied similarly to both mineral and peat soils (see below).

We added two additional emission sources when calculating carbon dynamics on peatlands (Equation 5.2): below-ground carbon emissions from peat burning and oxidation (decomposition) due to draining (Hooijer et al., 2010; Page et al., 2002).

ΔCpeat = CAGB(i) + Coxid(j) + Cburn - CAGB(OP) (Equation 5.2)

where ΔCpeat is the net carbon emissions in peat soils, CAGB (i) is the AGB carbon stock under land cover class i, Coxid (j) is carbon emissions from oxidation under peat depth j and Cburn is carbon emissions from peat burning. We used the average value of AGB carbon of old growth peat swamp forest (174.35 + 40.47 MgC per hectare) from seven sites across Borneo (Budiharta et al., 2014b) as input for intact forest on peat soils. For logged forest, we assumed that 54.6% AGB carbon is retained as in mineral soils (Dewi et al., 2009) resulting in an input value of 95.2 + 54.7 MgC per hectare. We

Chapter 5. Restoration offsetting in oil-palm landscape assigned values of AGB carbon for scrub, agroforest and non-forest similar to those in mineral soils with 57.0 + 39.61, 41 + 15 and 10 + 8 MgC per hectare respectively.

As carbon emissions from peat oxidation increase with drainage depth at a rate of 2.5 MgC per hectare per year for every 10 cm of additional depth (Couwenberg et al., 2010), we differentiated two levels of emissions from this source. For shallow peat soils (peat depth up to 50 cm), we used carbon emissions of 12.5 MgC per hectare per year, while for deep peat soils (peat depth more than 50 cm) we employed 20 MgC per hectare per year, assuming the recommended maximum drainage depth was 80 cm (Ministry of Agriculture, 2009). We used the peatlands base map developed by Sekala and Wetland International (Gingold et al., 2012) to assign peat depth.

Carbon emissions from peat burning have a large uncertainty as these variables are largely influenced by management practices of oil-palm planters and environmental conditions, such as prolonged drought during the El Niño events (Casson, 2000; Obidzinski et al., 2012). We therefore used estimates of 217.5 MgC per hectare to account for the annual probability of burning on drained peatlands in Southeast Asia (Hooijer et al., 2006; Venter et al., 2009b). This value is comparable to the average carbon emissions from peat burning across Indonesia by another study (Carlson et al., 2013) with 203 MgC per hectare. We then used the low and high values (72-386 MgC per hectare) to account for uncertainty (Carlson et al., 2013).

5.3.2.2. The loss of native vegetation by oil-palm plantations

We assumed biodiversity loss as the clearing of native vegetation to be replaced by monoculture oil- palm plantations. We used floristic ecoregions to represent the potential distribution of native vegetation types in Kalimantan (Raes, 2009). Raes (2009) classified Borneo into floristic ecoregions based on species distribution modelling using the MaxEnt algorithm (Phillips and Dudík, 2008) of more than 2,270 vascular plant species, using 44,000 herbarium records. Raes (2009) then clustered the resultant matrix of species distributions using a hierarchical clustering analysis and generated eleven floristic ecoregions, of which all occur in Kalimantan, using an indicator species analysis.

We masked the floristic ecoregion map (Raes, 2009) with the oil-palm plantation map (Gaveau et al., in review). As existing vegetation cover varies in terms of condition, due to anthropogenic and environmental factors such as logging and forest fires (Etter et al., 2011; Klein et al., 2009), we used an intactness-adjusted area (IAA) as the metric for native vegetation loss (Habib et al., 2013). The IAA was calculated as follows:

IAA(i) = A(i) x I(j) (Equation 5.3)

Chapter 5. Restoration offsetting in oil-palm landscape where IAA(i) is intactness-adjusted area for floristic ecoregion i, A(i) is the extent of area lost due to oil-palm establishment under floristic ecoregion i, and I(j) is the intactness index for land cover class j. We used species richness of native trees to generate the parameters of a floristic intactness index with the rationale that Borneo’s terrestrial ecosystems were historically composed of tree-dominated ecosystems (i.e. forests) with limited evidence of the prevalence of other vegetation types in the past (e.g. savannahs; Raes et al., 2014). We assumed that intact forest serves as a baseline system with an intactness index of 1. We assigned an average value of intactness index of 0.77 for logged forest, as species-area curves per hectare showed that this land cover type retains 74-80% of tree species of intact forest (Cannon et al., 1998; Imai et al., 2012). For scrub, we assumed that burned forest has 30% floristic similarity in trees to intact forest (Slik et al., 2008), resulting in 0.3 for the intactness index. An intactness index of 0.23 was assigned to agroforest according to the average similarity indices between primary forest and forest garden systems in Maluku, Indonesia (Kaya et al., 2002). We assigned a zero value of the intactness index for non-forested areas as tree species richness is extremely low, especially on I. cylindrica grassland (Potter, 1996).

5.3.2.3. The loss of mammal habitat by oil-palm plantations

We also measured biodiversity loss as the loss of original habitat of mammal species impacted by oil- palm development. We employed recently developed habitat suitability map of 81 mammal species (Struebig et al., 2015b) belonging to three groups: carnivores (23 species), primates (13 species) and bats (45 species). Struebig et al. (2015b) employed the Maximum Entropy (MaxEnt) algorithm (Phillips and Dudík, 2008) to map an environmental envelope for each species using bioclimatic variables (i.e. climates, topographic, and distances to water, wetlands and limestone karst). Struebig et al. (2015b) then corrected the resultant environmental envelope map with mammal sensitivity to land cover (Wilting et al., 2010) and through experts verification to produce habitat suitability map for each species. For our analysis, we used a habitat suitability map with strict treatment of possible omission errors (i.e. 25%), reflecting the core habitat inside the known geographical range of the species (Struebig et al., 2015a). We calculated habitat loss for each mammal by masking its habitat suitability map onto the oil-palm plantation map (Gaveau et al., in review).

5.3.3. Potential areas for restoration

We defined potential areas for restoration offsetting as areas that were currently deforested or degraded. We employed the land cover map (Gaveau et al., 2014b) generated from ALOS PALSAR data (Hoekman et al., 2010) to classify land cover into nine categories. For our analysis, deforested and degraded areas were those under the class of non-forest, agroforest, scrub/burned forest and logged forest. Within these areas, we mapped ‘future’ landscapes assuming all degraded areas are

Chapter 5. Restoration offsetting in oil-palm landscape restored accounting for the benefits of restoration in term of carbon, re-establishment of native vegetation and mammal habitat suitability. We also calculated the costs of restoration in a spatially- explicit manner.

5.3.4. Quantifying the benefits of restoration

5.3.4.1. The benefit on carbon

We calculated and mapped the carbon benefit based on the difference in value between the initial condition and the restored state (Evans et al., 2015; Maron et al., 2013; Figure 5.1). We differentiated potential sources of carbon benefits from restoration between mineral soils and peatlands. For mineral soils, the carbon benefit was formulated as:

ΔCmineral = Cseq(ij) - CAGB(i) (Equation 5.4)

where ΔCmineral is net carbon gain from restoration arising from Cseq(ij), the total AGB carbon sequestered on a restored site currently under land cover class i and floristic ecoregion j, and CAGB(i) the initial AGB carbon stock under land cover class i.

We calculated carbon sequestration using the 3-PG (physiological principle for predicting growth) model (Budiharta et al., 2014b; Landsberg and Waring, 1997) over 25 years—equating to one oil- palm cycle. Soil texture classes, fertility ratings and maximum and minimum plant available soil water were obtained from the Harmonised World Soil Database (FAO/IIASA/ISRIC/ISSCAS/JRC, 2012). Climate variables including monthly temperature, monthly precipitation and vapour pressure deficit were obtained from the WorldClim database (Hijmans et al., 2005), and solar radiation data from the POWER project (The Prediction of Worldwide Energy Resource; NASA, 2013). We parameterised physiological inputs for the model for each floristic ecoregion (Budiharta et al., 2014a; Budiharta et al., 2014b).

For restoration occurring on peatlands, two additional potential sources of carbon gain were added (Equation 5.5):

ΔCpeat = Cseq(i) + Coxid(j) + Cburn - CAGB(i) (Equation 5.5)

where ΔCpeat is net carbon gain from restoration on peatlands, Cseq(i) is the total AGB carbon sequestered on a restored peatland site currently under land cover class i, Coxid(j) is the avoided carbon emissions from peat oxidation under peat depth j, Cburn is the avoided carbon emissions from peat burning, and CAGB(i) is the initial AGB carbon stock under land cover class i. Using this equation, we assumed that in the absence of restoration, degraded peatlands would be drained and burned for

Chapter 5. Restoration offsetting in oil-palm landscape agriculture, and thereafter emissions would occur (Gaveau et al., 2014a; Harrison et al., 2009). We explored the sensitivity of the assumption that restoration would fail to prevent fires on peat by omitting Cburn from Equation 5.5 (i.e. fires would also occur on peat of restored peatlands). We assigned parameter values for avoided carbon emissions from peat oxidation and burning, similar to when calculating carbon loss (see Section 5.3.2.1).

5.3.4.2. The benefit on the re-establishment of native vegetation

We calculated the benefit of restoration on native vegetation establishment by subtracting the restored state by the initial state before restoration occurs (Evans et al., 2015; Figure 5.2). We assumed that restoration would fully recover native vegetation on the degraded areas and thus accumulate area in the intact condition class and determined the net gain in extent (measured as the change in intactness- adjusted area, ΔIAA(i)) for each floristic ecoregion i (Equation 5.6):

ΔIAA(i) = A(i) [1 - I(j)] (Equation 5.6)

Chapter 5. Restoration offsetting in oil-palm landscape where A(i) is the extent of area restored under floristic ecoregion i and I(j) is intactness index of the initial state for land cover class j. We assigned parameter values for the intactness index for each land cover class as per the calculations for native vegetation loss (see Section 5.3.2.2).

Figure 5.2. Net gain in term of re-establishment of native vegetation from restoration on degraded areas measured as intactness-adjusted area (IAA) of floristic ecoregion. Note that two floristic ecoregions (i.e. hill forest and montane forest of the upper Kapuas) are not mapped as no oil-palm plantations have been established.

5.3.4.3. The benefit on the re-creation of mammal habitat

The contribution of each potential restoration offset site to mammal habitat was calculated as the extent of degraded areas (Gaveau et al., 2014b) that occurred within historical suitable habitat (i.e. prior to industrialisation in Kalimantan, which commenced in the 1950s). As such, we assumed that restoration would fully recover the degraded areas to their pre-1950 condition. Historically suitable habitat was spatially modelled using the MaxEnt algorithm (Phillips and Dudík, 2008) and bioclimatic variables as predictors (Struebig et al., 2015b), and was then corrected with historical land cover (Struebig et al., 2015a). We employed a strict commission error threshold of 25% to assign habitat suitability of restored site into binary category.

Chapter 5. Restoration offsetting in oil-palm landscape

5.3.5. The costs of restoration

We calculated the restoration cost as a combination of the implementation and opportunity costs (Figure 5.3). The implementation cost was based on the standard cost of forest rehabilitation in Indonesia as prescribed by the Ministry of Forestry (Budiharta et al., 2014a; Ministry of Forestry, 2012d). This cost captures expenses related to planting activities and maintenance, including fire prevention. For restoration occurring on degraded peatlands, we also accounted for the cost of hydrological rehabilitation assuming dam construction was required to decommission canals (Budiharta et al., 2014a; Kalimantan Forest Carbon Partnership, 2009).

The opportunity cost was defined as the revenue forgone due to restoration for alternative forms of land management (Table A5.1). We considered oil-palm plantations and logging as the most relevant alternative land uses in the region (Carlson et al., 2013; Gaveau et al., 2014b). We employed the Net Present Value (NPV) of oil-palm plantations managed by listed companies as the baseline opportunity cost (Irawan et al., 2013). We differentiated the NPV of oil-palm plantation on the basis of land suitability mapped using 11 biophysical variables (Table A5.2). We added potential revenues from timber extraction during land clearing, if the areas suitable for oil-palm plantations overlapped with extant forest (Venter et al., 2009b). For areas not suitable for oil palm, the opportunity cost was derived from timber revenue if it occurred on logged and burned forest, and from agroforestry products if it occurred on agroforest (Table A5.1).

Chapter 5. Restoration offsetting in oil-palm landscape

Figure 5.3. The cost of restoration in degraded areas accounting for the implementation and opportunity costs. Degraded areas were defined as non-forest, agroforest, burned forest and logged. The implementation cost covered expenses for planting and maintenance, and cost for hydrological restoration if it occurred on degraded peatlands. Opportunity cost was assumed as forgone revenues from oil-palm plantations, or logging or agroforestry assigned on the basis of biophysical suitability and existing land cover class.

5.3.6. Prioritising areas for restoration offsetting

We prioritised potential areas for restoration offsetting using the decision-support tool Zonation v.4 (Moilanen et al., 2014). For each feature (i.e. carbon, floristic ecoregions and mammal habitat) we determined the loss incurred due to the development of oil-palm plantations (as detailed above) and employed this as a target in the prioritisation analysis. A target-based algorithm sought the most cost- efficient combination of areas to meet these targets. We also investigated the resultant priority areas by seeking to compensate for the loss of: (a) floristic ecoregions only; (b) mammal habitat only; (c) floristic ecoregions and mammal habitat; (d) carbon only; (e) carbon and floristic ecoregions; (f) carbon and mammal habitat. All input layers (and the resultant priority maps) had a spatial resolution of 100 hectares to align with the minimum size of industrial-scale oil-palm plantations (Gaveau et al., in review).

Chapter 5. Restoration offsetting in oil-palm landscape

5.4. Results

Between 1973 and 2013, 4.65 million hectares of industrial oil-palm plantations were established in Kalimantan (Gaveau et al., in review). We estimated net emissions of 0.66 GtC (0.36–0.96 GtC) of carbon over a 25-year planting cycle (Figure 5.4). This equates to 11% of the total emissions from land use and land cover change in Indonesia (Busch et al., 2015), highlighting that mitigating carbon impacts by oil-palm plantations in Kalimantan would contribute substantially to achieving the national emissions reduction target. While oil-palm plantations on peatlands occupied only 14.3% of the total plantation area they contributed 74.8% of total carbon emissions from oil-palm development (Figure 5.4). Carbon emitted from peatland conversion (averaged across land cover classes and peat depths) was 745 tC per hectare, more than five times higher than mean emissions from converting forests (of intact and logged condition) on mineral soils (136 tC per hectare) (Figure 5.5). Establishing oil-palm plantations on mineral soils in non-forest areas has resulted in a net carbon gain (29 tC per hectare).

0.25

0.2

0.15

0.1

0.05

0 Total carbon emissions Totalcarbon emissions (Gt C)

-0.05

Intact Intact

Scrub Scrub

Logged Logged

Agroforest Agroforest

NonForest NonForest Mineral soils Peatlands

Chapter 5. Restoration offsetting in oil-palm landscape

1200

1000

800

600

400

200

Net carbon emissions Net carbonemissions (t C/ha) 0

-200

Intact Intact Intact

Scrub Scrub Scrub

Logged Logged Logged

Agroforest Agroforest Agroforest

NonForest NonForest NonForest Shallow Deep Mineral soils Peatlands

Figure 5.5. Average net carbon emissions per hectare from oil-palm plantation development for each soil type, peat depth, and land-cover type over a 25-year planting cycle. Error bars represent lower and upper estimates of carbon emissions.

Industrial oil-palm plantations have converted the equivalent of 1.85 million hectares of intact floristic ecoregions (intactness-adjusted area/IAA; Table 5.1). Lowland forest of southern Kalimantan is the ecoregion that has been most extensively replaced with oil palm (1.03 million hectares of total extent; IAA of 0.453 million hectares). Heath forest has had the greatest proportional loss with 18.7% of the historical extent being converted to oil-palm plantations. The most intensely impacted ecoregion (with the highest IAA relative to oil-palm extent) is lowland forest of northern Kalimantan as plantations have typically replaced intact and logged forests.

It is estimated that the habitat of 78 mammals (96.3% of the sample) has been impacted by industrial oil-palm plantations with an average of 7.59% (+ 2.94) of originally-suitable habitat having been converted (Table A5.3). For charismatic mammals such as Bornean orangutan (Pongo pygmaeus) and Proboscis monkey (Nasalis larvatus), oil-palm plantations have replaced more than 10% of their originally-suitable habitat across Kalimantan.

Table 5.1. Extent of native vegetation in Kalimantan replaced by oil-palm plantations. Intactness-adjusted area (IAA) represents ecological integrity of extant native vegetation. Native vegetation is represented by floristic ecoregion using clustering analyses (Raes, 2009). For the purpose of this paper, some nomenclatures of the ecoregions were modified from the original dataset (Raes, 2009).

Historical Extent occupied by oil-palm plantations (000 ha) IAA relative IAA Floristic ecoregion name extent Intact Logged Scrub Agroforest Non- Uncertain Total Per cent to oil-palm (000 ha) (000 ha) forest extent occupied (%) extent (%)

Freshwater swamp forest 4,971 71 99 156 69 95 23 515 10.38 211 41.0 Peat swamp forest 4,575 67 113 161 152 123 21 639 13.98 238 37.3 Heath forest 4,205 95 137 222 142 174 16 788 18.74 300 38.1 Lowland forest of western Kalimantan 4,865 14 11 0.308 80 37 41 186 3.82 41 22.5 Lowland forest of central Kalimantan 4,991 40 45 27 139 34 71 358 7.19 115 32.3 Lowland forest of southern Kalimantan 10,346 142 301 173 118 226 67 1,030 9.96 453 44.1 Lowland forest of northern Kalimantan 1,083 32 79 0.450 19 15 0 147 13.63 98 66.9 Lowland forest of eastern Kalimantan 8,056 132 141 466 44 85 11 881 10.94 391 44.4

72 Hill forest 445 0 0 0 0 0 0 0 0.00 0 N/A Chapter

Montane forest of the upper Kapuas 3,529 0 0 0 0 0 0 0 0.00 0 N/A Montane forest of eastern Kalimantan 5,765 0 0.479 0 0 0.246 0 0.725 0.01 0.368 50.9

5. Restoration offsetting oil in palm landscape

Chapter 5. Restoration offsetting in oil-palm landscape

To offset the combined biodiversity losses by industrial oil-palm plantations developed between 1973 and 2013, the oil-palm industry would need to restore up to 8.74% of Kalimantan’s landmass. Offsetting the loss of floristic ecoregions alone (equivalent to the IAA) would require the restoration of 2.193 million hectares at a cost of US$3.59 billion (Figures 5.6a and 5.6b). Priority areas for offsetting this loss are dispersed across Kalimantan (Figure 5.7a). To offset the loss of mammal habitat, a total area of 4.596 million hectares would be required at a predicted cost of US$7.647 billion (Figures 5.6a and 5.6b). Priority areas for offsetting would include severely degraded lowland forests in East Kalimantan which were subject to severe fires during the El Niño events (Figure 5.7b). Simultaneously offsetting the losses of floristic ecoregions and mammal habitat indicates the most extensive areas required for offsetting among all scenarios (4.661 million hectares) and the highest cost (US$7.736 billion) (Figures 5.6a, 5.6b and 5.7c).

To offset the emitted carbon from the creation of industrial oil-palm plantations would require restoration of 0.789 million hectares (0.420–1.157 million hectares) and incur costs of US$1.33 billion (US$0.703–2.04 billion) assuming that restoration would effectively avoid emissions from further peat fires (Figures 5.6a and 5.6b). The areas selected for offsetting of carbon impacts are primarily degraded areas of deep peat, including the failed Ex-Mega Rice Project (EMRP) in Central Kalimantan (Figure 5.7d). Restoration of these areas avoids carbon emissions from peat oxidation and burning, and incur low opportunity costs due to their unsuitability for oil palm and timber extraction, although the cost of hydrological rehabilitation is high (Figure 5.3). If we assume that the restoration would fail to prevent peat fires, the required area for compensation would increase to 1.057 million hectares (0.554–1.592 million hectares) with the cost of restoration rising to US$1.847 billion (US$0.969–2.93 billion) (Figures A5.1 and A5.2).

When attempting to achieve the offset targets for restoration of floristic ecoregions and carbon simultaneously, the extent of offsetting areas is similar to that required when compensating the loss of floristic ecoregions only (Figure 5.6a), but the priority areas are altered to include degraded peatlands in the EMRP (Figure 5.7e). This spatial shift would incur a higher cost of US$3.711 billion (US$3.566–4.111 billion) (Figures 5.6b). When carbon offset targets were included with mammal habitat targets, the extent of offsets increases to 4.642 million hectares incurring costs of US$7.736 billion (Figures 5.6a, 5.6b and 5.7f).

Chapter 5. Restoration offsetting in oil-palm landscape

a b 6 10

5 8

4 6 3 4

2 Extent Extent (Mha)

Cost (US$ billion) 2 1

0 0 F M F&M C C&F C&M F M F&M C C&F C&M Scenario Scenario

Figure 5.6. Resources required to offset the impacts of oil-palm plantation in Kalimantan. a, Extent of landscape selected for restoration offsetting. b, Total offsetting costs accounting for opportunity and implementation costs. Each offsetting scenario aims to compensate for the loss of: floristic ecoregion (scenario F); mammal habitat (scenario M); floristic ecoregion and mammal habitat (scenario F&M); carbon (scenario C); carbon and floristic ecoregion (scenario C&F); carbon and mammal habitat (scenario C&M). Error bars represent the range of results accounting for lower and higher estimates of total carbon emissions.

a b c

75 d e f

Chapter

5. Restoration offsetting oil in pa

Figure 5.7. Priority areas for offsetting the impacts of oil-palm development via restoration in Kalimantan. Each figure represents different scenarios lm landscape for compensating the loss of: (a) floristic ecoregion; (b) mammal habitat; (c) floristic ecoregion and mammal habitat; (d) carbon; (e) carbon and floristic ecoregion; (f) carbon and mammal habitat. Label refers to province: West Kalimantan (WK), Central Kalimantan (CK), South Kalimantan (SK), East Kalimantan (EK) and North Kalimantan (NK).

Chapter 5. Restoration offsetting in oil-palm landscape

Figure 5.8. Overlapping priority areas between offsetting the impacts of oil-palm development on carbon emissions and compensating biodiversity loss (i.e. mammal’s habitat and floristic ecoregion combined) in Kalimantan.

5.5. Discussion

Our results indicate that full compensation for two elements of biodiversity loss could be achieved through restoration of an area equivalent to the overall extent of Kalimantan’s industrial oil-palm plantation estate, resulting in an offset ratio of 1:1. We have however assumed that restoration will successfully recover native vegetation and the habitat of mammals. In reality, there are long time- lags and uncertainties in restoring sites to the level of being intact systems, sometimes requiring centuries (Curran et al., 2013), although there is some evidence in the tropics that species rapidly colonise the restored sites within three decades (e.g. Ansell et al., 2011; Edwards et al., 2009; Gilroy et al., 2014). There have been non-spatially explicit methodologies to account for time-lags and uncertainties in restoration offsetting which commonly use multipliers in calculating offset ratio (e.g. Gibbons et al., 2015; Laitila et al., 2014). We anticipate that accounting for these variables in a spatial analysis will result in the requirement for areas much larger than that which has been impacted to require offsetting in order to mitigate impacts (Maron et al., 2012).

We also accounted only for a sample of mammals and ignored habitat connectivity in our prioritisation analysis (Lentini et al., 2013). Spatial distribution of biodiversity richness and

Chapter 5. Restoration offsetting in oil-palm landscape endemism poorly overlap across taxa in Kalimantan (e.g. mammals, birds, plants; Murray et al., 2015), and the incorporation of a complete set of taxa along with their connectivity requirements will likely also increase the extent of offsetting areas and costs. We therefore further caution that an equal offset ratio indicated by our analysis represents a minimum value to compensate biodiversity values that have been lost due to oil-palm development in Kalimantan. Even using this conservative estimate, the extent of offsetting areas required to fully compensate biodiversity losses will likely be increasingly constrained by regional development targets aiming for the expansion of oil palm and industrial timber plantations, logging and mining (Abood et al., 2015; Runting et al., 2015). The high costs incurred also raises questions about the economic capacity of the oil-palm industry.

We show that restoring degraded peatlands with deep peat thickness is a cost-effective strategy to offset carbon emissions from industrial oil-palm development in Kalimantan. Assuming that the net present value (NPV) of oil-palm plantation per hectare is US$6,355 (Irawan et al., 2013), total NPV of the oil-palm industry in Kalimantan would be US$29.6 billion, suggesting a potential opportunity of the industry to invest in carbon offsetting. The relatively small extent of offsetting areas (0.789– 1.057 million hectares) has likely less complication for land use policies in Kalimantan. Also, in response to recent catastrophic peat fires, the Indonesian Government have pledged to restore two million hectares of degraded peatlands across the country and is urging the contribution from the oil- palm industry (Hidayat and Primandari, 2015).

Degraded peatlands have been the subject of agriculture activities, either by large companies or small- scale farmers, through draining and burning (Gaveau et al., 2014a; Harrison et al., 2009), which has led to the release of carbon emissions between 0.81–2.57 GtC in one El Niño event alone (Page et al., 2002). Peat fires also cause up to US$33 billion in economic losses and severe public health problems (Chan, 2015). Our findings imply that no extractive activities should take place in degraded peatlands, and rather these areas could be restored under the carbon offsetting scheme. In implementing peatland restoration, however, social contexts must be considered to enhance effectiveness and mitigate perverse social impacts (Budiharta et al., in review), for example reforestation by employing local people and using locally-important native species that generate non-timber forest products, such as Dyera lowii.

There is only 0.23 million hectares of the priority areas for restoration aiming to compensate biodiversity losses overlaps with those for offsetting carbon emissions (Figure 5.8). The lack of spatial congruence between priority areas for compensating carbon emissions and biodiversity loss suggests that win-win solutions are likely to be problematic. This finding mirrors similar situation in REDD+ policy implementations in Kalimantan where there is spatial mismatch between areas for

Chapter 5. Restoration offsetting in oil-palm landscape climate change mitigation (by protecting carbon-rich sites such as peatlands) or biodiversity conservation (by focusing on species-rich areas)(Murray et al., 2015; Paoli et al., 2010). The oil-palm industry thus faces a difficult choice between offsetting targeting carbon or biodiversity. In the context of Kalimantan, the most feasible offsetting scenario is influenced by the capacity of oil-palm sector and the political context. Considering the comparatively smaller extent of offsetting areas and costs, and current public concern about peat fires, carbon offsetting will likely take priority over biodiversity offsetting.

The RSPO uses a site-by-site approach in their offsetting mechanism which does not provide clear guidance where to implement restoration offsetting (RSPO, 2014). The cost-effectiveness of a site- by-site approach in compensating ecological impacts is yet untested and this needs to be compared with landscape approach employed in this study. The RSPO offset policy is also voluntary and thus could potentially result in leakage as offsets implemented by some oil-palm plantations owners could be negated by the un-compensated ecological impacts from plantations that do not adopt the policy and continue to convert natural forests and peatlands. If this is the case, the no-net loss principle in offsets would not be achieved.

5.6. Conclusion

Oil-palm plantations in Kalimantan are predicted to expand between 6.9–9.4 million hectares (Abood et al., 2015; Carlson et al., 2013; Runting et al., 2015) over the next five years. To minimise carbon emissions and thus offset future liabilities, our results imply that there should be no oil-palm expansion on peatlands including areas in a degraded condition. However, shifting expansion away from peatlands will likely exacerbate the impacts on biodiversity, burdening the oil-palm sector to offset more of its biodiversity impact. Unless any additional mechanisms are introduced (e.g. strong premium price for low-biodiversity-impact oil palm), the oil-palm industry is likely to have limited capacity to fully implement biodiversity offsetting.

Chapter 6. Enhancing feasibility in restoration

Chapter 6

Enhancing feasibility: incorporating a socio-ecological systems framework into restoration planning

Submitted as:

Chapter 6. Enhancing feasibility in restoration

6. Enhancing feasibility: incorporating a socio-ecological systems framework into restoration planning

6.1. Abstract

Forest restoration is the counterforce to deforestation. In many parts of the world it mitigates forest loss and degradation, but success rates vary. Socio-political variables are important predictors of effectiveness of restoration activities, indicating that restoration strategies need to be locally adapted. Yet, contextual assessments of the biophysical, social and political characteristics of forest restoration are rare. Here, we integrate a social-ecological systems framework with systematic decision-making to inform forest restoration planning. We illustrate this approach through a prioritization analysis in a community-based forest restoration context in Paser District, East Kalimantan, Indonesia. We compare the solutions of our integrated framework with those identified on the basis of biophysical criteria alone. We discover that incorporating a socio-political context alters the selection of priority areas. While the social feasibility and political permissibility can be enhanced, ecological benefits are likely to be reduced and/or opportunity costs of alternative land uses are to be increased. Our conceptual framework allows the appraisal of potential trade-offs between social and ecological outcomes of alternative options, and has the potential to evaluate the efficiency of existing policies. Empirical testing in a range of contexts is required to ensure broad applicability and transferability of our conceptual framework.

6.2. Introduction

Tropical deforestation has been the primary contributor to forest loss globally over the past decade (Hansen et al., 2013). An estimated 500-600 million hectares (or 30-40%) of tropical forest is considered to be in a degraded state (Blaser et al., 2011). In Southeast Asia, primary forest represents a small proportion of total forest cover, with most forests having experienced some form of logging or extraction (Sodhi et al., 2009). As such, the goods and services provided by forests such as timber and non-timber products, habitat for biodiversity and water regulation, have deteriorated, often with severe consequences for forest-dependent communities (Abram et al., 2014; Lamb et al., 2005; Wells et al., 2013). Forest restoration to mitigate forest loss and/or degradation has had varied success, with performance strongly moderated by country-specific socio-ecological and political contexts (Food and Agriculture Organisation, 2011; Lamb, 2010; Lamb et al., 2005; Meyfroidt and Lambin, 2011).

When implementing forest restoration, either through afforestation or reforestation, context is especially important in countries such as Indonesia that are ecologically and socio-culturally

Chapter 6. Enhancing feasibility in restoration heterogeneous (Fearon, 2003; Lamb, 2010; Nagendra, 2007; Ostrom and Cox, 2010). For example, the Government of Indonesia has implemented a variety of forest restoration programs since the 1950s, mainly using a top-down approach (Murniati et al., 2007). While most restoration programs are considered unsuccessful (Murniati et al., 2007), remarkable success has been achieved in central Java through reforestation using teak (Tectona grandis) on severely degraded lands on drought-prone limestone soils (Nawir et al., 2007b). This achievement is largely due to familiarity with teak planting (with the practice dating back to the 1800s), highly motivated communities seeking to enhance the provision of water, and compatibility of the program with the capacity of local communities (e.g. the communities have other sources of income while the planted trees reach a harvestable size). Conversely, a lack of fit towards prescribed species (including teak) with local experiences and resource needs contributes to limited success of reforestation programs in other regions such as in Sumatra and Kalimantan (Indonesian Borneo), along with other factors including unclear tenure and complicated funding mechanisms (Nawir et al., 2007a).

The methods available for planning restoration activities are increasingly sophisticated, which account for both spatial and temporal heterogeneity (e.g. Birch et al., 2010; Budiharta et al., 2014a). Most of restoration design studies have focused on ecological criteria with few analyses incorporating social elements (e.g. Jellinek et al., 2014; Orsi and Geneletti, 2010). Furthermore, spatial heterogeneity in the social context of where restoration will be undertaken has not been captured. Instead, the social or political feasibility of restoration is assumed to be homogenous across large geographic extents, even including entire nations or continents, ignoring the local context of the planning region (e.g. Carwardine et al., 2015; Egoh et al., 2014). While contextual assessments that account for both the socio-political and ecological characteristics of a region are becomingly available in the environmental management literature (e.g. Basurto et al., 2013 in fisheries; Baur, 2013 in pasture; Cox, 2014 in irrigation systems), similar analyses for planning forest restoration activities are lacking.

The social-ecological systems (SES) framework proposed by Elinor Ostrom (Ostrom, 2009; Ostrom and Cox, 2010) has potential utility for diagnosing the local nuances of natural resource management and informing restoration projects. In this framework, the motivations for restoration are seen as being locally unique, analogous to the unique symptoms associated with a patient in medical practice. While individuals may exhibit similar symptoms, the treatments prescribed by a medical practitioner vary depending on the individual physiological attributes of the patient, such as age and blood pressure. Similarly, restoration activities are more likely to be effective if the planning process is based on a diagnostic style of investigation. This approach would seek to characterise the socio-political and

Chapter 6. Enhancing feasibility in restoration ecological context and provide insight into the opportunities and constraints that could influence the effectiveness of restoration activities.

Here we develop an analytical framework for operationalising a comprehensive and systematic approach to restoration planning that employs Ostrom’s SES framework in conjunction with methods for systematic decision-making (Figure 6.1). This approach enables priority areas for restoration to be identified by integrating information on ecological suitability, social feasibility and political permissibility. We demonstrate its application to restoration planning in Paser District in the province of East Kalimantan, Indonesia, where a recently developed community forestry program aims to deliver benefits for forest-dependent people (Sardjono et al., 2013).

6.3. Methods

6.3.1. Analytical framework

Chapter 6. Enhancing feasibility in restoration

6.3.2. Case study

Paser District has a total extent of 1.16 million hectares (Figure 6.2; Paser Statictics Service, 2014a). In 2013, approximately 256,000 people inhabited 144 villages within 10 sub-districts (Paser Statictics Service, 2014a). The population is represented by numerous ethnic groups including Paser indigenous people (Dayak Paser) who mainly live in forest frontier areas, and migrant communities such as those from Java (Javanese), Sulawesi (Buginese) and southern Borneo (Banjarese) who commonly reside in coastal zones and lower plains (Bakker, 2008; Belcher et al., 2004). “Dayak’’ is an anthropological term that refers broadly to indigenous ethnic groups living in the interior of the Island of Borneo who mainly practise cultivation (Levang et al., 2005). Dayak Paser are believed to be part of the wider Dayak Benuaq, who inhabit the upper Mahakam River and have been practicing complex rotational cultivation systems for centuries (Gönner, 2000; Saragih, 2011).

Chapter 6. Enhancing feasibility in restoration

Paser District was previously densely forested, but is undergoing rapid land-use change and economic transition from a forest-based economy to a more diverse economy including rapid development of oil palm plantation (Dewi et al., 2005; Gaveau et al., 2014b; Saragih, 2011). Logging operations in Paser began in the 1970s, involving 13 concessions at its peak (Saragih, 2011). Through logging however, many areas have depleted timber stocks resulting in degraded forest, which has led to these areas being converted to non-forested land uses mainly for oil palm plantation. However, some logged areas remain in degraded conditions and are not currently managed by any timber concessions (Ministry of Forestry, 2012b; Saragih, 2011). As of 2011, there were eight broad land uses within the Paser District (Figure 6.2) with this landscape including traditional communities who are dependent on forest resources and subsistence dryland agriculture (often referred to as agroforestry); as well as more modernised communities who employ various livelihood practises (Dewi et al., 2005).

6.3.3. Diagnosing the social-ecological systems that represent an opportunity for restoration

The goal of restoring degraded lands in order to support rural livelihoods (Figure 1) align with the Indonesia’s government vision that forest restoration programmes should enhance community empowerment and livelihoods (Article 42 par. 2 of Law No. 41/1999). The explanation (Penjelasan) of this law also states a paradigm shift in forest management from a timber focused one, to a view of deriving multiple forest benefits. This vision is supported by Article 48 par. 1 of Law No. 26/2007 on National Spatial Planning, which states that spatial planning for rural areas should maintain ecosystem services, conserve biodiversity and sustain food security.

Using background information about the planning region (Dewi et al., 2005; Paser Statictics Service, 2014a, b; Saragih, 2011), we apply Ostrom’s (2009) diagnostic approach to investigate social- ecological systems that represent potential restoration opportunities (Gregory et al., 2012; Moon et al., 2014). For this study, we focused on local tree-based agroforestry systems that are compatible with our stated goal, which is to restore degraded lands while providing livelihoods for local communities (Figure 6.1). Although we acknowledge other land management such as oil palm plantation could provide livelihood benefits, we do not include this as restoration opportunity in our study as it is unlikely to deliver ecological benefits (Carlson et al., 2013; Fitzherbert et al., 2008). The agroforestry systems we use herein, are locally known as awa pangeramu or simpukng and awa pangekulo or kebotn (herein called simpukng and kebotn as these are more popular terms)(Gönner, 2000; Saragih, 2011). The elaboration of Ostrom’s framework (i.e. Ostrom and Cox, 2010) is used to unpack the social-ecological systems of simpukng and kebotn into multi-scale subsystems (i.e. resource system, resource units, governance system, users, action situation and outcomes) to identify the contexts for restoration prioritisation (Figure 6.1; Supplementary Table A6.1).

Chapter 6. Enhancing feasibility in restoration

Simpukng and kebotn are integral parts of a wider system of agroforestry in Paser which has been practiced by Dayak Paser for more than 300 years (Gönner, 2000). Both resource systems are commonly established through enrichment planting in degraded forests or young secondary forests using seedlings and wildlings of various tree species grown alongside other naturally regenerating vegetation (Saragih, 2011). Simpukng forms a complex agroforest, containing more than 90 species of tree per hectare, mainly important fruits such as a variety of durian species (Durio spp), jackfruit (Artocarpus spp.) and rambutan (Nephelium spp.), creating a multi-layered canopy similar to natural forest (Crevello, 2003; Gönner, 2000). Kebotn generally has a simpler floristic structure and composition, as the vegetation is dominated by perennial crops such as rubber (Hevea brasiliensis) and coffee (Coffea spp) (Crevello, 2003; García-Fernández and Casado, 2005).

A variety of products generated from simpukng and kebotn provide a broad range of livelihood options in response to social and ecological uncertainties such as market dynamics, forest fires, prolonged droughts and pest outbreaks (Gönner, 2010). As access to markets and dispensaries in rural areas is limited, this form of land management also provides a variety of goods for subsistence needs. For example, agroforestry systems in rural Paser provide between 47 and 88 plant species for consumption, 38 to 56 species for medicinal purposes and between 15 to 23 species for traditional ceremonies (Saragih, 2011).

The governance system in Paser District (as in many regions of Indonesia) has developed from the interplay between informal and formal systems, represented by customary rules and state law, respectively. As agroforestry has been developed by Dayak Paser over many generations, this system has been integrated with their unwritten traditional customary practices (adat). Community matters, such as decisions and disputes related to land tenure, are discussed in a meeting (berinok) led by the community’s customary head (kepala adat) (Saragih, 2011). Trust and reciprocity are common, and group members are often willing to share labour resources voluntarily (gotong-royong) on farming- related works (Gönner, 2000; Saragih, 2011). However, many agroforestry lands in Paser and other areas in Indonesia are under legal ownership of the central government, and controlled by the Ministry of Forestry as part of the “Forest Estate” (Peluso, 1995; Saragih, 2011).

6.3.4. Identifying the context for restoration activities

6.3.4.1. Ecological context

There are two important ecological considerations for the establishment of simpukng and kebotn: biophysical suitability and compatibility with existing land cover (Table A6.1). Biophysical suitability depends on species-specific responses to factors such as topography, climate and soil

Chapter 6. Enhancing feasibility in restoration

(Ritung et al., 2007). The biophysical requirements for eleven of the most important tree species commonly cultivated in simpukng and kebotn in Paser were available (Table A6.2; Saragih, 2011). Eleven biophysical attributes of each site were used to classify suitability for each species into four categories: not suitable, marginal, moderate and high –for which ordinal values from 0 to 3 were assigned (Table A6.3; Ministry of Agriculture, 2012a; Ritung et al., 2007). Data on slope was generated from the CIAT-CGIAR digital elevation model v.4.1 based on the Shuttle Radar Topography Mission (Jarvis et al., 2008). Annual average temperature and rainfall data were obtained from the WorldClim database (Hijmans et al., 2005). Data on soil properties including drainage, soil texture, coarse material, base saturation and organic carbon were extracted from the Harmonized World Soil Database version 1.2 (FAO/IIASA/ISRIC/ISSCAS/JRC, 2012). Soil cation exchange capacity and pH were obtained from ISRIC (ISRIC – World Soil Information, 2013), while soil depth data were obtained from the World Resources Institute (Gingold et al., 2012). All datasets had 1x1 km spatial resolution, except for slope and soil depth, which was resampled to the 1x1 km resolution.

To identify land covers compatible with simpukng and kebotn agroforestry, we primarily used a land cover map derived from LANDSAT ETM+7 (Ministry of Forestry, 2012c). This map classifies land cover into 23 categories. To simplify, we merged some classes into broader categories; for example, the classes for irrigated paddy field (sawah), fish pond, freshwater swamp and mangrove were merged into “wetlands’’. In addition, locations of oil-palm and industrial timber plantations were identified from LANDSAT ETM and ALOS PALSAR data (Gaveau et al., 2014b). In the final map, we defined compatible land cover types as logged forest, severely degraded forest/bush (semak belukar) and mixed garden (kebun campur), and excluded primary forest, wetlands, oil-palm plantation, industrial timber plantation, settled and cleared areas, and mines.

6.3.4.2. Socio-economic context

We identified four important socio-economic characteristics for simpukng and kebotn, for which spatial information could be collated and synthesised (Table A6.1): 1. The communities’ economic dependency on agroforestry systems, measured by the prevalence of agroforestry as the main occupation. 2. The communities’ cultural dependency on agroforestry systems, expressed as the use of non-timber forest products (NTFPs) to support subsistence needs and traditional customs. 3. The communities’ preference regarding the role of agroforestry compared to other competing land uses. 4. The location of simpukng and kebotn.

Chapter 6. Enhancing feasibility in restoration

Economic dependency on agroforestry was calculated using the proportion of the village workforce whose primary occupation was in the agroforestry sector (Dewi et al., 2005), based on 2010 national census data (Indonesian Bureau of Statistics, 2014). This dataset differentiates occupational sectors into 19 categories. Agroforestry is not defined as a distinct sector and thus we employed a combination of the agriculture, horticulture and plantation sectors (Paser Statictics Service, 2014b). Forestry was excluded as this sector is associated specifically with jobs in timber industries such as logging concessions and wood processing. As the plantation sector statistics relate to a combination of oil palm and agroforestry (e.g. rubber and coconut), the proportion of the workforce employed in agroforestry plantations was estimated using the extent of agroforest relative to the total extent of plantations (Paser Statictics Service, 2014a). We calculated an adapted version of the Importance Value Index that is commonly used in ecology (Cottam and Curtis, 1956) to obtain an index ranging from 0 (no workers in the agroforestry sector) to 1 (all workers in the agroforestry sector).

Cultural dependency on agroforestry was estimated by the proportion of people in each village who use agroforestry products, assuming that enrichment planting of 11 tree species described above would also generate agroforestry products in a broader sense (i.e. various NTFPs including fruits, vegetables and medicinal plants; Saragih, 2011). These estimates were provided by Abram et al. (2014) who mapped local villagers’ perceptions on forest ecosystem services and land cover changes using interview data for 1837 individuals within 185 villages in Kalimantan conducted between 2008 and 2012. Spatial predictions of these perceptions throughout the landscape used 39 social-economic and ecological predictors. For our analysis, we incorporated a map representing the use of a combined set of 29 NTFPs, which included products such as small/fire wood, rubber, fruits and vegetables. Specifically, this map classified the use of at least one or more of these NTFPs into three classes: low, moderate and high.

Among communities in Paser District, there are conflicting opinions on land management between maintaining traditional agroforestry system or converting their lands to oil palm (Belcher et al., 2004; Saragih, 2011). To represent communities’ preference for agroforestry over other land uses we used spatial estimates of villagers’ negative perceptions of large-scale clearing for agriculture (e.g. oil- palm plantation) and forestry (e.g. industrial timber plantations) (Abram et al., 2014) and classified this perception into three levels: weak, moderate or strongly negative.

To allow for movement of labour and harvested goods, both simpukng and kebotn are generally established in the vicinity of transportation networks, such as roads and rivers (Gönner, 2000), but not necessarily in close proximity to towns. This is because the commercial products, such as dried rubber latex and coffee seeds, can be prepared locally and do not require sophisticated post-harvest

Chapter 6. Enhancing feasibility in restoration technologies and processing factories (Saragih, 2011), unlike oil palm fruits (Meijaard and Sheil, 2013). Instead, in the case of simpukng and kebotn, traders from nearby towns regularly visit villages or local markets to purchase crops (Saragih, 2011). We mapped the proximity to transportation networks as the Euclidean distance to roads (including logging and village roads) and rivers, using data obtained from the Development Planning Agency (Bappeda) of East Kalimantan. We classified distance from transportation networks into four classes (Gönner, 2000): 0-0.5 km; 0.5-1 km; 1-2 km, and >2 km.

6.3.4.3. Political context

There are two possible legal contexts for the establishment of simpukng and kebotn: one on lands outside the Forest Estate (Area Penggunaan Lain/APL), and the other on state lands inside the Forest Estate and under the institutional arrangement of community forest (hutan kemasyarakatan/HKm) (Table A6.1). In community forest, a farmer group is given rights to manage and utilise forest resources over a 35 year contract which can be extended every five years when the contract expires (Ministry of Forestry, 2007). However, community forest involving tree planting, such as required for establishing simpukng and kebotn, can only be granted in areas designated as Production Forest (i.e. Forest Estate designated to generate forest products such as logging concession) without an active concession.

To identify areas legally allowable for establishing simpukng and kebotn, we employed the forest classification map (Ministry of Forestry, 2012a) to delineate the Forest Estate. The Forest Estate was then further differentiated into Production Forest and protected areas (which consist of Protection Forest for hydrological protection and Conservation Forest for biodiversity conservation such as national park). A map of forest concessions (Ministry of Forestry, 2012b) was used to delineate areas of Production Forest without an active concession.

6.3.5. Opportunity costs of restoration

We defined costs as the profits forgone from alternative land uses (i.e. opportunity cost bear by competing land-based industries eminent in Paser District) if a land parcel is managed for simpukng and kebotn (Wilson et al., 2012). We considered two competing land uses when estimating the opportunity cost: oil palm plantations and logging. We assumed that if a land parcel is suitable for oil palm, the opportunity cost is calculated as the net present value (NPV) of oil palm plus additional revenues generated from harvesting timber during land clearing. Conversely, if a land parcel is not suitable for oil palm, we assumed the opportunity cost is equal to the value of timber that would be obtained either through selective logging (of existing logged forest) or clear cutting (of existing mixed

Chapter 6. Enhancing feasibility in restoration gardens). We used data from Irawan et al. (2013) as a baseline NPV for oil-palm plantation (i.e. US$6,355 per hectare). We adjusted the NPV of oil palm according to land suitability assuming that highly suitable land would produce full yield (i.e. 100% NPV), while moderately and marginally suitable land would generate 75% and 50% of full of the NPV, respectively. Land suitability for oil palm was mapped using a similar method as for the agroforestry species described above. We estimated the NPV of timber extraction for each compatible land cover using data from Ruslandi et al. (2011). For logged forest, the NPV of logging was adjusted to US$1,292 per hectare as commercial timber volumes in already-logged forests in Kalimantan are in average 57% of unlogged areas (Putz et al., 2012). For areas that are currently under mixed gardens, a conservative opportunity cost of US$300 per hectare was assigned (assuming smallholder timber extraction since large-scale commercial logging is unlikely to be feasible), while for severely degraded forest we assumed no timber of value remains (Budiharta et al., 2014a).

6.3.6. Prioritisation analysis and planning scenarios

We assigned an ordinal value for biophysical suitability, cultural dependency, community preference and accessibility; and an index for economic dependency (continuous over the range 0–1). We employed the decision support tool Zonation v.4 (Moilanen et al., 2009a) to produce a ranking of priority areas for restoration (through the establishment of simpukng and kebotn agroforestry). We prioritised land parcels with the highest aggregate values for social feasibility (across four social variables) and biophysical suitability (across 11 fruit and perennial crops) while minimising opportunity cost (Moilanen et al., 2009a). All spatial inputs had a 1x1 km grid size to produce planning unit with spatial extent of 100 hectares, reflecting the scale of land management systems (e.g. community forest), and the resolution of the spatial datasets used to generate the social and biophysical suitability layers.

We tested four scenarios to explore the influence of different socio-ecological perspectives on restoration priorities, and to investigate the trade-offs in term of social and ecological benefits gained, and costs incurred under each scenario. In Scenario 1, we accounted only for biophysical suitability layers. We incorporated both social and biophysical suitability layers in Scenario 2, with equal weight for these two classes of information. In Scenario 3, we incorporated the political context by assuming that priority areas for restoration are constrained by current legal systems for land use. For this scenario, areas that occur in either the non-Forest Estate or in Production Forest without active concessions were prioritised over areas in Production Forest with active logging concessions and protected areas. In Scenario 4, we allowed agroforestry restoration on all Production Forests, disregarding the location of active logging concessions. For this scenario, we assumed that if priority

Chapter 6. Enhancing feasibility in restoration areas occur in active concessions, the land use would need to change from a logging concession to a community forest.

For each scenario, restoration sites were selected as the top 20% of potential areas in the ranked prioritisation. For each set of selected sites, we calculated the value for each variable (e.g. biophysical suitability of a specific crop, index of economic dependency on agroforestry) as a proportion of its summed value across the areas potentially available for restoration, and the overall value across all variables in each class (social or ecological) was calculated as the mean of these proportions. We also calculated cost-effectiveness of each scenarios formulated as total cost required if the summed values of all variables in each social or ecological class was retained.

6.4. Results

We identified 515,300 hectares of land in Paser District that could potentially be restored for agroforestry to achieve our goal. These areas are suitable for at least one of 11 the fruit/crop types of commonly employed in simpukng and kebotn for which suitability criteria are known, and the present land cover is compatible (Figure 6.3). ). The top 20% priority sites (103,060 hectares) under Scenario 1 (considering only biophysical criteria) had the overall greatest biophysical suitability but low social feasibility (Figure 6.4). This is due to the selection of areas in the eastern part of the region that were adjacent to oil-palm plantations (Figure 6.3a) which have a low cultural dependency on agroforestry and high economic dependency on the oil-palm sector. Consequently, the areas selected under Scenario 1 incurred the greatest opportunity cost, equating to US$605 million. Consequently, the areas selected under Scenario 1 incurred the greatest opportunity cost, equating to US$605 million (Figure 6.4). The areas selected also encompassed at least 8,500 and 1,500 hectares of Conservation Forest and Protection Forest respectively.

When incorporating both social and ecological variables concurrently (Scenario 2), priority areas showed a more distributed pattern, with many high priority areas in the central and south-west portions of the Paser District (Figure 6.3b). The overlap in land areas selected in Scenarios 1 versus 2 was only 54.9% (56,600 hectares) of top 20% priority areas, and these areas were primarily in the central-west of the district. As expected, the integration of social variables increased the social feasibility (Figure 6.4) and the areas selected had higher economic and cultural dependency on agroforestry (Figure 6.3b). Conversely, the biophysical suitability of the areas selected was lower than for other scenarios (Figure 6.4), mainly due to the selection of extensive areas in the southwestern tip of Paser District. Scenario 2 incurred the lowest opportunity cost of US$480 million, which was approximately 20% less than Scenario 1. Under this scenario, at least 1,400 and 12,200 hectares of prioritised areas occurred in Conservation Forest and Protection Forest areas respectively, 90

Chapter 6. Enhancing feasibility in restoration reflecting the high level of cultural and economic dependency on Protection Forests (e.g. for collection of non-timber forest products).

Under Scenario 3, both social and ecological variables were incorporated, but the selection was constrained to areas where restoration was legally permitted (i.e. non-Forest Estate/APL, or Production Forest without an active logging concessions; Figure 6.3c). As a result, only 46,500 hectares (45.1%) of the priority areas under Scenario 2 were selected under Scenario 3. Scenario 3 also had low social feasibility and incurred a higher opportunity cost (US$598 million; Figure 6.4) in part due to the selection of areas in the central and eastern part of the region where the opportunity cost for the oil-palm sector is high. Under this scenario, at least 23,500 hectares occurred in Production Forest without an active concession, and could therefore be allocated for community forest to establish simpukng and kebotn.

Scenario 4, where we assumed a legal framework that allows agroforestry restoration in any area of Production Forest (regardless of concessions), delivered the best outcome across ecological, social and cost criteria. This would offer greater social feasibility, and a reduction in opportunity costs in the order of US$70.6 million (11.8%) compared to Scenario 3 (Figure 4). Under this scenario at least 10,800 hectares occurred in Production Forest with no concessions, while 49,800 hectares overlapped with active logging concessions. Many of these overlapping areas occurred on severely logged forests with limited opportunity costs for timber harvesting.

Chapter 6. Enhancing feasibility in restoration

Figure 6.3. Maps of priority areas for agroforestry restoration under four social-ecological scenarios: (a) Scenario 1 (ecological variables only); (b) Scenario 2 (ecological and social variables); (c) Scenario 3 (ecological and social variables with existing political constraints); (d) Scenario 4 (ecological and social variables with constraints associated with land use removed). The coloured areas show the potential areas for agroforestry restoration from the ecological context (i.e. areas suitable for at least one fruit/crop type, and where the land cover is compatible).

Chapter 6. Enhancing feasibility in restoration

Figure 6.4. Outcomes for each scenario in regards to biophysical suitability, social feasibility and opportunity cost for the highest priority areas for agroforestry restoration (areas within the top 20%). For each class of values (biophysical or social), the aggregate value is the mean proportion of the landscape’s summed potential value that is encompassed by the selected restoration areas. The proportion of total landscape value was calculated separately for each variable in each class, and then the mean taken across all variables within the class.

Figure 6.5. Cost-effectiveness across four scenarios formulated as total cost required to retain total values of (a) crop suitability and (b) social variables, when a given proportion of the available landscape is selected (X axis).

Scenario 1 was the most cost-effective in selecting areas with high biophysical suitability (Figure 6.5a), while Scenario 2 delivered the greatest returns in relation to social feasibility (Figure 6.5b). Differences between scenarios were greatest in terms of social feasibility (i.e. there were large

Chapter 6. Enhancing feasibility in restoration differences in the costs per unit of social feasibility gained). Constrained by the existing political context, Scenario 3 incurred the greatest cost at the point when all legally allowable areas (i.e. on non-Forest Estate or in Production Forest without active concessions) have been selected (Figure 6.5b: the dashed line).

6.5. Discussion

Existing studies on assessment and systematic planning for restoration emphasise biophysical aspects (Birch et al., 2010; Budiharta et al., 2014a), with much less attention given to the social characteristics of the region (McElwee, 2009). We proposed an integrated framework to diagnose socio-political and ecological context of a region and to employ the diagnostic tool to inform the development of forest restoration plans. We discovered that the incorporation of socio-political and ecological contexts in restoration planning altered the selection of priority areas for restoration (Figure 6.3). This integration revealed trade-offs between the potential social and ecological outcomes, as well as the opportunity costs incurred (Figure 6.4). Our analysis also provides policy direction to Indonesia’s government in the context of allocating lands in the Forest Estate to “community forests” with the aim of delivering socio-economic benefits for forest-dependent communities.

We found that considering biophysical criteria alone (Scenario 1) would deliver the greatest ecological benefits in term of biophysical suitability for crops and fruit trees. Priority areas under this scenario had the highest aggregated suitability across 11 species, representing greater opportunities to plant diverse species and attain higher growth rates and yields. However, as the social feasibility was comparatively low, there is a greater risk that agroforestry restoration on these prioritised areas may be unsuccessful in the long term. This failure could arise from a preference for alternative livelihood options, such as in the oil palm or mining sectors, rather than planting and maintaining trees. The planted areas would then likely be converted into oil palm plantations, if communities are in favour of or do not oppose these developments. Many examples exist of restoration programmes being jeopardised because the main actors (e.g. local communities) are either reluctant to implement the prescribed activities, or they undertake activities that are counterproductive to the overall goal even opposing the programmes (Boedhihartono and Sayer, 2012; Nawir et al., 2007b; Wuethrich, 2007).

Analyses by Le et al. (2014) in the Philippines indicate that community dependency on forest, either as a main income source or to fulfil subsistence needs, has a strong positive relationship with the success of reforestation. In our analysis however, incorporating the variable of community dependency on forest aspect along with two social variables (i.e. community preference on land use and agroforestry location) in order to increase feasibility (Scenario 2) would be at the expense of the 94

Chapter 6. Enhancing feasibility in restoration biophysical suitability of restored areas. This implies that in some priority areas, such as in the southwestern tip of Paser District, rural farmers would have fewer tree species available for planting that could achieve maximum growth or productivity. Furthermore, despite having the greatest social feasibility and the lowest opportunity cost, agroforestry restoration would be legally prohibited and undesired in some areas (e.g. protected areas) and likely to lead to land use conflict.

While the political context is often neglected in restoration studies (Baker et al., 2013), we discovered that integrating the existing legal framework changed the priority areas for restoration, and incurred substantially greater opportunity costs. Despite its poorer performance, Scenario 3 was closest to the actual situation of the current social-ecological system, and so its application would likely entail the smallest legal and political ramifications, for example by avoiding lengthy constitutional amendments, and potential tenure disputes and social conflicts (Baker et al., 2013). Our analysis reveals that current policies may lead to inefficient restoration programme design, because regulations for Forest Estate lands only allow agroforestry restoration (in the form of community forest) to occur in Production Forest areas without active concessions. Our results suggest that re-allocating management of areas with logging concessions and high social-cultural values to community forests (Scenario 4), would increase the cost-effectiveness of agroforestry restoration. As these areas occur mainly in severely degraded forests, the transfer of management would incur minimal forgone opportunity costs from timber harvesting, and this could mitigate potential opposition from large- scale logging concession holders (Saragih, 2011). Inefficiency in land-use policies is not restricted to Paser District as it also occurs across Kalimantan (Law et al., 2014; Runting et al., 2015; Sumarga and Hein, 2015). In reality however, changing land-use systems in the region would be challenged by current political practices, which puts great emphasis on patronage and favouritism toward elites, rather than fair and balanced distribution of benefits across stakeholders (Faisal, 2013). For example, severely logged forest in some regions is being speculatively held by timber concessionaires for future opportunities that may arise (e.g. land banking for oil palm plantation or mining; Kartodihardjo and Supriono, 2000).

In our case study, agroforestry restoration using indigenous resource systems offers an opportunity to restore degraded lands in the Paser District. Similar approaches in the Philippines with institutional arrangements for community forestry show promising results, covering 37% (5.9 million hectares) of total state forest lands with more than 4,800 community groups involved (Poffenberger, 2006). This strategy has contributed to a steady increase of forest cover (i.e. 0.75% annually) in the Philippines between 1990 and 2010 (Food and Agriculture Organisation, 2011). The Indonesian Government has set a target for designating two million hectares of their Forest Estate as community forest by 2014 but only 438,000 hectares have been allocated so far (Ministry of Forestry, 2014; Sardjono et al.,

Chapter 6. Enhancing feasibility in restoration

2013). Our planning framework provides guidance on how such a policy target could be achieved, for example, in Paser District this could involve allocating 23,000 hectares of Production Forest without active logging concession to community forest. Further government support could be in the form of providing funds (e.g. using the Reforestation Fund and national forest rehabilitation programme/GERHAN) for communities to undertake tree planting and maintenance using performance-based funding mechanisms (Government of Indonesia, 2002, 2008). This financial support is important, because there would be time lag for the farmers to generate economic benefits from the fruits or crops planted (Nawir et al., 2007b).

To provide more realistic and meaningful solutions, we would ideally incorporate all of the social variables proposed in Ostrom’s diagnostic approach, such as leadership and network structures among stakeholders (e.g. Guerrero et al., 2015), to complement the four social variables employed in this study. The presence of well-coordinated networks involving communities, local forestry service and academic institutions proved to be an important factor for the success of a reforestation programme using teak species in central Java, Indonesia (Nawir et al., 2007b). We were not able to include such variables in our analysis, due to limited availability of this information in a data-poor region, especially in a spatial format.

We demonstrated the operationalisation of our analytical framework using a single restoration goal and a specific social-ecological system (i.e. framed around agroforestry) as a restoration strategy for illustrative purpose. However, multiple restoration goals along with different restoration strategies that interact with related social-ecological systems (e.g. oil-palm plantations and logging concessions) could be incorporated into the framework. This is feasible, pending available data on social and political aspects, to simultaneously incorporate various goals related to the conservation of wildlife habitat and enhancement of carbon stocks using, for example, native species plantings and natural regrowth as restoration strategies, and ecosystem restoration concessions as an institutional arrangement (Budiharta et al., 2014a).

We quantified restoration benefits by assigning an ordinal scale for some variables (e.g. biophysical suitability of agroforestry species, cultural dependency, and accessibility to transportation networks), this judgement does not necessarily represent values held by all stakeholders. As such, we suggest that for more transparent and defensible implementation, stakeholders’ values be captured during the planning processes, for example through participatory workshop processes (Gregory et al., 2012). Several rounds of learning processes during the workshops might be required to familiarise stakeholders with our analytical framework and to find consensus among them to yield sound decisions that have wide acceptance (Boedhihartono and Sayer, 2012). Nonetheless, empirical testing

Chapter 6. Enhancing feasibility in restoration in a range of contexts and case study systems is required to ensure broad applicability and transferability of our framework.

6.6. Conclusion

Understanding the social-ecological context of a restoration programme is likely to have a strong impact on its effectiveness, especially in socio-culturally heterogeneous regions such as Indonesia. While social and political systems are often viewed as complex, it is possible to explicitly incorporate them into systematic planning for restoration by utilising a social-ecological system framework and systematic decision making. Despite the potential trade-offs between the social feasibility and ecological benefits of restoration, we demonstrate that this integration results in more realistic and feasible solutions. There are, however, challenges in translating existing social variables into relevant and accurate spatially-explicit representations, especially in data-poor regions such as in tropical developing countries.

6.7. Acknowledgements

We thank Enrico Di Minin for discussions on Zonation and Muhammad Yusuf for providing occupational statistics data. SB was supported by an Australian Awards Scholarship and Australian Research Council Centre of Excellence for Environmental Decisions (ARC-CEED). KAW was supported by an Australian Research Council Future Fellowship. EM was supported by an Arcus Foundation grant. JW was supported by the ARC-CEED.

Chapter 7. Discussion

Chapter 7

General Discussion

No publication associated with this chapter

Chapter 7. Discussion

7. General Discussion

This thesis advances the application of decision science frameworks to guide the restoration of degraded tropical forests to enhance the delivery of ecosystem services. My thesis addresses critical questions on how we can make better decisions for landscape-scale restoration by: (a) addressing landscape heterogeneity in terms of degradation condition, which determines the appropriate restoration action, the cost of these actions, and the benefit that will be accrued through time; (b) leveraging restoration within competing land uses using emerging policy for offsetting; and (c) enhancing feasibility by accounting for the social and political dimensions related to restoration. In this chapter, general conclusions are drawn and several limitations are identified, which could provide useful avenues for future research.

7.1. Identifying contextual ecosystem services for restoration

Ecosystem services pertinent to a region and to stakeholders are often defined by values and knowledge systems (Díaz et al., 2015a; Díaz et al., 2015b). We might consider that an ecosystem service is important for our region but it may not be the case for other areas (e.g. religious values held relating to a sacred forest). Even within one region, the value of one ecosystem service can be perceived differently between stakeholders. For instance, based on empirical evidence (e.g. Phalan et al., 2011) we might jump to a conclusion that traditional agroforestry systems practised by an indigenous community are an old-fashioned way to manage land for provisioning services. We might also perceive that agroforestry systems should be replaced by modern intensive-agriculture (e.g. oil- palm plantation) to generate greater monetary benefits and that to deliver better conservation outcomes some lands should be spared for biodiversity protection. However, the community might not accept this argument due to its own socio-cultural perceptions, which do not fit with the economic and conservation motives we value. In this regard, the commonly believed ideology that science is value-free may not always be applicable in ecosystem services studies.

Ecosystem services considered in my thesis were identified from a workshop attended by scientists and policy-makers, and seeking to capture context-specific values (“Development, Environment and the People of Kalimantan”, Banjarmasin, South Kalimantan, November 2012). In this workshop, it was argued forest loss and degradation were the major environmental problems in Kalimantan causing significant impacts on biodiversity and local communities’ livelihoods. Chapter 2 of my thesis (Budiharta and Meijaard, in press) confirmed that habitat loss, primarily due to logging and forest conversion for oil-palm plantation, was the most important threat to biodiversity in Kalimantan, affecting more than 82.1% of threatened animal species and 60.5% of threatened plant species.

Chapter 7. Discussion

Chapter 2 also indicated that deforestation and forest degradation was likely to affect local livelihoods. Various fruit and crop trees are managed in agroforestry systems and the associated products such as durian, honey, rubber and medicinal plants provide essential food and income sources for local communities. As shown in Chapter 5, 0.78 Mha (16.8%) of oil-palm plantations in Kalimantan have replaced agroforestry lands. Understandably, this conversion leads to social conflict between oil-palm companies and indigenous people (Abram et al., in review). Thus, the conservation of biodiversity (Chapter 4 and 5) and the sustained provision of livelihoods (Chapter 6) were a strong focus of this thesis.

Another ecosystem service important in Kalimantan is carbon storage. This arises not only in the context of global climate but also due to the problems associated with fire and haze, which cause severe health and well-being impacts in a regional context (Betha et al., 2013; Brown, in press; Resosudarmo, in press). Existing strategies for carbon emissions reduction in Indonesia through REDD+ have been focused on the protection of primary forests and peatlands (Murdiyarso et al., 2012; Murdiyarso et al., 2011; Venter et al., 2013). Yet, protecting peatlands without additional management (e.g. hydrological rehabilitation, reforestation) will not reduce emissions due to the continual occurrence of fires on degraded peatlands (Gaveau et al., 2014a). Addressing carbon emissions through restoring forests and peatlands in Kalimantan will make an important contribution to the achievement of the national target (Budiharta et al., 2014a; Busch et al., 2015; Koh et al., 2011). For example, mitigating carbon emissions from industrial oil-palm plantations in Kalimantan through restoration offsets alone could contribute an emissions reduction of 11% of Indonesia’s total emissions from land use change at relatively low cost (Chapter 5).

While I have attempted to identify three key ecosystem services in Kalimantan, there are other important ecosystem services excluded from this thesis. For example, flooding impacts 65% of villages and 1.5 million people in Kalimantan and there is evidence of increasing flood events associated with the loss of forest cover (Wells et al., 2013). The occurrence of severely degraded forests and forested watersheds are found to be important predictors of flooding in the region (Wells et al., 2013). Reforestation programs have been in place in Kalimantan, yet there is a lack of a systematic approach to explicitly address flooding (Wells et al., 2013). Including hydrological services as a restoration benefit would expand the significance of my thesis in the region, for example restoration could be targeted to restore degraded areas surrounding water recharge and discharge areas (Soejono et al., 2013). However, there will be challenges associated with modelling the benefit of forest restoration for hydrological regulation as current knowledge in the field of forest hydrology is limited in tropical regions (van Dijk and Keenan, 2007; Van Dijk et al., 2009). Further, this analysis

100

Chapter 7. Discussion will require fine resolution climate, terrain and soil data, which is lacking for Kalimantan (van Dijk and Keenan, 2007).

7.2. Restoration planning in data-poor regions

One common problem when working in data-poor regions is the limited tools and information available to support the analyses. Often, regions of global importance from a particular environmental perspective are the most understudied (e.g. Indonesia in the context of biodiversity conservation; Meijaard et al., 2015; Wilson et al., 2016). Despite the vital roles of Kalimantan for REDD+ policy implementation (e.g. Murray et al., 2015), a limited number of models have been developed to estimate the potential carbon sequestration benefits from forest restoration. In Chapter 3, I assessed the utility of the 3-PG (physiological principles for predicting growth) model to estimate the above- ground biomass (AGB) of Bornean forests. My findings provide evidence for the use of a simple yet robust carbon model for informing the development of restoration policies and practices under REDD+ in data-poor regions.

In the context of Borneo, and Indonesia and Malaysia more broadly, 3-PG could be applied to inform the development of Ecosystem Restoration Concessions (ERC) by facilitating the quantification and trading of carbon that is sequestered (e.g. FACE the Future, 2011; Indriatmoko et al., 2014). The relative simplicity in computational operations and data requirements offers practicality compared with more complex carbon models developed for tropical forests (e.g. Huth et al., 1997; Phillips et al., 2003). The developers of ERCs (and REDD+ projects) could run the 3-PG model using readily available data to calculate the accumulated carbon during the project timeframe. For broader application, the model could be trialled in other regions, such as Sumatra and Sulawesi, but this assessment would require field data for parameterisation and validation.

Similar efforts to reveal information and methodological tools for restoration planning in tropical regions should be expanded to other fields of research. For example, the lack of information on specific-species responses to land-cover change was problematic in determining restoration benefits in terms of habitat suitability for wildlife (Chapters 4 and 5). Ideally, dynamic models (e.g. Santika et al., 2015a; Thomson et al., 2009; Vesk et al., 2008) could be employed to predict the trajectory of biodiversity recovery during restoration, which have been rare in most tropical species except for charismatic mammals such as orangutan (Wilson et al., 2014). More studies are needed to reveal similar information for other species, at least for species with greater risk of extinction.

The problems I encountered in Chapter 6 due to limited spatial data for key social variables reveals the urgent need for more concerted effort to generate information on the social aspects of restoration.

101

Chapter 7. Discussion

This is a challenging task as restoration ecology is largely rooted in the natural sciences (Hobbs and Norton, 1996). Nonetheless, as the advancement of conservation sciences has been evolving toward a multidisciplinary approach (Groom et al., 2006; Kareiva and Marvier, 2012), knowledge in social and economic themes is likely to improve in the future. To achieve this, inclusive research collaboration with social scientists should be widely promoted to capture social and political information relevant to restoration.

The development of new methodologies using limited information necessitates creativity, and this is particularly the case for emerging disciplines such as restoration ecology and ecosystem services. Nevertheless, analyses conducted in pioneering fields in data poor locations inspires the generation of further knowledge and motivates the collection of new information to fill knowledge gaps.

7.3. Research contributions: advancing decision science in landscape-scale restoration to inform policy

7.3.1. Accounting for landscape heterogeneity in restoration planning

One of the unresolved questions in restoration ecology is how to address the heterogeneity of the landscape context with regard to plausible restoration strategies, associated costs and the delivery of benefits. In Chapter 4, I developed a novel framework for systematic restoration planning that considered a broad range of landscape conditions of tropical degraded forest. I demonstrated how landscape condition could be defined, plausible restoration actions could be assigned, restoration cost could be calculated and temporally explicit restoration benefits could be estimated. Employing a newly developed decision-support algorithm, I provided guidance for prioritising restoration investments with multiple objectives. My study should change the widely accepted perspective that views a degraded forest as a wasteland, which is better converted to other land uses (e.g. oil palm plantation; Santika et al., 2015b; Smit et al., 2013). I show that not all degraded areas are equal, with some degraded forests (e.g. moderately and highly degraded forest) important areas for restoration, while others (e.g. deforested/critically degraded forest) may be suitable for conversion to productive land uses.

Chapter 4 advanced the existing literature on the application of decision science in restoration ecology for the delivery of ecosystem services (e.g. Bryan and Crossman, 2008; Crossman and Bryan, 2009; Egoh et al., 2014; Orsi et al., 2011; Renwick et al., 2014). These studies generally assume that the landscape condition is homogeneous, which results in uniform restoration strategies being implemented across the landscape. A correctly defined measure of landscape condition used in restoration planning will affect how restoration actions are developed and how costs and benefits are 102

Chapter 7. Discussion calculated, which could alter decision outcomes (Evans et al., 2015; Maron et al., 2013). My research has revealed that a failure to consider landscape heterogeneity is likely to produce sub-optimal solutions, which are less cost-efficient (Evans et al., 2015; Kotiaho and Moilanen, 2015). Constrained by available resources (e.g. lands required for other land uses and budget), restoration should target certain degradation conditions that deliver the greatest restoration benefits.

In Chapter 4, I provided an example of how multiple ecosystem services (i.e. carbon sequestration and habitat for biodiversity) were incorporated and scenarios were developed to reflect heterogeneity in term of objectives, values and beneficiaries (Díaz et al., 2015a; Díaz et al., 2015b). I demonstrated that maximising the delivery of one ecosystem service would forgo that of other elements. By incorporating multiple objectives simultaneously, I discovered that compromised solutions were possible that achieved minimal trade-offs among benefits. While trade-off and synergy issues have been extensively discussed in the ecosystem services literature (e.g. Chan et al., 2006; Egoh et al., 2011; Howe et al., 2014; Law et al., 2014; Nelson et al., 2009; Runting et al., 2015; Venter et al., 2009a), these issues are rarely investigated in restoration ecology. My research has progressed the field of restoration planning beyond a static approach, which assumes the state of ecosystem services remains constant through time, to accounting for the reality that the delivery of services changes as restoration proceeds. An innovative aspect of my framework is that I considered the time-dependent delivery of ecosystem services as a result of restoration, which determines the nature of the trade-offs incurred (Pouzols et al., 2012).

In the context of REDD+, the results of Chapter 4 indicated that in achieving carbon objectives, restoring highly degraded forests is preferable to restoring moderately degraded forests, which are more intact in terms of their degradation condition. This intriguing finding leads to a critical question of whether, in Kalimantan, restoring degraded forest might be more cost-effective in accruing carbon benefit than protecting primary forest, owing to the relatively low restoration costs of planting and maintenance compared with the high opportunity cost of standing timber under a protection-focused strategy. This hypothesis needs to be tested to investigate the best strategy for REDD+ in Kalimantan, either through forest protection (Venter et al., 2009b), sustainable logging (Griscom et al., 2014) or forest restoration (Budiharta et al., 2014a), or a combination of the three. Surprisingly, this integrated analysis has not been conducted despite extensive studies on REDD+.

In Chapter 4, I defined landscape heterogeneity from an ecological perspective to identify priority areas that would deliver the greatest carbon and biodiversity benefits. Yet, questions arise: instead of providing support, what if local communities who live in close proximity to the restored areas oppose the restoration as they perceive that carbon sequestration will not benefit them and restoration will

103

Chapter 7. Discussion reduce their access to forest? What if the priority areas occur on lands that have been allocated to logging concessions, which may trigger conflict between the restoration project and the existing logging company? These questions imply that socio-political dimensions need to be considered in restoration planning to enhance the likelihood of success. In Chapter 6, I demonstrated how socio- political heterogeneity, when making decisions for restoration, could be addressed.

The impact of heterogeneity in costs was not comprehensively resolved in Chapter 4. I defined opportunity cost only as revenues forgone from one form of land management: timber extraction, and ignored the other prevalent land uses in the region including oil-palm plantations (Carlson et al., 2013). Certain degraded areas might be preferable for the development of oil-palm plantations due to high biophysical suitability, which is largely influenced by climate and soils (Gingold et al., 2012; Ritung et al., 2007; Smit et al., 2013). Areas with higher suitability potentially produce greater palm oil yields and profits (Gingold et al., 2012)—assuming capital inputs are the same across all suitability levels—which increases the opportunity cost of restoration. Accounting for the profits from oil-palm plantations is likely to change priority areas for restoration and this motivated Chapters 5 and 6.

7.3.2. Leveraging restoration through offsetting policy

Restoration poorly competes with other forms of land management. In Chapter 5, I investigated the functionality of restoration offsetting to leverage restoration in a rapidly-changing forest-frontier landscape under pressure from oil-palm plantation expansion. I calculated the losses of carbon and biodiversity from industrial oil-palm plantations, determined potential areas for restoration along with the costs and benefits, and identified cost-effective areas for restoration to offset such losses. I discovered opportunities to offset carbon emissions by focusing on restoration in degraded peatlands with deep peat thickness including the Ex-Mega Rice Project in Central Kalimantan. In contrast, offsetting biodiversity losses would be constrained by the high cost incurred and the vast extent of the areas involved, which also face competing interests from mining, oil-palm plantations and logging. I recommend that offsets can be directed to compensate carbon emissions to help solving the severe ecological, economic and social problems associated with peat fires in the region.

My study is the first to assess environmental offsetting in an agri-environmental context at a large spatial extent. Existing studies have been focused on the mining sector and urban development, which typically require less lands than the agriculture sector and have greater economic capacity to adopt offsetting policy (Gordon, 2015; Kiesecker et al., 2009b; Madsen et al., 2010; Sonter et al., 2014). My conservative estimates indicated that the extent of offsetting areas required to compensate

104

Chapter 7. Discussion biodiversity losses from oil-palm plantations is equivalent to the impacted sites (i.e. offset ratio of 1:1). As the agriculture sector is generally less profitable, often requiring subsidies, its financial capacity for biodiversity offsetting is likely limited. To improve the capacity of the agriculture sector, markets could assist by putting a premium price for agricultural products that implement biodiversity offsets (Dinerstein et al., 2013). I suggest that future analysis could investigate the price rise required and explore whether consumers would be willing to pay such a price.

If leveraging biodiversity offsets using a premium-price is not achievable, I propose an alternative strategy via corridor offsetting. This strategy assumes that biodiversity offsets could potentially deliver greater biodiversity benefits if offsets were targeted to connect protected areas although the offset areas will be smaller than required in a full compensation scenario. The corridor approach has potential for regions such as Kalimantan as many existing protected areas in lowland zones are isolated and at risk when temperatures increase due to climate change, shifting habitat suitability from lowland to upland areas (Struebig et al., 2015b). Therefore, offsetting corridors could potentially serve as refuge sites for wildlife, linking lowland and upland protected areas (Scriven et al., 2015; Struebig et al., 2015b). I recommend that this analysis is a priority for upcoming research.

The retrospective framework of carbon offsets presented in Chapter 5 provides useful insights for global carbon policies, as recently enacted in the Paris Agreement in Conference of the Parties 21. Article 3.2 of the agreement states “Parties may take action to implement and support, including by scaling up resources, policy approaches and positive incentives for reducing emissions from deforestation and forest degradation…as well as alternative policy approaches, such as joint mitigation approaches for the integral and sustainable management of forests” (http://unfccc.int/resource/docs/2015/cop21/eng/da02.pdf). While I used an oil-palm example, my backcasting approach employed in Chapter 5 could be expanded to other sectors, for instance emissions from transportation sector to be offset through forest and peatland restoration. A broader application of inter-sectoral carbon offsets will generate more investments, which can be used to further leverage restoration. For Indonesia, the offsetting policy could help the country to achieve the Intended Nationally Determined Contribution under the Paris Agreement with unconditional reduction of 29% of greenhouse gases against a business as usual scenario by 2030 (http://www4.unfccc.int/submissions/INDC/Published%20Documents/Indonesia/1/INDC_REPUBL IC%20OF%20INDONESIA.pdf).

Chapter 5 demonstrated the interesting finding that there is a strong trade-off between the delivery of carbon and biodiversity benefits between scenarios compensating only for carbon emissions and biodiversity losses. This differs with the result of Chapter 4 in which trade-offs between both features

105

Chapter 7. Discussion are less prevalent. The difference is likely because I used a provincial-scale case study in Chapter 4 which may be inadequate for capturing the heterogeneity that occurs in the Kalimantan-wide analysis of Chapter 5. Also, Chapter 4 only considered carbon benefits in term of carbon sequestration, which could undermine the benefits of avoided emissions from soil carbon as accounted for in Chapter 5. These findings provide evidence that the concept of a win-win (or compromised) solution cannot be generalised in ecosystem services analyses, instead we should acknowledge that trade-offs are inevitable and could vary in intensity (Howe et al., 2014). My analyses suggest that the level of trade-off is influenced by the scale employed and the way benefits are defined, highlighting that trade- offs are context specific and value laden (Howe et al., 2014).

While I find restoring highly degraded peatlands with deep peat thickness as a cost-effective way for carbon offsetting, there are several issues to consider for effective implementation on the ground. These were not explicitly addressed in Chapter 5. One key factor of successful peatland restoration is political support at local, regional and national level to clarify and enforce land-tenure systems (Page et al., 2009). This is to break the vicious cycle of tenure disputes which often leads to infringements such as land-encroachment and peat burning (Gaveau et al., 2014a; Harrison et al., 2009). Appropriate institutional arrangement consisting of capable stakeholders is required to implement restoration so that success stories (e.g. Central Kalimantan Peatlands Project) can be replicated (CKPP, 2016). Reforestation should also be developed in line with social characteristics of a region to provide livelihoods for local communities, for example using indigenous agroforestry as a framework for restoration (Limin et al., 2008; Page et al., 2009).

In Chapter 5, I only considered offsets using a restoration scheme that aligns with the overall topic of my thesis and the primary offsetting mechanism prescribed by the Roundtable on Sustainable Palm Oil (RSPO, 2014). Another offsetting scheme entails protecting areas with ecological value equivalent to the damages on the impacted sites (Kiesecker et al., 2009a; Maron et al., 2015). In Kalimantan, however, most intact forests have been protected by existing policies including various regulations on protected areas and REDD+. Incorporating the protection of intact forests into offsetting analysis is possible, but to calculate additionality from offsets would need additional variables describing the effectiveness of protected areas management and REDD+ across Kalimantan (Curran et al., 2004; Fuller et al., 2004; Murray et al., 2015; Santika et al., 2015b). Also, for offset schemes focused on habitat protection, the prediction of future land use dynamics would be required to calculate the offset benefits by assuming what would have occurred in the absence of offsets (e.g. the landscape would be deforested; Maron et al., 2013). A range of approaches could be employed to predict the dynamics of land use (e.g. Brun et al., 2015; Sonter et al., 2014) but this would require numerous social and economic variables as predictors and for casual relationships to be understood 106

Chapter 7. Discussion

(Baylis et al., 2015; Ferraro et al., 2015). Future study integrating restoration and protection schemes simultaneously will provide comprehensive insights to identify the best combination of what offset scheme to apply and where.

7.3.3. Utilising social-ecological systems framework to enhance feasibility in restoration

Many landscape-scale restoration projects are regarded as ineffective in achieving the intended goals (Murniati et al., 2007; Wuethrich, 2007). Clearly, there is a lack of knowledge on how feasibility is defined to enhance the likelihood of success. In Chapter 6, I developed an innovative approach to enhance feasibility in restoration planning. I integrated a social-ecological systems framework with systematic decision theory to develop context-specific restoration plans. I discovered that incorporating the socio-political context alters the selection of priority areas, enhances feasibility, and potentially reduces ecological benefits and/or increases restoration costs compared with considering biophysical considerations alone. My integrated approach could also be used to assess the efficiency of existing policies related to restoration.

Chapter 6 demonstrates the gap on the use of simplifying assumptions in existing restoration planning studies (and conservation more broadly) which have routinely overlooked social and political perspectives (Baker et al., 2013; Cowling and Wilhelm-Rechmann, 2007; Knight et al., 2013). As such, it is not surprising that many plans remain as paper plans or failed plans (Ban et al., 2013; Murniati et al., 2007). While I provided a case study using few socio-political variables, I discovered that only half the priority areas were the same with, and without, these variables. I argue that this finding could be used as evidence to transform how restoration plans should be developed to equally consider ecological, social and political contexts to enhance implementation effectiveness.

Feasibility is a relatively new perspective being explored in conservation sciences (Knight et al., 2011) with few studies having attempted to define feasibility in restoration planning (e.g. Curran et al., 2012; Jellinek et al., 2014; Orsi and Geneletti, 2010). Yet, as feasibility strongly relates to social and political dimensions with many complex inter-connected variables, there has been no prescriptive way to identify the most prevalent variables. As I demonstrated in Chapter 6, a social-ecological systems framework (i.e. Ostrom, 2009) could be useful to provide insight into variables affecting feasibility in restoration. Ostrom’s diagnostic approach has been developed through a long-term field observation in various locations worldwide encompassing a wide-range of themes from forestry to fishery (Basurto et al., 2013; Nagendra, 2007). Certainly, Ostrom’s prescription is not a silver bullet, but for now, it is the most widely used framework in environmental management which can be utilised for making decisions in restoration.

107

Chapter 7. Discussion

In my study, the simple yet well-structured framework by Ostrom has helped to diagnose potential restoration strategies based on social and political characteristics of the study region. The diagnostic processes in Ostrom’s framework involves a somewhat subjective judgement, although this judgement is developed through comprehensive and systematic investigation using empirical evidence (Basurto et al., 2013). This is similar to the situation in medical practice in which not all the empirically-proven treatments can be applied by the doctor and accepted by the patient and society due to the values they hold, for example, in the case of organ transplantation. Instead of viewing the judgement process as a caveat, I considered it as the strength of Ostrom’s framework combining values and scientific evidence, which are all relevant in ecosystem services analysis (Díaz et al., 2015a; Díaz et al., 2015b).

I acknowledge several limitations in Chapter 6. While I accounted for variables reflecting the value of indigenous agroforestry systems for the provision of livelihoods (e.g. social dependency, communities’ preferences), these variables do not explicitly show the monetary benefits of the system. Recent literature demonstrating how monetary value of non-timber forest products from agroforest is spatially calculated (e.g. Schaafsma et al., 2014; Sumarga et al., 2015) could enhance the practical implications of my study. I recommend that my analysis could be expanded to combine socio-cultural values and monetary benefits of restoration for livelihoods provision.

I also did not account for the dynamics of the social-ecological systems of interest. Socio-political situations could also rapidly change in response to internal and external drivers, including improved education, change in lifestyle, improved access to information and market-price fluctuation (Feintrenie and Levang, 2009; Fox et al., 2014). Ideally, the dynamics of social variables could be incorporated in this study, for example using a dynamic systems approach (e.g. Medrilzam et al., 2014) to develop causal relationships between socio-political drivers and restoration outcomes. Nonetheless, the incorporation of a dynamic systems method will require complex socio-ecological models with large data requirements.

7.4. Synthesis and future directions

In general, my thesis demonstrates that the decision in landscape-scale restoration is not simply a binary answer: restore or not restore. By employing decision science, I have shown how and where we should do restoration. The use of carefully developed frameworks when making decisions for restoration can identify opportunities for the effective implementation of landscape-scale restoration. While my thesis focuses on a tropical setting, the conceptual frameworks could be applied to similar restoration problems globally.

108

Chapter 7. Discussion

Addressing heterogeneity is the first step when designing landscape-scale restoration programs. Heterogeneity appears in many facets, from ecological perspectives (e.g. landscape condition, biophysical variables), to economic (e.g. restoration cost), social (e.g. community support) and political (e.g. legal considerations). Heterogeneity could also be in the form of a restoration goal representing a broad range of beneficiaries and values (e.g. global communities, local people), expected benefits (e.g. carbon sequestration, biodiversity conservation, livelihood provision), and policy strategy (e.g. REDD+, offsets, indigenous agroforestry). Ideally, all aspects of heterogeneity are considered in decision-making processes to produce close-to-reality and effective restoration plans. This, however, will require robust development of methods (e.g. models) with large amounts of spatial information, which for many regions is not always available. Greater effort to collect data and develop innovative methods are therefore required for a comprehensive account of heterogeneity.

Determining the appropriate scale of planning is the key consideration when addressing landscape heterogeneity. This is problematic as we have to consider the scale at the outset of planning processes with regard to overarching restoration goals and policy contexts. For example, restoration planning employed at a national and regional level will be different to that at a provincial and district level. This difference could affect the congruence of solutions produced at different scales, particularly when spatial distribution of restoration benefits and costs have high variability and large divergence. Currently, research aiming to resolve the problems of scale-mismatch is still lacking in decision- making in restoration ecology. In this context, we need to expand our understanding of how decision outcomes resulting from broad-scale analysis can be best aligned with fine-scale studies.

The way restoration costs are defined will affect restoration decisions. Assuming restoration cost to be uniform across a landscape can be erroneous considering the fact that landscapes are heterogeneous. Restoration cost should be factored in by accounting for implementation cost on the basis of landscape condition, and foregone revenues of alternative land management (i.e. opportunity cost) in regard to biophysical potential. Many studies in restoration ecology have been focused on the implementation costs with limited attention to opportunity costs. The common problem when defining opportunity cost is how the assumption of alternative land uses is developed. Opportunity cost could vary significantly in value and spatial distribution, depending on the assumptions made. For example, accounting only for timber revenues will differ to using comprehensive measures for all plausible land uses (e.g. mining, oil-palm plantation). Knowledge pertaining to the temporal dynamics of restoration costs has also been poorly developed. For instance, as degraded forest is restored towards an intact condition, the implementation cost will decline but the opportunity cost will escalate as the trees reach a harvestable size. Therefore, future research in restoration ecology should be aimed at building knowledge on cost accounting through time. 109

Chapter 7. Discussion

Similarly, restoration benefits should be carefully defined and calculated. To correctly define the net accumulated benefit, we have to consider the difference in value between the initial condition and the restored state, and not only refer to the end state as commonly assumed. Developing appropriate assumptions when quantifying restoration benefits could also affect decision outcomes. For example, the objective of carbon can be formulated in two ways: accounting only for carbon sequestration from the restored forest, or combining carbon sequestration with avoided emissions assuming that restoration would not have occurred. As restoration involves time-lags, a temporally-explicit approach should be considered when calculating restoration benefits. Benefits gained in the near future will have a different value to those accumulated in the distant future because of delayed outputs and uncertainties, reflecting a discounting principle as widely used in economic theory. The discounting concept of non-monetary benefits in restoration (e.g. biodiversity improvement, carbon sequestered) is relatively new in restoration ecology, and needs further exploration.

While affecting the likelihood of success in restoration, social and political perspectives have been rarely considered in restoration planning studies. The investigation of social feasibility and political permissibility for incorporation with ecological assessments is clearly required. As socio-political variables are complex, context-specific and rapidly changing, a robust framework (e.g. social- ecological systems) could help to define feasibility in restoration. Addressing this complexity will need an improved understanding developed from empirical evidence to investigate the relationships between social and ecological variables in restoration, which unfortunately is still limited. Future research should be focused on revealing these relationships (e.g. the influence of poverty on restoration success) to provide a basis for planning and decision-making processes.

Finally, as ecosystem services encompass a broad range of elements, determining what relevant components to consider in making decisions for restoration is essential. Ideally, the identification of key components of ecosystem services is developed through a participatory approach involving various stakeholders and scientific disciplines to capture diverse values and knowledge systems. Consensus-based decision-making has been lacking in restoration ecology due to the much-focused- on site-scale analysis, viewed mainly from the perspective of ecologists (e.g. focusing only on biodiversity). This problem suggests inclusive approaches that accommodate broad interests should be fostered when making decisions for landscape-scale restoration to produce socially and politically robust solutions.

110

Chapter 7. Discussion

7.5. Concluding remarks

The year of 2020 is approaching. This will be the moment of truth for the Aichi targets of the CBD, including whether the target of restoring at least 15% of degraded ecosystems globally will be achieved. The achievement of ambitious restoration policy targets requires strong support from science and this thesis has contributed to the required body of knowledge. The rapidly growing paradigm on ecosystem services is an opportunity to reform our traditional egocentric ideology to view the benefits of restoration in a more comprehensive manner. Decision science can inform policy and decision-making in large-scale restoration programs. Yet, our understanding is far from complete. New ideas and knowledge to improve decision making are needed to ensure the multitude of values derived from of ecosystems are restored for future generations.

111

References

112

References

Abeli, T., Gentili, R., Rossi, G., Bedini, G., Foggi, B. (2009) Can the IUCN criteria be effectively applied to peripheral isolated plant populations? Biodiversity and Conservation 18, 3877-3890.

Abood, S.A., Lee, J.S.H., Burivalova, Z., Garcia-Ulloa, J., Koh, L.P. (2015) Relative Contributions of the Logging, Fiber, Oil Palm, and Mining Industries to Forest Loss in Indonesia. Conservation Letters 8, 58-67.

Abram, K.A., Meijaard, E., Wilson, K.A., Davis, J.T., Wells, J.A., Ancrenaz, M., Budiharta, S., Durrant, A., Fakhruzzi, A., Runting, R.K., Gaveau, D., Mengersen, K. (in review) Oil palm- community conflict mapping in Indonesia: A case for better community liaison in planning for development initiatives. Global Environmental Change.

Abram, N.K., Meijaard, E., Ancrenaz, M., Runting, R.K., Wells, J.A., Gaveau, D., Pellier, A.-S., Mengersen, K. (2014) Spatially explicit perceptions of ecosystem services and land cover change in forested regions of Borneo. Ecosystem Services 7, 116-127.

Achard, F., Eva, H.D., Stibig, H.-J., Mayaux, P., Gallego, J., Richards, T., Malingreau, J.-P. (2002) Determination of Deforestation Rates of the World's Humid Tropical Forests. Science 297, 999-1002.

Adger, W.N. (2006) Vulnerability. Global Environmental Change 16, 268-281.

Ådjers, G., Hadengganan, S., Kuusipalo, J., Nuryanto, K., Vesa, L. (1995) Enrichment planting of dipterocarps in logged-over secondary forests: effect of width, direction and maintenance method of planting line on selected Shorea species. Forest Ecology and Management 73, 259-270.

Agrawal, A., Cashore, B., Hardin, R., Shepherd, G., Benson, C., Miller, D., (2013) Economic contributions of forests, Background 1, United Nations Forum on Forests (UNFF), 10th Session, Istanbul, Turkey.

Ahmad, A.H., Khairul-Adha, A.R., (2007) State of knowledge on freshwater fishes of Malaysia, in: Chua, L. (Ed.), Status of Biological Diversity in Malaysia and Threat Assessment of Plant Species in Malaysia. Forest Research Institute Malaysia, Kuala Lumpur, Malaysia, pp. 83–90.

Alexander, S., Nelson, C.R., Aronson, J., Lamb, D., Cliquet, A., Erwin, K.L., Finlayson, C.M., de Groot, R.S., Harris, J.A., Higgs, E.S., Hobbs, R.J., Robin Lewis, R.R., Martinez, D., Murcia, C. (2011) Opportunities and Challenges for Ecological Restoration within REDD+. Restoration Ecology 19, 683-689.

Alfred, R., Ahmad, A.H., Payne, J., Williams, C., Ambu, L.N., How, P.M., Goossens, B. (2012) Home Range and Ranging Behaviour of Bornean Elephant (Elephas maximus borneensis) Females. PLoS ONE 7, e31400.

Angelsen, A., Jagger, P., Babigumira, R., Belcher, B., Hogarth, N.J., Bauch, S., Borner, J., Smith- Hall, C., Wunder, S. (2014) Environmental Income and Rural Livelihoods: A Global-Comparative Analysis. World Development 64, S12-S28.

113

References

Ansell, F.A., Edwards, D.P., Hamer, K.C. (2011) Rehabilitation of Logged Rain Forests: Avifaunal Composition, Habitat Structure, and Implications for Biodiversity-Friendly REDD+. Biotropica 43, 504-511.

Aoyagi, R., Imai, N., Kitayama, K. (2013) Ecological significance of the patches dominated by pioneer trees for the regeneration of dipterocarps in a Bornean logged-over secondary forest. Forest Ecology and Management 289, 378-384.

Aronson, J., Brancalion, P.H.S., Durigan, G., Rodrigues, R.R., Engel, V.L., Tabarelli, M., Torezan, J.M.D., Gandolfi, S., de Melo, A.C.G., Kageyama, P.Y., Marques, M.C.M., Nave, A.G., Martins, S.V., Gandara, F.B., Reis, A., Barbosa, L.M., Scarano, F.R. (2011) What Role Should Government Regulation Play in Ecological Restoration? Ongoing Debate in São Paulo State, Brazil. Restoration Ecology 19, 690-695.

Ashton, P.M.S., (1989) Sundaland, in: Campbell, D.G., Hammond, H.D. (Eds.), Floristic inventory of tropical countries. New York Botanical Garden, New York.

Asner, G.P., Rudel, T.K., Aide, T.M., Defries, R., Emerson, R. (2009) A Contemporary Assessment of Change in Humid Tropical Forests. Conservation Biology 23, 1386-1395.

Baker, S., Eckerberg, K., Zachrisson, A. (2013) Political science and ecological restoration. Environmental Politics 23, 509-524.

Bakker, L., (2008) “Can We Get Hak Ulayat?”: Land and Community in Pasir and Nunukan, East Kalimantan, Ten Years After: Reformasi and New Social Movements in Indonesia, 1998-2008, University of California, Berkeley.

Ball, I.R., Possingham, H.P., Watts, M., (2009) Marxan and relatives: Software for spatial conservation prioritization, in: Moilanen, A.J., Wilson, K.A., Possingham, H.P. (Eds.), Spatial conservation prioritisation: Quantitative methods and computational tools. Oxford University Press, Oxford, UK, pp. 185-195.

Ban, N.C., Mills, M., Tam, J., Hicks, C.C., Klain, S., Stoeckl, N., Bottrill, M.C., Levine, J., Pressey, R.L., Satterfield, T., Chan, K.M.A. (2013) A social–ecological approach to conservation planning: embedding social considerations. Frontiers in Ecology and the Environment 11, 194-202.

Bank Indonesia, (2013) Bank Indonesia Rate. Bank Indonesia, Jakarta, Indonesia. http://www.bi.go.id/id/moneter/bi-rate/data/Default.aspx (accessed 28 December 2013).

Barr, C., Dermawan, A., Purnomo, H., Komarudin, H. (2010) Financial governance and Indonesia's Reforestation Fund during the Soeharto and post-Soeharto periods, 1989-2009: a political economic analysis of lessons for REDD+. Center for International Forestry Research (CIFOR), Bogor, Indonesia.

Barr, C.M., Sayer, J.A. (2012) The political economy of reforestation and forest restoration in Asia– Pacific: Critical issues for REDD+. Biological Conservation 154, 9-19.

Basurto, X., Gelcich, S., Ostrom, E. (2013) The social–ecological system framework as a knowledge classificatory system for benthic small-scale fisheries. Global Environmental Change 23, 1366-1380. 114

References

Baur, I. (2013) Adapting to Socioeconomic Developments by Changing Rules in the Governance of Common Property Pastures in the Swiss Alps. Ecology and Society 18.

Baylis, K., Honey-Rosés, J., Börner, J., Corbera, E., Ezzine-de-Blas, D., Ferraro, P.J., Lapeyre, R., Persson, U.M., Pfaff, A., Wunder, S. (2015) Mainstreaming Impact Evaluation in Nature Conservation. Conservation Letters, n/a-n/a.

BBOP, (2012) Biodiversity offset design handbook—updated, in: Treweek, J.O., ten Kate, K. (Eds.). BBOP, Washington, D.C., USA.

Beaman, J.H., Burley, J.S., (2003) Progress in the floristic inventory of Borneo, in: Padoch, C., Peluso, N. (Eds.), Borneo in Transition: People, Forests, Conservation and Development. Oxford University Press, New York, pp. 93–113.

Belcher, B., Rujehan, Imang, N., Achdiawan, R. (2004) Rattan, rubber, or oil palm: cultural and financial considerations for farmers in Kalimantan. Economic Botany 58, S77-S87.

Benayas, J.M.R., Newton, A.C., Diaz, A., Bullock, J.M. (2009) Enhancement of Biodiversity and Ecosystem Services by Ecological Restoration: A Meta-Analysis. Science 325, 1121-1124.

Berry, N., Phillips, O., Lewis, S., Hill, J., Edwards, D., Tawatao, N., Ahmad, N., Magintan, D., Khen, C., Maryati, M., Ong, R., Hamer, K. (2010) The high value of logged tropical forests: lessons from northern Borneo. Biodiversity and Conservation 19, 985-997.

Betha, R., Pradani, M., Lestari, P., Joshi, U.M., Reid, J.S., Balasubramanian, R. (2013) Chemical speciation of trace metals emitted from Indonesian peat fires for health risk assessment. Atmospheric Research 122, 571-578.

Birch, J.C., Newton, A.C., Aquino, C.A., Cantarello, E., Echeverría, C., Kitzberger, T., Schiappacasse, I., Garavito, N.T. (2010) Cost-effectiveness of dryland forest restoration evaluated by spatial analysis of ecosystem services. Proceedings of the National Academy of Sciences 107, 21925- 21930.

Birdlife International, (2014) Forests of Hope site - Harapan Rainforest, Indonesia. http://www.birdlife.org/worldwide/projects/forests-hope-site-harapan-rainforest-indonesia.

Blaser, J., Sarre, A., Poore, D., Johnson, S., (2011) Status of Tropical Forest Management 2011, ITTO Technical Series No. 38. International Tropical Timber Organization, Yokohama.

Blignaut, J., Aronson, J., de Wit, M. (2014) The economics of restoration: looking back and leaping forward. Annals of the New York Academy of Sciences 1322, 35-47.

Boedhihartono, A., Sayer, J., (2012) Forest Landscape Restoration: Restoring What and for Whom?, in: Stanturf, J., Lamb, D., Madsen, P. (Eds.), Forest Landscape Restoration. Springer Netherlands, pp. 309-323.

Boer, R., (2012) Sustainable forest management, forest-based carbon, carbon stock, CO2 sequestration and green products to reduce emissions from deforestation and forest degradation,

115

References

Enhancing Forest Carbon Stock to Reduce Emissions from Deforestation and Degradation through Sustainable Forest Management (SFM) Initiatives in Indonesia. Indonesia Ministry of Forestry - International Tropical Timber Organization, Jakarta, Indonesia.

BOS Foundation, (2013) Borneo Orangutan Survival Foundation Annual Report 2012. Borneo Orangutan Survival Foundation, Bogor, Indonesia. http://orangutan.or.id/wp- content/uploads/2013/09/Lowres-Annual-Report-2012.pdf.

Brearley, F.Q., Press, M.C., Scholes, J.D. (2003) Nutrients obtained from leaf litter can improve the growth of dipterocarp seedlings. New Phytologist 160, 101-110.

Brockhaus, M., Obidzinski, K., Dermawan, A., Laumonier, Y., Luttrell, C. (2012) An overview of forest and land allocation policies in Indonesia: Is the current framework sufficient to meet the needs of REDD+? Forest Policy and Economics 18, 30-37.

Brown, K., (2005) Addressing Trade-Offs in Forest Landscape Restoration, Forest Restoration in Landscapes. Springer New York, pp. 59-64.

Brown, S., Lugo, A.E. (1982) The Storage and Production of Organic Matter in Tropical Forests and Their Role in the Global Carbon Cycle. Biotropica 14, 161-187.

Brown, S., Lugo, A.E. (1990) Tropical secondary forests. Journal of Tropical Ecology 6, 1-32.

Brown, T., (in press) Local Climate Change Impact and Adaptation, in: Resosudarmo, B.P., Imansyah, M.H., Napitupulu, L. (Eds.), Development, Environment and the People of Kalimantan. Indonesia Regional Science Association (IRSA), Jakarta, Indonesia.

Brun, C., Cook, A.R., Lee, J.S.H., Wich, S.A., Koh, L.P., Carrasco, L.R. (2015) Analysis of deforestation and protected area effectiveness in Indonesia: A comparison of Bayesian spatial models. Global Environmental Change 31, 285-295.

Bryan, B.A., Crossman, N.D. (2008) Systematic regional planning for multiple objective natural resource management. Journal of Environmental Management 88, 1175-1189.

Budiharta, S., Meijaard, E., (in press) State of Kalimantan's Biodiversity, in: Resosudarmo, B.P., Imansyah, M.H., Napitupulu, L. (Eds.), Development, Environment and the People of Kalimantan. Indonesian Regional Science Association (IRSA), Jakarta, Indonesia.

Budiharta, S., Meijaard, E., Erskine, P.D., Rondinini, C., Pacifici, M., Wilson, K.A. (2014a) Restoring degraded tropical forests for carbon and biodiversity. Environmental Research Letters 9, 114020.

Budiharta, S., Meijaard, E., Wells, J.A., Abram, N., Wilson, K.A. (in review) Enhancing feasibility: incorporating a socio-ecological systems framework into restoration planning. Environmental Science and Policy.

Budiharta, S., Slik, F., Raes, N., Meijaard, E., Erskine, P.D., Wilson, K.A. (2014b) Estimating the Aboveground Biomass of Bornean Forest. Biotropica 46, 507-511.

116

References

Budiharta, S., Widyatmoko, D., Irawati, Wiriadinata, H., Rugayah, Partomihardjo, T., Ismail, Uji, T., Keim, A.P., Wilson, K. (2011) The processes that threaten Indonesian plants. Oryx 45, 172-179.

Bullock, J.M., Aronson, J., Newton, A.C., Pywell, R.F., Rey-Benayas, J.M. (2011) Restoration of ecosystem services and biodiversity: conflicts and opportunities. Trends in Ecology & Evolution 26, 541-549.

Busch, J., Ferretti-Gallon, K., Engelmann, J., Wright, M., Austin, K.G., Stolle, F., Turubanova, S., Potapov, P.V., Margono, B., Hansen, M.C., Baccini, A. (2015) Reductions in emissions from deforestation from Indonesia’s moratorium on new oil palm, timber, and logging concessions. Proceedings of the National Academy of Sciences 112, 1328-1333.

Calmon, M., Brancalion, P.H.S., Paese, A., Aronson, J., Castro, P., da Silva, S.C., Rodrigues, R.R. (2011) Emerging Threats and Opportunities for Large-Scale Ecological Restoration in the Atlantic Forest of Brazil. Restoration Ecology 19, 154-158.

Caniago, I., Stephen, F.S. (1998) Medicinal plant ecology, knowledge and conservation in Kalimantan, Indonesia. Economic Botany 52, 229-250.

Cannon, C.H., Leighton, M. (2004) Tree species distributions across five habitats in a Bornean rain forest. Journal of Vegetation Science 15, 257-266.

Cannon, C.H., Peart, D.R., Leighton, M. (1998) Tree Species Diversity in Commercially Logged Bornean Rainforest. Science 281, 1366-1368.

Cannon, C.H., Peart, D.R., Leighton, M., Kartawinata, K. (1994) The structure of lowland rainforest after selective logging in West Kalimantan, Indonesia. Forest Ecology and Management 67, 49-68.

Cao, S.X., Zhong, B.L., Yue, H., Zeng, H.S., Zeng, J.H. (2009) Development and testing of a sustainable environmental restoration policy on eradicating the poverty trap in China's Changting County. Proceedings of the National Academy of Sciences of the United States of America 106, 10712-10716.

Cardoso, P., Borges, P.A.V., Triantis, K.A., Ferrández, M.A., Martín, J.L. (2011) Adapting the IUCN Red List criteria for invertebrates. Biological Conservation 144, 2432-2440.

Carlson, K.M., Curran, L.M., Asner, G.P., Pittman, A.M., Trigg, S.N., Adeney, J.M. (2013) Carbon emissions from forest conversion by Kalimantan oil palm plantations. Nature Climate Change 3, 283- 287.

Carwardine, J., Hawkins, C., Polglase, P., Possingham, H.P., Reeson, A., Renwick, A.R., Watts, M., Martin, T.G. (2015) Spatial Priorities for Restoring Biodiverse Carbon Forests. BioScience 65, 372- 382.

Casson, A. (2000) The Hesitant boom: Indonesia's oil palm sub-sector in an era of economic crisis and political change. CIFOR, Bogor, Indonesia.

117

References

Chan, C., Lamb, A., Shim, P., Wood, J. (1994) Orchids of Borneo: Volume 1. Introduction and a selection of species. Royal Botanic Gardens, Kew, UK.

Chan, F., (2015) $47b? Indonesia counts costs of haze, The Straits Times. The Straits Times, Singapore.

Chan, K.M.A., Shaw, M.R., Cameron, D.R., Underwood, E.C., Daily, G.C. (2006) Conservation Planning for Ecosystem Services. PLoS Biol 4, e379.

Chazdon, R.L. (2008) Beyond Deforestation: Restoring Forests and Ecosystem Services on Degraded Lands. Science 320, 1458-1460.

Chazdon, R.L., Brancalion, P.H.S., Lamb, D., Laestadius, L., Calmon, M., Kumar, C. (2015) A Policy-Driven Knowledge Agenda for Global Forest and Landscape Restoration. Conservation Letters, n/a-n/a.

CKPP, (2016) The Central Kalimantan Peatland Project. (http://ckpp.wetlands.org/).

Clearwater, M.J., Susilawaty, R., Effendi, R., van Gardingen, P.R. (1999) Rapid photosynthetic acclimation of Shorea johorensis seedlings after logging disturbance in Central Kalimantan. Oecologia 121, 478-488.

Coates, B.J., Bishop, K.D. (1997) A Guide to the Birds of Wallacea: Sulawesi, the Moluccas and the Lesser Sunda Islands, Indonesia. Dove Publications, Alderley.

Cole, R.J., Holl, K.D., Keene, C.L., Zahawi, R.A. (2011) Direct seeding of late-successional trees to restore tropical montane forest. Forest Ecology and Management 261, 1590-1597.

Convention on Biological Diversity, (2011) Conference of the Parties Decision X/2: Strategic plan for biodiversity 2011–2020. United Nations. https://www.cbd.int/sp/targets/rationale/default.shtml.

Corlett, R.T. (2007) The Impact of Hunting on the Mammalian Fauna of Tropical Asian Forests. Biotropica 39, 292-303.

Cottam, G., Curtis, J.T. (1956) The Use of Distance Measures in Phytosociological Sampling. Ecology 37, 451-460.

Couwenberg, J., Dommain, R., Joosten, H. (2010) Greenhouse gas fluxes from tropical peatlands in south-east Asia. Global Change Biology 16, 1715-1732.

Cowling, R.M., Wilhelm-Rechmann, A. (2007) Social assessment as a key to conservation success. Oryx 41, 135-136.

Cox, M. (2014) Applying a Social-Ecological System Framework to the Study of the Taos Valley Irrigation System. Human Ecology 42, 311-324.

118

References

Crevello, S.M., (2003) Local land use on Borneo: applications of indigenous knowledge systems and natural resource utilization among the Benuaq Dayak of Kalimantan, Indonesia, The School of Renewable Natural Resources. Louisiana State University, Louisiana.

Crossman, N.D., Bryan, B.A. (2006) Systematic landscape restoration using integer programming. Biological Conservation 128, 369-383.

Crossman, N.D., Bryan, B.A. (2009) Identifying cost-effective hotspots for restoring natural capital and enhancing landscape multifunctionality. Ecological Economics 68, 654-668.

Crossman, N.D., Bryan, B.A., Summers, D.M. (2011) Carbon Payments and Low-Cost Conservation. Conservation Biology 25, 835-845.

Curran, L.M., Trigg, S.N., McDonald, A.K., Astiani, D., Hardiono, Y.M., Siregar, P., Caniago, I., Kasischke, E. (2004) Lowland Forest Loss in Protected Areas of Indonesian Borneo. Science 303, 1000-1003.

Curran, L.M., Webb, C.O. (2000) Experimental tests of the spatiotemporal scale of seed predation in mast-fruiting Dipterocarpaceae. Ecological Monographs 70, 129-148.

Curran, M., Hellweg, S., Beck, J. (2013) Is there any empirical support for biodiversity offset policy? Ecological Applications 24, 617-632.

Curran, P., Smedley, D., Thompson, P., Knight, A.T. (2012) Mapping Restoration Opportunity for Collaborating with Land Managers in a Carbon Credit-Funded Restoration Program in the Makana Municipality, Eastern Cape, South Africa. Restoration Ecology 20, 56-64.

Das, I., Lakim, M., Lim, K.K.P., Hui, T.H. (2008) New Species of Anomochilus from Borneo (Squamata: Anomochilidae). Journal of Herpetology 42, 584-591. de Bruyn, M., Stelbrink, B., Morley, R.J., Hall, R., Carvalho, G.R., Cannon, C.H., van den Bergh, G., Meijaard, E., Metcalfe, I., Boitani, L., Maiorano, L., Shoup, R., von Rintelen, T. (2014) Borneo and Indochina are Major Evolutionary Hotspots for Southeast Asian Biodiversity. Systematic Biology 63, 879-901.

De Groot, R.S., Blignaut, J., Van Der Ploeg, S., Aronson, J., Elmqvist, T., Farley, J. (2013) Benefits of Investing in Ecosystem Restoration. Conservation Biology 27, 1286-1293. de Jong, W. (2010) Forest Rehabilitation and Its Implication for Forest Transition Theory. Biotropica 42, 3-9. de Jong, W., Chokkalingam, U., Smith, J., Sabogal, C. (2001) Tropical secondary forests in Asia: introduction and synthesis. Journal of tropical forest science 13, 563-576.

DeFries, R.S., Houghton, R.A., Hansen, M.C., Field, C.B., Skole, D., Townshend, J. (2002) Carbon emissions from tropical deforestation and regrowth based on satellite observations for the 1980s and 1990s. Proceedings of the National Academy of Sciences 99, 14256-14261.

119

References

Dewi, S., Belcher, B., Puntodewo, A. (2005) Village economic opportunity, forest dependence, and rural livelihoods in East Kalimantan, Indonesia. World Development 33, 1419-1434.

Dewi, S., Khasanah, N., Rahayu, S., Ekadinata, A., Van Noordwijk, M. (2009) Carbon footprint of Indonesian palm oil production: a pilot study. World Agroforestry Centre-ICRAF, SEA Regional Office, Bogor, Indonesia 8.

Díaz, S., Demissew, S., Carabias, J., Joly, C., Lonsdale, M., Ash, N., Larigauderie, A., Adhikari, J.R., Arico, S., et al. (2015a) The IPBES Conceptual Framework — connecting nature and people. Current Opinion in Environmental Sustainability 14, 1-16.

Díaz, S., Demissew, S., Joly, C., Lonsdale, W.M., Larigauderie, A. (2015b) A Rosetta Stone for Nature’s Benefits to People. PLoS Biol 13, e1002040.

Dinerstein, E., Varma, K., Wikramanayake, E., Powell, G., Lumpkin, S., Naidoo, R., Korchinsky, M., Del Valle, C., Lohani, S., Seidensticker, J., Joldersma, D., Lovejoy, T., Kushlin, A. (2013) Enhancing Conservation, Ecosystem Services, and Local Livelihoods through a Wildlife Premium Mechanism. Conservation Biology 27, 14-23.

Dudley, N., Morrison, J., Aronson, J., Mansourian, S., (2005) Why Do We Need to Consider Restoration in a Landscape Context?, Forest Restoration in Landscapes. Springer New York, pp. 51- 58.

Ebeling, J., Yasué, M. (2009) The effectiveness of market-based conservation in the tropics: Forest certification in Ecuador and Bolivia. Journal of Environmental Management 90, 1145-1153.

Edwards, D.P., Ansell, F.A., Ahmad, A.H., Nilus, R., Hamer, K.C. (2009) The Value of Rehabilitating Logged Rainforest for Birds. Conservation Biology 23, 1628-1633.

Edwards, D.P., Laurance, S.G. (2012) Green labelling, sustainability and the expansion of tropical agriculture: Critical issues for certification schemes. Biological Conservation 151, 60-64.

Egoh, B.N., Paracchini, M.L., Zulian, G., Schägner, J.P., Bidoglio, G. (2014) Exploring restoration options for habitats, species and ecosystem services in the European Union. Journal of Applied Ecology 51, 899-908.

Egoh, B.N., Reyers, B., Rouget, M., Richardson, D.M. (2011) Identifying priority areas for ecosystem service management in South African grasslands. Journal of Environmental Management 92, 1642- 1650.

Eschenbach, C., Glauner, R., Kleine, M., Kappen, L. (1998) Photosynthesis rates of selected tree species in lowland dipterocarp rainforest of Sabah, Malaysia. Trees 12, 356-365.

ESRI, (2013) ArcGIS Desktop: Release 10.2. Environmental Systems Research Institute Redlands, CA.

120

References

Etter, A., McAlpine, C.A., Seabrook, L., Wilson, K.A. (2011) Incorporating temporality and biophysical vulnerability to quantify the human spatial footprint on ecosystems. Biological Conservation 144, 1585-1594.

Evans, M.C., Tulloch, A.I.T., Law, E.A., Raiter, K.G., Possingham, H.P., Wilson, K.A. (2015) Clear consideration of costs, condition and conservation benefits yields better planning outcomes. Biological Conservation 191, 716-727.

Eyring, K.L., Healy, C.J., Reyda, F.B. (2012) A New Genus and Species of Cestode (Rhinebothriidea) from Mobula kuhlii (Rajiformes: Mobulidae) from Malaysian Borneo. Journal of Parasitology 98, 584-591.

FACE the Future, (2011) INFAPRO Rehabilitation of logged-over dipterocarp forest in Sabah, Malaysia, Verified Carbon Standard Project Description Template Version 1.7. Verified Carbon Standard.

Fachrizal, A., Pahlevi, A., (2013) Illegal wildlife trade syndicate uncovered in West Kalimantan, hundreds of protected animal organs confiscated, Mongabay Indonesia. http://www.mongabay.co.id/2013/04/26/sindikat-perdagangan-ilegal-terbongkar-di-kalbar-ratusan- organ-satwa-dilindungi-disita/.

Faisal, M., (2013) State Structural and Institutional Transformation and the New Dynamics of Business Power, Corruption and Clientelism in Indonesia, School of Political Science and International Studies. The University of Queensland, Australia, Brisbane, Queensland.

Falloon, P., Smith, P. (2003) Accounting for changes in soil carbon under the Kyoto Protocol: need for improved long-term data sets to reduce uncertainty in model projections. Soil Use and Management 19, 265-269.

FAO/IIASA/ISRIC/ISSCAS/JRC, (2012) Harmonized World Soil Database (version 1.2). FAO, Rome, Italy and IIASA, Laxenburg, Austria. http://webarchive.iiasa.ac.at/Research/LUC/External- World-soil-database/HTML/ (accessed 25 July 2013).

Fearon, J.D. (2003) Ethnic and Cultural Diversity by Country. Journal of Economic Growth 8, 195- 222.

Feintrenie, L., Levang, P. (2009) Sumatra’s Rubber Agroforests: Advent, Rise and Fall of a Sustainable Cropping System. Small-Scale Forestry 8, 323-335.

Ferraro, P.J., Hanauer, M.M., Miteva, D.A., Nelson, J.L., Pattanayak, S.K., Nolte, C., Sims, K.R.E. (2015) Estimating the impacts of conservation on ecosystem services and poverty by integrating modeling and evaluation. Proceedings of the National Academy of Sciences 112, 7420-7425.

Fikacek, M. (2012) A new species of leaf-litter inhabiting Georissus from Borneo (Coleoptera: Hydrophiloidea: Georissidae). Zootaxa 3316, 63-68.

Fitzherbert, E.B., Struebig, M.J., Morel, A., Danielsen, F., Brühl, C.A., Donald, P.F., Phalan, B. (2008) How will oil palm expansion affect biodiversity? Trends in Ecology & Evolution 23, 538-545.

121

References

Food and Agriculture Organisation (2010) Global Forest Resource Assessment 2010: Main Report. Food and Agriculture Organization of The United Nations, Rome.

Food and Agriculture Organisation (2011) State of The Worlds Forests. Food and Agriculture Organisation of The United Nation, Rome.

Food and Agriculture Organisation, (2015) FAOSTAT. FAO, Rome, Italy. http://faostat.fao.org/ (accessed 11 November 2015).

Forest Trends, (2014) REDDX: Global State of REDD+ Finance. Forest Trends. http://reddx.forest- trends.org/.

Fox, J., Castella, J.-C., Ziegler, A.D. (2014) Swidden, rubber and carbon: Can REDD+ work for people and the environment in Montane Mainland Southeast Asia? Global Environmental Change 29, 318-326.

Froese, R., Pauly, D., (2012) FishBase. World Wide Web electronic publication, Version (08/2012). www.fishbase.org (accessed 14 March 2013).

Fuller, D.O., Jessup, T.C., Salim, A. (2004) Loss of Forest Cover in Kalimantan, Indonesia, Since the 1997–1998 El Niño. Conservation Biology 18, 249-254.

Galabuzi, C., Eilu, G., Mulugo, L., Kakudidi, E., Tabuti, J.R.S., Sibelet, N. (2014) Strategies for empowering the local people to participate in forest restoration. Agroforestry Systems 88, 719-734.

García-Fernández, C., Casado, M.A. (2005) Forest recovery in managed agroforestry systems: The case of benzoin and rattan gardens in Indonesia. Forest Ecology and Management 214, 158-169.

Gardner, T.A., Von Hase, A., Brownlie, S., Ekstrom, J.M.M., Pilgrim, J.D., Savy, C.E., Stephens, R.T.T., Treweek, J.O., Ussher, G.T., Ward, G., Ten Kate, K. (2013) Biodiversity Offsets and the Challenge of Achieving No Net Loss. Conservation Biology 27, 1254-1264.

Gaveau, D.L., Sheil, D., Husnayaen, Salim, M.A., Arjasakusuma, S., Ancrenaz, M., Pacheco, P., Meijaard, E. (in review) Should Borneo's deforestation be blamed on the oil-palm industry.

Gaveau, D.L.A., Curran, L.M., Paoli, G.D., Carlson, K.M., Wells, P., Besse-Rimba, A., Ratnasari, D., Leader-Williams, N. (2012) Examining protected area effectiveness in Sumatra: importance of regulations governing unprotected lands. Conservation Letters 5, 142-148.

Gaveau, D.L.A., Kshatriya, M., Sheil, D., Sloan, S., Molidena, E., Wijaya, A., Wich, S., Ancrenaz, M., Hansen, M., Broich, M., Guariguata, M.R., Pacheco, P., Potapov, P., Turubanova, S., Meijaard, E. (2013) Reconciling Forest Conservation and Logging in Indonesian Borneo. PLoS ONE 8, e69887.

Gaveau, D.L.A., Salim, M.A., Hergoualc'h, K., Locatelli, B., Sloan, S., Wooster, M., Marlier, M.E., Molidena, E., Yaen, H., DeFries, R., Verchot, L., Murdiyarso, D., Nasi, R., Holmgren, P., Sheil, D. (2014a) Major atmospheric emissions from peat fires in Southeast Asia during non-drought years: evidence from the 2013 Sumatran fires. Nature Scientific Reports 4.

122

References

Gaveau, D.L.A., Sloan, S., Molidena, E., Yaen, H., Sheil, D., Abram, N.K., Ancrenaz, M., Nasi, R., Quinones, M., Wielaard, N., Meijaard, E. (2014b) Four Decades of Forest Persistence, Clearance and Logging on Borneo. PLoS ONE 9, e101654.

Giam, X., Ng, T., Yap, V., Tan, H. (2010) The extent of undiscovered species in Southeast Asia. Biodiversity and Conservation 19, 943-954.

Gibbons, P., Evans, M.C., Maron, M., Gordon, A., Le Roux, D., von Hase, A., Lindenmayer, D.B., Possingham, H.P. (2015) A loss-gain calculator for biodiversity offsets and the circumstances in which no net loss is feasible. Conservation Letters, n/a-n/a.

Gibbs, H.K., Brown, S., Niles, J.O., Foley, J.A. (2007) Monitoring and estimating tropical forest carbon stocks: making REDD a reality. Environmental Research Letters 2, 045023.

Gibbs, H.K., Ruesch, A.S., Achard, F., Clayton, M.K., Holmgren, P., Ramankutty, N., Foley, J.A. (2010) Tropical forests were the primary sources of new agricultural land in the 1980s and 1990s. Proceedings of the National Academy of Sciences 107, 16732-16737.

Gilroy, J.J., Woodcock, P., Edwards, F.A., Wheeler, C., Baptiste, B.L.G., Medina Uribe, C.A., Haugaasen, T., Edwards, D.P. (2014) Cheap carbon and biodiversity co-benefits from forest regeneration in a hotspot of endemism. Nature Climate Change 4, 503-507.

Gingold, B., Rosenbarger, A., Muliastra, Y.I.K.D., Stolle, F., Sudana, I.M., Manessa, M.D.M., Murdimanto, A., Tiangga, S.B., Madusari, C.C., Douard, P., (2012) How to identify degraded land for sustainable palm oil in Indonesia. Working Paper World Resources Institute and Sekala, Washington D.C.

Goldstein, J.H., Pejchar, L., Daily, G.C. (2008) Using return-on-investment to guide restoration: a case study from Hawaii. Conservation Letters 1, 236-243.

Gönner, C., (2000) Resource Management in a Dayak Benuaq Village: Strategies, Dynamics and Prospects. Deutsche Gesellschaft für Technische Zusammenarbeit (GTZ), Eschborn, Germany.

Gönner, C. (2010) Surfing on Waves of Opportunities: Resource Use Dynamics in a Dayak Benuaq Community in East Kalimantan, Indonesia. Society & Natural Resources 24, 165-173.

Gordon, A. (2015) Implementing backcasting for conservation: Determining multiple policy pathways for retaining future targets of endangered woodlands in Sydney, Australia. Biological Conservation 181, 182-189.

Government of Indonesia, (2002) Government Regulation No. 35/2002: Reforestation Fund. Secretary of State of Indonesia, Jakarta, Indonesia.

Government of Indonesia, (2008) Government Regulation No. 78/2008: Forest rehabilitation and reclamation. Secretary of State of Indonesia, Jakarta, Indonesia.

Green, R.E., Cornell, S.J., Scharlemann, J.P.W., Balmford, A. (2005) Farming and the Fate of Wild Nature. Science 307, 550-555.

123

References

Gregory, R., Failing, L., Harstone, M., Long, G., McDaniels, T., Ohlson, D. (2012) Structured Decision Making: A Practical Guide to Environmental Management Choices. John Wiley & Sons, West Sussex, UK.

Griscom, B., Ellis, P., Putz, F.E. (2014) Carbon emissions performance of commercial logging in East Kalimantan, Indonesia. Global Change Biology 20, 923-937.

Groom, M.J., Meffe, G.K., Carroll, C.R. (2006) Principles of conservation biology. Sinauer Associates Sunderland.

Guariguata, M.R., Ostertag, R. (2001) Neotropical secondary forest succession: changes in structural and functional characteristics. Forest Ecology and Management 148, 185-206.

Guerrero, A.M., McAllister, R.R.J., Corcoran, J., Wilson, K.A. (2013) Scale Mismatches, Conservation Planning, and the Value of Social-Network Analyses. Conservation Biology 27, 35-44.

Guerrero, A.M., McAllister, R.R.J., Wilson, K.A. (2015) Achieving Cross-Scale Collaboration for Large Scale Conservation Initiatives. Conservation Letters 8, 107-117.

Gullison, R.E., Frumhoff, P.C., Canadell, J.G., Field, C.B., Nepstad, D.C., Hayhoe, K., Avissar, R., Curran, L.M., Friedlingstein, P., Jones, C.D., Nobre, C. (2007) Tropical Forests and Climate Policy. Science 316, 985-986.

Habib, T.J., Farr, D.R., Schneider, R.R., Boutin, S. (2013) Economic and Ecological Outcomes of Flexible Biodiversity Offset Systems. Conservation Biology 27, 1313-1323.

Hamidy, A., Matsui, M., Nishikawa, K., Belabut, D.M. (2012) Detection of cryptic taxa in Leptobrachium nigrops (Amphibia, Anura, Megophryidae), with description of two new species. Zootaxa 3398, 22-39.

Hansen, M.C., Potapov, P.V., Moore, R., Hancher, M., Turubanova, S.A., Tyukavina, A., Thau, D., Stehman, S.V., Goetz, S.J., Loveland, T.R., Kommareddy, A., Egorov, A., Chini, L., Justice, C.O., Townshend, J.R.G. (2013) High-Resolution Global Maps of 21st-Century Forest Cover Change. Science 342, 850-853.

Hardiansyah, G., (2011) The Potency of Intensive Sylviculture System (TPTII) to Support Reduced Emissions from Deforestation and Forest Degradation (REDD) (A case study in concession of PT. Sari Bumi Kusuma in Central Kalimantan), Graduate School, Bogor Agricultural University, Bogor, Indonesia.

Harris, N.L., Brown, S., Hagen, S.C., Saatchi, S.S., Petrova, S., Salas, W., Hansen, M.C., Potapov, P.V., Lotsch, A. (2012) Baseline Map of Carbon Emissions from Deforestation in Tropical Regions. Science 336, 1573-1576.

Harrison, M.E., Page, S.E., Limin, S. (2009) The global impact of Indonesian forest fires. Biologist 56, 156-163.

124

References

Harrison, R.D., Swinfield, T. (2015) Restoration of logged humid tropical forests: An experimental programme at Harapan Rainforest, Indonesia. Tropical conservation science 8, 4-16.

Heriansyah, I., Hamid, H.A., Nuruddin, A.A., Abdu, A., Ibrahim, S. (2013) Measuring the Short-term Success of Hill Dipterocarp Forest Restoration: The Use of Organic Materials. Journal of Agricultural Science 5, 230-240.

Hidayat, A., Primandari, T., (2015) 50 trillion for first phase of peatland restoration, Tempo. Tempo Inti Media, Jakarta, Indonesia. http://en.tempo.co/read/news/2015/11/26/056722486/50-Tn-for-First- Phase-of-Peatland-Restoration.

Hijmans, R.J., Cameron, S.E., Parra, J.L., Jones, P.G., Jarvis, A. (2005) Very high resolution interpolated climate surfaces for global land areas. International Journal of Climatology 25, 1965- 1978.

Hobbs, R.J., Norton, D.A. (1996) Towards a Conceptual Framework for Restoration Ecology. Restoration Ecology 4, 93-110.

Hoekman, D.H., Vissers, M.A.M., Wielaard, N. (2010) PALSAR Wide-Area Mapping of Borneo: Methodology and Map Validation. Selected Topics in Applied Earth Observations and Remote Sensing, IEEE Journal of 3, 605-617.

Hooijer, A., Page, S., Canadell, J.G., Silvius, M., Kwadijk, J., Wösten, H., Jauhiainen, J. (2010) Current and future CO2 emissions from drained peatlands in Southeast Asia. Biogeosciences 7, 1505- 1514.

Hooijer, A., Silvius, M., Wosten, H., Page, S., (2006) Peat CO2, assessment of CO2 emissions from drained peatland of SE Asia. Delft Hydraulics, Delft, The Netherlands.

Houghton, R.A., (2005) Tropical deforestation as a source of greenhouse gas emissions, in: Mountinho, P., Schwartzman, S. (Eds.), Tropical deforestation and climate change. Istituto de Perquisa Ambiental de Amazônia, Belem, Brasil, pp. 13-21.

Howe, C., Suich, H., Vira, B., Mace, G.M. (2014) Creating win-wins from trade-offs? Ecosystem services for human well-being: A meta-analysis of ecosystem service trade-offs and synergies in the real world. Global Environmental Change 28, 263-275.

Huth, A., Ditzer, T. (2000) Simulation of the growth of a lowland Dipterocarp rain forest with FORMIX3. Ecological Modelling 134, 1-25.

Huth, A., Ditzer, T., Bossel, H. (1997) Rain Forest Growth Model FORMIX3: A Tool for Forest Management Planning Towards Sustainability. Center for Environmental Systems Research, University of Kassel, Kassel, Germany.

Imai, N., Seino, T., Aiba, S.-i., Takyu, M., Titin, J., Kitayama, K. (2012) Effects of selective logging on tree species diversity and composition of Bornean tropical rain forests at different spatial scales. Plant Ecology 213, 1413-1424.

125

References

Indonesian Bureau of Statistics, (2012) Regional income, East Kalimantan in Figures 2011. Indonesian Bureau of Statistics, Samarinda, Indonesia.

Indonesian Bureau of Statistics, (2014) National Census 2010. Indonesian Bureau of Statistics, Jakarta.

Indonesian Bureau of Statistics, (2015a) Indonesian Oil Palm Statistics 2014. Indonesian Bureau of Statistics, Jakarta.

Indonesian Bureau of Statistics, (2015b) National Income of Indonesia. Indonesian Bureau of Statistics, Jakarta, Indonesia.

Indriatmoko, Y., Atmadja, S., Ekaputri, A.D., Komalasari, M. (2014) Rimba Raya Biodiversity Reserve Project, Central Kalimantan, Indonesia. Center for International Forestry Research (CIFOR), Bogor, Indonesia.

IPCC, (2006) IPCC Guidelines for National Greenhouse Gas Inventories. Intergovernmental Panel on Climate Change. http://www.ipcc-nggip.iges.or.jp/public/2006gl/vol4.html. .

Irawan, S., Tacconi, L., Ring, I. (2013) Stakeholders' incentives for land-use change and REDD+: The case of Indonesia. Ecological Economics 87, 75-83.

ISRIC – World Soil Information, (2013) SoilGrids: an automated system for global soil mapping. ISRIC – World Soil Information. http://soilgrids1km.isric.org/ (accessed 02 March 2015).

ITTO (2002) ITTO Guidelines for the Restoration, Management and Rehabilitation of Degraded and Secondary Tropical Forests. ITTO, Yokohama.

IUCN, (2012) IUCN Red List of Threatened Species, Version 2012.1. www.iucnredlist.org (accessed 10 October 2012).

IUCN, (2013) IUCN Red List of Threatened Species, Version 2013.2. www.iucnredlist.org (accessed 10 October 2013).

Jarvis, A., Reuter, H.I., Nelson, A., Guevara, E., (2008) Hole-filled seamless SRTM data V4. International Centre for Tropical Agriculture (CIAT). http://srtm.csi.cgiar.org (accessed 4 February 2015).

Jellinek, S., Rumpff, L., Driscoll, D.A., Parris, K.M., Wintle, B.A. (2014) Modelling the benefits of habitat restoration in socio-ecological systems. Biological Conservation 169, 60-67.

Jenkins, A., Kullander, F.F., Tan, H.H., (2009a) Acrochordonichthys chamaeleon. IUCN Red List of Threatened Species, Version 2012.2. www.iucnredlist.org (accessed 8 February 2013).

Jenkins, A., Kullander, F.F., Tan, H.H., (2009b) Chendol lubricus. IUCN Red List of Threatened Species, Version 2012.2. www.iucnredlist.org (accessed 8 February 2013).

126

References

Jenkins, A., Kullander, F.F., Tan, H.H., (2009c) Pseudomystus myersi. IUCN Red List of Threatened Species, Version 2012.2. www.iucnredlist.org (accessed 8 February 2013).

Joshi, L., Wijaya, K., Sirait, M., Mulyoutami, E., (2004) Indigenous systems and ecological knowledge among Dayak people in Kutai Barat, East Kalimantan–a preliminary report. ICRAF Southeast Asia Working Paper.

Kalimantan Forest Carbon Partnership, (2009) Strategic Peatland Rehabilitation Plan: for Block A (North-West) in the Ex-Mega Rice Project Area, Central Kalimantan. Kalimantan Forests and Climate Partnership, Jakarta, Indonesia.

Kareiva, P., Marvier, M. (2012) What Is Conservation Science? BioScience 62, 962-969.

Kartodihardjo, H., Supriono, A., (2000) The Impact of sectoral development on natural forest conversion and degradation: the case of timber and tree crop plantations in Indonesia. CIFOR, Bogor, Indonesia.

Kasperson, R.E., Dow, K., Archer, E., Caceres, D., Downing, T., Elmqvist, T., Eriksen, S., Folke, C., Han, G., Iyengar, K., Vogel, C., Wilson, K., Ziervogel, G., (2005) Vulnerable Peoples and Places, in: Hassan, R., Scholes, R., Ash, N. (Eds.), Ecosystems and Human Well-Being: Current State and Trends. Island Press, Washington, DC, pp. 143–164.

Katz, E., (1997) NWFPs in Bulungan, East Kalimantan, Indonesia, in: Mittelman, A.J., Lai, C.K., Byron, N., Michon, G., Katz, E. (Eds.), Non-wood forest products outlook study for Asia and the Pacific: towards 2010. Food and Agriculture Organization, Bangkok, Thailand, pp. 38-46.

Kaya, M., Kammesheidt, L., Weidelt, H.J. (2002) The forest garden system of Saparua island Central Maluku, Indonesia, and its role in maintaining tree species diversity. Agroforestry Systems 54, 225- 234.

Kettle, C. (2010) Ecological considerations for using dipterocarps for restoration of lowland rainforest in Southeast Asia. Biodiversity and Conservation 19, 1137-1151.

Kier, G., Mutke, J., Dinerstein, E., Ricketts, T.H., Küper, W., Kreft, H., Barthlott, W. (2005) Global patterns of plant diversity and floristic knowledge. Journal of Biogeography 32, 1107-1116.

Kiesecker, J.M., Copeland, H., Pocewicz, A., McKenney, B. (2009a) Development by design: blending landscape-level planning with the mitigation hierarchy. Frontiers in Ecology and the Environment 8, 261-266.

Kiesecker, J.M., Copeland, H., Pocewicz, A., Nibbelink, N., McKenney, B., Dahlke, J., Holloran, M., Stroud, D. (2009b) A Framework for Implementing Biodiversity Offsets: Selecting Sites and Determining Scale. BioScience 59, 77-84.

Klein, C.J., Wilson, K.A., Watts, M., Stein, J., Carwardine, J., Mackey, B., Possingham, H.P. (2009) Spatial conservation prioritization inclusive of wilderness quality: A case study of Australia’s biodiversity. Biological Conservation 142, 1282-1290.

127

References

Knight, A.T., Rodrigues, A.S.L., Strange, N., Tew, T., Wilson, K.A., (2013) Designing effective solutions to conservation planning problems, in: Macdonald, D.W., Willis, K.J. (Eds.), Key Topics in Conservation Biology 2. John Wiley & Sons, pp. 362-383.

Knight, A.T., Sarkar, S., Smith, R.J., Strange, N., Wilson, K.A. (2011) Engage the hodgepodge: management factors are essential when prioritizing areas for restoration and conservation action. Diversity and Distributions 17, 1234-1238.

Koh, L.P., Ghazoul, J. (2010) Spatially explicit scenario analysis for reconciling agricultural expansion, forest protection, and carbon conservation in Indonesia. Proceedings of the National Academy of Sciences 107, 11140-11144.

Koh, L.P., Ghazoul, J., Butler, R.A., Laurance, W.F., Sodhi, N.S., Mateo-Vega, J., Bradshaw, C.J.A. (2010) Wash and Spin Cycle Threats to Tropical Biodiversity. Biotropica 42, 67-71.

Koh, L.P., Miettinen, J., Liew, S.C., Ghazoul, J. (2011) Remotely sensed evidence of tropical peatland conversion to oil palm. Proceedings of the National Academy of Sciences 108, 5127-5132.

Korpelainen, H., Ådjers, G., Kuusipalo, J., Nuryanto, K., Otsamo, A. (1995) Profitability of rehabilitation of overlogged dipterocarp forest: A case study from South Kalimantan, Indonesia. Forest Ecology and Management 79, 207-215.

Kotiaho, J.S., Moilanen, A. (2015) Conceptual and operational perspectives on ecosystem restoration options in the European Union and elsewhere. Journal of Applied Ecology 52, 816-819.

Kottelat, M. (2012) Rasbora rheophila, a new species of fish from northern Borneo (Teleostei: Cyprinidae). Revue Suisse de Zoologie 119, 77-87.

Kottelat, M., Hui, T.H. (2011) Crossocheilus elegans, a new species of fish from northern Borneo (Teleostei: Cyprinidae). The Raffles Bulletin of Zoology 59, 195-199.

Kottelat, M., Tan, H.H. (2012) Rasbora cryptica, a new species of fish from Sarawak, Borneo (Teleostei: Cyprinidae). Ichthyological Exploration of Freshwaters 23, 37-44.

Kottelat, M., Whitten, A.J., Kartikasari, S.N., Wirjoatmodjo, S. (1996) Freshwater Fishes of Western Indonesia and Sulawesi. Periplus Editions, Hong Kong.

Kujala, H., Whitehead, A.L., Morris, W.K., Wintle, B.A. (2015) Towards strategic offsetting of biodiversity loss using spatial prioritization concepts and tools: A case study on mining impacts in Australia. Biological Conservation 192, 513-521.

Kumagai, T., Saitoh, T.M., Sato, Y., Morooka, T., Manfroi, O.J., Kuraji, K., Suzuki, M. (2004) Transpiration, canopy conductance and the decoupling coefficient of a lowland mixed dipterocarp forest in Sarawak, Borneo: dry spell effects. Journal of Hydrology 287, 237-251.

Laitila, J., Moilanen, A., Pouzols, F.M. (2014) A method for calculating minimum biodiversity offset multipliers accounting for time discounting, additionality and permanence. Methods in Ecology and Evolution 5, 1247-1254.

128

References

Lamb, A., (1991) Orchids of Sabah and Sarawak, in: Kiew, R. (Ed.), The state of nature conservation in Malaysia. Malayan Nature Society & IDRC, Canada., pp. 78-88.

Lamb, D. (2010) Regreening the bare hills: Tropical forest restoration in the Asia-Pacific region. Springer Verlag.

Lamb, D., Erskine, P.D., Parrotta, J.A. (2005) Restoration of Degraded Tropical Forest Landscapes. Science 310, 1628-1632.

Lamb, D., Stanturf, J., Madsen, P., (2012) What Is Forest Landscape Restoration?, in: Stanturf, J., Lamb, D., Madsen, P. (Eds.), Forest Landscape Restoration. Springer Netherlands, pp. 3-23.

Landsberg, J., Sands, P., (2011) The 3-PG Process-Based Model, in: Landsberg, J., Sands, P. (Eds.), Physiological Ecology of Forest Production. Elsevier, pp. 241-282.

Landsberg, J.J., Waring, R.H. (1997) A generalised model of forest productivity using simplified concepts of radiation-use efficiency, carbon balance and partitioning. Forest Ecology and Management 95, 209-228.

Landsberg, J.J., Waring, R.H., Coops, N.C. (2003) Performance of the forest productivity model 3- PG applied to a wide range of forest types. Forest Ecology and Management 172, 199-214.

Langner, A., Miettinen, J., Siegert, F. (2007) Land cover change 2002–2005 in Borneo and the role of fire derived from MODIS imagery. Global Change Biology 13, 2329-2340.

Langner, A., Titin, J., Kitayama, K., (2013) The Application of Satellite Remote Sensing for Classifying Forest Degradation and Deriving Above-Ground Biomass Estimates, in: Kitayama, K. (Ed.), Co-benefits of Sustainable Forestry. Springer Japan, pp. 23-40.

Laurance, W.F., Carolina Useche, D., Rendeiro, J., Kalka, M., Bradshaw, C.J.A., Sloan, S.P., Laurance, S.G., Campbell, M., Abernethy, K., Alvarez, P., et al. (2012) Averting biodiversity collapse in tropical forest protected areas. Nature 489, 290-294.

Law, E.A., Bryan, B.A., Meijaard, E., Mallawaarachchi, T., Struebig, M., Wilson, K.A. (2014) Ecosystem services from a degraded peatland of Central Kalimantan: implications for policy, planning, and management. Ecological Applications 25, 70-87.

Law, E.A., Meijaard, E., Bryan, B.A., Mallawaarachchi, T., Koh, L.P., Wilson, K.A. (2015) Better land-use allocation outperforms land sparing and land sharing approaches to conservation in Central Kalimantan, Indonesia. Biological Conservation 186, 276-286.

Law, E.A., Wilson, K.A. (2015) Providing context for the land-sharing and land-sparing debate. Conservation Letters, n/a-n/a.

Le, H.D., Smith, C., Herbohn, J. (2014) What drives the success of reforestation projects in tropical developing countries? The case of the Philippines. Global Environmental Change 24, 334-348.

129

References

Leaman, D.J., (1996) The medicinal ethnobotany of the Kenyah of East Kalimantan (Indonesian Borneo). University of Ottawa, Ottawa.

Lentini, P.E., Gibbons, P., Carwardine, J., Fischer, J., Drielsma, M., Martin, T.G. (2013) Effect of Planning for Connectivity on Linear Reserve Networks. Conservation Biology 27, 796-807.

Levang, P., Dounias, E., Sitorus, S. (2005) Out of the forest, out of poverty? Forests, Trees and Livelihoods 15, 211-235.

Levang, P., Sitorus, S., Gaveau, D., Sunderland, T. (2012) Landless Farmers, Sly Opportunists, and Manipulated Voters: The Squatters of the Bukit Barisan Selatan National Park (Indonesia). Conservation and Society 10, 243-255.

Limin, S., Susanto, A.R., Gaman, S., Prawira, Y., Darma, F., Rumbih, P., Ermiasi, Y., Kutsin, K., Selviana, E.I., Suparjo, (2008) Buying Living Tree System (BLTS) for Increasing Local Community Welfare and Accelerating the Rehabilitation of Critical Land, International Symposium and Workshop on Tropical Peatland, 19-22 August 2008, Kuching, Malaysia.

Liu, Y.Y., van Dijk, A.I.J.M., de Jeu, R.A.M., Canadell, J.G., McCabe, M.F., Evans, J.P., Wang, G. (2015) Recent reversal in loss of global terrestrial biomass. Nature Climate Change 5, 470-474.

MacKinnon, K., Hatta, G., Halim, H., Mangalik, A. (1996) The Ecology of Kalimantan. Periplus Editions.

Madsen, P., Carroll, N., Moore Brands, K., (2010) State of biodiversity markets report: offset and compensation programs worldwide. Ecosystem Marketplace, Washington, DC.

Mansourian, S., Lamb, D., Gilmour, D., (2005) Overview of Technical Approaches to Restoring Tree Cover at the Site Level, Forest Restoration in Landscapes. Springer New York, pp. 241-249.

Mansourian, S., Vallauri, D. (2014) Restoring Forest Landscapes: Important Lessons Learnt. Environmental Management 53, 241-251.

Marjokorpi, A., Otsamo, R. (2006) Prioritization of Target Areas for Rehabilitation: A Case Study from West Kalimantan, Indonesia. Restoration Ecology 14, 662-673.

Maron, M., Gordon, A., Mackey, B., Posssingham, H.P., Watson, J.E.M. (2015) Stop misuse of biodiversity offsets. Nature 523, 401–403.

Maron, M., Hobbs, R.J., Moilanen, A., Matthews, J.W., Christie, K., Gardner, T.A., Keith, D.A., Lindenmayer, D.B., McAlpine, C.A. (2012) Faustian bargains? Restoration realities in the context of biodiversity offset policies. Biological Conservation 155, 141-148.

Maron, M., Rhodes, J.R., Gibbons, P. (2013) Calculating the benefit of conservation actions. Conservation Letters 6, 359-367.

130

References

Martinez-Harms, M.J., Bryan, B.A., Balvanera, P., Law, E.A., Rhodes, J.R., Possingham, H.P., Wilson, K.A. (2015) Making decisions for managing ecosystem services. Biological Conservation 184, 229-238.

Maswar, Mokhtaruddin, A.M., Majid, N.M., Hanum, I.F., Yusoff, M.K., Azani, A.M., Kobayashi, S., (2001) Evaluation of methods for rehabilitation of logged-over lowland forest in Pasoh, Negeri Sembilan, Malaysia, Rehabilitation of degraded tropical forest ecosystems: workshop proceedings, 2-4 November 1999, Bogor, Indonesia. CIFOR, Bogor, Indonesia, pp. 135-140.

Matsui, M. (2011) Taxonomic revision of one of the Old World's smallest frogs, with description of a new Bornean Microhyla (Amphibia, Microhylidae). Zootaxa 2814, 33-49.

McBride, M.F., Wilson, K.A., Burger, J., Fang, Y.-C., Lulow, M., Olson, D., O’Connell, M., Possingham, H.P. (2010) Mathematical problem definition for ecological restoration planning. Ecological Modelling 221, 2243-2250.

McElwee, P. (2009) Reforesting “Bare Hills” in Vietnam: Social and Environmental Consequences of the 5 Million Hectare Reforestation Program. AMBIO: A Journal of the Human Environment 38, 325-333.

Medrilzam, M., Dargusch, P., Herbohn, J., Smith, C. (2014) The socio-ecological drivers of forest degradation in part of the tropical peatlands of Central Kalimantan, Indonesia. Forestry 87, 335-345.

Meijaard, E., Abram, N.K., Wells, J.A., Pellier, A.-S., Ancrenaz, M., Gaveau, D.L.A., Runting, R.K., Mengersen, K. (2013) People’s Perceptions about the Importance of Forests on Borneo. PLoS ONE 8, e73008.

Meijaard, E., Achdiawan, R., Wan, M., Taber, A. (2014) Rattan: The decline of a once-important non-timber forest product in Indonesia. Center for International Forestry Research (CIFOR), Bogor, Indonesia.

Meijaard, E., Albar, G., Nardiyono, Rayadin, Y., Ancrenaz, M., Spehar, S. (2010a) Unexpected Ecological Resilience in Bornean Orangutans and Implications for Pulp and Paper Plantation Management. PLoS ONE 5, e12813.

Meijaard, E., Buchori, D., Hadiprakarsa, Y., Utami-Atmoko, S.S., Nurcahyo, A., Tjiu, A., Prasetyo, D., Nardiyono, Mengersen, K., et al. (2011) Quantifying Killing of Orangutans and Human- Orangutan Conflict in Kalimantan, Indonesia. PLoS ONE 6, e27491.

Meijaard, E., Cardillo, M., Meijaard, E.M., Possingham, H.P. (2015) Geographic bias in citation rates of conservation research. Conservation Biology, n/a-n/a.

Meijaard, E., Nijman, V. (2003) Primate Hotspots on Borneo: Predictive Value for General Biodiversity and the Effects of Taxonomy. Conservation Biology 17, 725-732.

Meijaard, E., Nijman, V. (2014) Secrecy considerations for conserving Lazarus species. Biological Conservation 175, 21-24.

131

References

Meijaard, E., Sheil, D., (2013) Oil-Palm Plantations in the Context of Biodiversity Conservation, in: Levin, S.A. (Ed.), Encyclopedia of Biodiversity (Second Edition). Academic Press, Waltham, pp. 600-612.

Meijaard, E., Sheil, D., Nasi, R., Augeri, D., Rosenbaum, B., Iskandar, D., Setyawati, T., Lammertink, A., Rachmatika, I., Wong, A., Soehartono, T., Stanley, S., O'Brien, T. (2005) Life after logging: reconciling wildlife conservation and production forestry in Indonesian Borneo. CIFOR, Bogor, Indonesia.

Meijaard, E., Sozer, R. (1996) Reported sightings of crocodiles in Kalimantan. Crocodile Specialist Group Newsletter 15, 12-14.

Meijaard, E., Welsh, A., Ancrenaz, M., Wich, S., Nijman, V., Marshall, A.J. (2010b) Declining Orangutan Encounter Rates from Wallace to the Present Suggest the Species Was Once More Abundant. PLoS ONE 5, e12042.

Meijaard, E., Wich, S., Ancrenaz, M., Marshall, A.J. (2012) Not by science alone: why orangutan conservationists must think outside the box. Annals of the New York Academy of Sciences 1249, 29- 44.

Meijaard, E., Wilson, K.A. (in review) Making oil-palm more beneficial. Science.

Menz, M.H.M., Dixon, K.W., Hobbs, R.J. (2013) Hurdles and Opportunities for Landscape-Scale Restoration. Science 339, 526-527.

Merrill, E.D. (1921) A Bibliographic Enumeration of Bornean Plants. Journal of the Straits Branch of the Royal Asiatic Society Special Number.

Meyfroidt, P., Lambin, E.F. (2011) Global Forest Transition: Prospects for an End to Deforestation. Annual Review of Environment and Resources 36, 343-371.

Michon, G., Aulong, S., Berenger, E., Clement, I., Goloubinoff, M., Katz, E., Sellato, B. (2005) Domesticating forests: how farmers manage forest resources. CIFOR, Bogor, Indonesia.

Millennium Ecosystem Assessment, (2005) Ecosystems and Human Well-being: Synthesis. Island Press, Washington, DC.

Ministry of Agriculture, (2009) Guidelines for oil palm cultivation on peat soils, No.14/Permentan/PL.110/2/2009. Ministry of Agriculture, Jakarta, Indonesia.

Ministry of Agriculture, (2012a) Land suitability criteria for agricultural commodities. Indonesian Agency for Agricultural Research and Development, Ministry of Agriculture, Bogor, Indonesia.

Ministry of Agriculture, (2012b) Land suitability criteria for agricultural commodities: oil palm. Indonesian Agency for Agricultural Research and Development, Ministry of Agriculture, Bogor, Indonesia.

132

References

Ministry of Forestry, (2007) Ministerial Decree P. 37/Menhut-II/2007 on Community Forestry. Ministry of Forestry, Jakarta.

Ministry of Forestry, (2008) Ministerial Decree P. 61/Menhut-II/2008 Terms and procedures for the application of Ecosystem Restoration Concesion in production forest. Ministry of Forestry, Jakarta.

Ministry of Forestry (2011) Forestry Statistics of Indonesia 2010. Ministry of Forestry, Jakarta, Indonesia.

Ministry of Forestry, (2012a) Indonesia's Forest Classification Map 2012. Forestry Planning Agency, Indonesian Ministry of Forestry, Jakarta, Indonesia.

Ministry of Forestry, (2012b) Indonesia's Forest Concession Map 2011. Directorate General of Forest Utilization, Indonesian Ministry of Forestry, Jakarta, Indonesia.

Ministry of Forestry, (2012c) Indonesia's Land Cover Map 2010. Forestry Planning Agency, Indonesian Ministry of Forestry, Jakarta, Indonesia.

Ministry of Forestry, (2012d) Regulation of the Directorate General of Watershed Management and Social Forestry No. P.6/V-SET/2012. Estimation of standard costs of programs for watershed management for the year 2013. The Directorate General of Watershed Management and Social Forestry, Jakarta, Indonesia.

Ministry of Forestry (2013a) Forestry Statistics of Indonesia 2012. Ministry of Forestry, Jakarta, Indonesia.

Ministry of Forestry, (2013b) Regulation of the Ministry of Forestry No.P.9/Menhut-II/2013. Procedure of Implementation, Supporting Activities and Incentive Payment on Rehabilitation of Forest and Land. Ministry of Forestry, Jakarta, Indonesia.

Ministry of Forestry (2014) Forestry Statistics of Indonesia 2013. Ministry of Forestry, Jakarta, Indonesia.

Mittermeier, R.A., Gil, P.R., Hoffman, M., Pilgrim, J.D., Brooks, T., Mittermeier, C.G., Lamoreux, J., da Fonseca, G.A.B. (2005) Hotspots Revisited: Earth's Biologically Richest and Most Endangered Terrestrial Ecoregions. Cemex, Mexico City.

Miyamoto, K., Suzuki, E., Kohyama, T., Seino, T., Mirmanto, E., Simbolon, H. (2003) Habitat differentiation among tree species with small-scale variation of humus depth and topography in a tropical heath forest of Central Kalimantan, Indonesia. Journal of Tropical Ecology 19, 43-54.

Moilanen, A., Kujala, H., Leathwick, J., (2009a) The Zonation framework and software for conservation prioritization, in: Moilanen, A.J., Wilson, K.A., Possingham, H.P. (Eds.), Spatial conservation prioritisation: Quantitative methods and computational tools. Oxford University Press, Oxford, UK, pp. 196-210.

Moilanen, A., Wilson, K., Possingham, H.P. (2009b) Spatial Conservation Prioritization: Quantitative Methods and Computational Tools. Oxford University Press.

133

References

Moilanen, A.J., Pouzols, F.M., Meller, L., Veach, V., Arponen, A., Leppanen, J., Kujala, H. (2014) Zonation Version 4 User Manual. Biodiversity Informatics Conservation Group, University of Helsinki, Helsinki, Finland.

Monela, G.C., Chamshama, S.A.O., Mwaipopo, R., Gamassa, D.M., (2005) A Study on the Social, Economic and Environmental Impacts of Forest Landscape Restoration in Shinyanga Region, Tanzania. Ministry of Natural Resources and Tourism Forestry and Beekeeping Division, The World Conservation Union Eastern Africa Regional Office, Nairobi, Kenya.

Monteith, J.L. (1972) Solar Radiation and Productivity in Tropical Ecosystems. Journal of Applied Ecology 9, 747-766.

Moon, K., Adams, V.M., Januchowski-Hartley, S.R., Polyakov, M., Mills, M., Biggs, D., Knight, A.T., Game, E.T., Raymond, C.M. (2014) A Multidisciplinary Conceptualization of Conservation Opportunity. Conservation Biology 28, 1484-1496.

Morel, A.C., Saatchi, S.S., Malhi, Y., Berry, N.J., Banin, L., Burslem, D., Nilus, R., Ong, R.C. (2011) Estimating aboveground biomass in forest and oil palm plantation in Sabah, Malaysian Borneo using ALOS PALSAR data. Forest Ecology and Management 262, 1786-1798.

Mori, T., (2001) Rehabilitation of degraded forests in lowland Kutai, East Kalimantan, Indonesia, Rehabilitation of degraded tropical forest ecosystems: workshop proceedings, 2-4 November 1999, Bogor, Indonesia. CIFOR, Bogor, Indonesia, pp. 17-36.

Mulyoutami, E., Rismawan, R., Joshi, L. (2009) Local knowledge and management of simpukng (forest gardens) among the Dayak people in East Kalimantan, Indonesia. Forest Ecology and Management 257, 2054-2061.

Murdiyarso, D., Brockhaus, M., Sunderlin, W.D., Verchot, L. (2012) Some lessons learned from the first generation of REDD+ activities. Current Opinion in Environmental Sustainability 4, 678-685.

Murdiyarso, D., Dewi, S., Lawrence, D., Seymour, F. (2011) Indonesia's forest moratorium: A stepping stone to better forest governance? Center for International Forestry Research (CIFOR), Bogor, Indonesia.

Murdiyarso, D., Hergoualc’h, K., Verchot, L.V. (2010) Opportunities for reducing greenhouse gas emissions in tropical peatlands. Proceedings of the National Academy of Sciences 107, 19655-19660.

Murniati, Nawir, A.A., Rumboko, L., Gumartini, T., (2007) The historical national overview and characteristics of rehabilitation initiatives, in: Nawir, A.A., Murniati, Rumboko, L. (Eds.), Forest rehabilitation in Indonesia: Where to After More Than Three Decades? Center for International Forestry Research (CIFOR), Bogor, Indonesia, pp. 75-112.

Murray, J.P., Grenyer, R., Wunder, S., Raes, N., Jones, J.P.G. (2015) Spatial patterns of carbon, biodiversity, deforestation threat, and REDD+ projects in Indonesia. Conservation Biology, n/a-n/a.

Myers, N., Mittermeier, R.A., Mittermeier, C.G., da Fonseca, G.A.B., Kent, J. (2000) Biodiversity hotspots for conservation priorities. Nature 403, 853-858.

134

References

Nagendra, H. (2007) Drivers of reforestation in human-dominated forests. Proceedings of the National Academy of Sciences 104, 15218-15223.

Nanang, M., Inoue, M. (2000) Local forest management in Indonesia: a contradiction between national forest policy and reality. International Review for Environmental Strategies 1, 175-191.

NASA, (2013) NASA Surface meteorology and Solar Energy, National Aeronautics and Space Administration. https://eosweb.larc.nasa.gov/cgi-bin/sse/[email protected] (accessed 30 July 2013).

Nawir, A.A., Murniati, Rumboko, L. (2007a) Forest rehabilitation in Indonesia: where to after three decades? Center for International Forestry Research (CIFOR), Bogor, Indonesia.

Nawir, A.A., Murniati, Rumboko, L., Hiyama, C., Gumartini, T., (2007b) Portraits of rehabilitation projects in Indonesia, in: Nawir, A.A., Murniati, Rumboko, L. (Eds.), Forest rehabilitation in Indonesia: Where to After More Than Three Decades? Center for International Forestry Research (CIFOR), Bogor, Indonesia, pp. 113-175.

Nederlandsch-Indische Vereeniging tot Natuurbescherming (1939) Nature in south and east Borneo. Fauna, flora and conservation in the southern and eastern of Borneo, Batavia.

Nelson, E., Mendoza, G., Regetz, J., Polasky, S., Tallis, H., Cameron, D., Chan, K.M.A., Daily, G.C., Goldstein, J., Kareiva, P.M., Lonsdorf, E., Naidoo, R., Ricketts, T.H., Shaw, M. (2009) Modeling multiple ecosystem services, biodiversity conservation, commodity production, and tradeoffs at landscape scales. Frontiers in Ecology and the Environment 7, 4-11.

Ng, H.H., Kottelat, M. (2012) Chaca serica, a new species of frogmouth catfish (Teleostei: Siluriformes) from southern Borneo. Zootaxa 3,258.

Nightingale, J.M., (2005) Modelling carbon dynamics within tropical rainforest environments using the 3-PG and 3-PGS ecosystem process models, School of Geography, Planning and Environmental Management.

Nightingale, J.M., Hill, M.J., Phinn, S.R., Davies, I.D., Held, A.A., Erskine, P.D. (2008) Use of 3- PG and 3-PGS to simulate forest growth dynamics of Australian tropical rainforests: I. Parameterisation and calibration for old-growth, regenerating and plantation forests. Forest Ecology and Management 254, 107-121.

Nightingale, J.M., Phinn, S.R., Held, A.A. (2004) Ecosystem process models at multiple scales for mapping tropical forest productivity. Progress in Physical Geography 28, 241-281.

Nijman, V. (2004) Effects of habitat disturbance and hunting on densities and biomass of the endemic Hose's leaf monkey Presbytis hosei (Thomas, 1989) (Mammalia: Primates: Cercopithecidae) in east Borneo. Contributions to Zoology 73.

Obidzinski, K., Andriani, R., Komarudin, H., Andrianto, A. (2012) Environmental and Social Impacts of Oil Palm Plantations and the Implications for Biofuel Production in Indonesia. Ecology and Society 17.

135

References

Olson, D.M., Dinerstein, E., Wikramanayake, E.D., Burgess, N.D., Powell, G.V.N., Underwood, E.C., D'Amico, J.A., Itoua, I., Strand, H.E., Morrison, J.C., Loucks, C.J., Allnutt, T.F., Ricketts, T.H., Kura, Y., Lamoreux, J.F., Wettengel, W.W., Hedao, P., Kassem, K.R. (2001) Terrestrial Ecoregions of the World: A New Map of Life on Earth. BioScience 51, 933-938.

Omeja, P.A., Chapman, C.A., Obua, J., Lwanga, J.S., Jacob, A.L., Wanyama, F., Mugenyi, R. (2011) Intensive tree planting facilitates tropical forest biodiversity and biomass accumulation in Kibale National Park, Uganda. Forest Ecology and Management 261, 703-709.

Orsi, F., Church, R.L., Geneletti, D. (2011) Restoring forest landscapes for biodiversity conservation and rural livelihoods: A spatial optimisation model. Environmental Modelling and Software 26, 1622- 1638.

Orsi, F., Geneletti, D. (2010) Identifying priority areas for Forest Landscape Restoration in Chiapas (Mexico): An operational approach combining ecological and socioeconomic criteria. Landscape and Urban Planning 94, 20-30.

Ostrom, E. (2009) A General Framework for Analyzing Sustainability of Social-Ecological Systems. Science 325, 419-422.

Ostrom, E., Cox, M. (2010) Moving beyond panaceas: a multi-tiered diagnostic approach for social- ecological analysis. Environmental Conservation 37, 451-463.

Otsamo, R. (1998) Effect of nurse tree species on early growth of Anisoptera marginata Korth. (Dipterocarpaceae) on an Imperata cylindrica (L.) Beauv. grassland site in South Kalimantan, Indonesia. Forest Ecology and Management 105, 303-311.

Otsamo, R., Otsamo, A., Ådjers, G., (1996) Reforestation experiences with dipterocarp species on grassland., in: Schulte, A., Schöne, D. (Eds.), Dipterocarp forest ecosystems: towards sustainable management. World Scientific Publishing, Singapore, pp. 464-477.

Page, S., Hosciło, A., Wösten, H., Jauhiainen, J., Silvius, M., Rieley, J., Ritzema, H., Tansey, K., Graham, L., Vasander, H., Limin, S. (2009) Restoration Ecology of Lowland Tropical Peatlands in Southeast Asia: Current Knowledge and Future Research Directions. Ecosystems 12, 888-905.

Page, S.E., Rieley, J.O., Banks, C.J. (2011) Global and regional importance of the tropical peatland carbon pool. Global Change Biology 17, 798-818.

Page, S.E., Siegert, F., Rieley, J.O., Boehm, H.-D.V., Jaya, A., Limin, S. (2002) The amount of carbon released from peat and forest fires in Indonesia during 1997. Nature 420, 61-65.

Paoli, G., Curran, L. (2007) Soil Nutrients Limit Fine Litter Production and Tree Growth in Mature Lowland Forest of Southwestern Borneo. Ecosystems 10, 503-518.

Paoli, G., Curran, L., Slik, J. (2008) Soil nutrients affect spatial patterns of aboveground biomass and emergent tree density in southwestern Borneo. Oecologia 155, 287-299.

136

References

Paoli, G.D., Carlson, K.M., Hooijer, A., Page, S.E., Curran, L.M., Wells, P.L., Morrison, R., Jauhiainen, J., Pittman, A.M., Gilbert, D., Lawrence, D. (2011) Policy perils of ignoring uncertainty in oil palm research. Proceedings of the National Academy of Sciences 108, E218.

Paoli, G.D., Wells, P.L., Meijaard, E., Struebig, M.J., Marshall, A.J., Obidzinski, K., Tan, A., Rafiastanto, A., Yaap, B., Slik, J.W.F. (2010) Biodiversity Conservation in the REDD. Carbon balance and management 5, 7.

Paser Statictics Service, (2014a) Paser District in Figures. Paser Statictics Board, Tana Paser, Indonesia. http://paserkab.bps.go.id/index.php?hal=publikasi (accessed 10 December 2014).

Paser Statictics Service, (2014b) Publications by Paser Statistics Service. Paser Statictics Board. http://paserkab.bps.go.id/index.php?hal=publikasi (accessed 10 December 2014), Tana Paser, Indonesia.

Pearson, T.R.H., Brown, S., Casarim, F.M. (2014) Carbon emissions from tropical forest degradation caused by logging. Environmental Research Letters 9, 034017.

Peluso, N.L. (1995) Whose woods are these? Counter-mapping forest territories in Kalimantan, Indonesia. Antipode 27, 383-406.

Peters-Stanley, M., Gonzalez, G., Yin, D., (2013) Covering New Ground: State of the Forest Carbon Markets 2013. Ecosystem Marketplace and Bloomberg New Energy Finance, Washington, DC.

Phalan, B., Onial, M., Balmford, A., Green, R.E. (2011) Reconciling Food Production and Biodiversity Conservation: Land Sharing and Land Sparing Compared. Science 333, 1289-1291.

Phillips, P.D., Brash, T.E., Yasman, I., Subagyo, P., van Gardingen, P.R. (2003) An individual-based spatially explicit tree growth model for forests in East Kalimantan (Indonesian Borneo). Ecological Modelling 159, 1-26.

Phillips, S.J., Dudík, M. (2008) Modeling of species distributions with Maxent: new extensions and a comprehensive evaluation. Ecography 31, 161-175.

Pinard, M.A., Barker, M.G., Tay, J. (2000) Soil disturbance and post-logging forest recovery on bulldozer paths in Sabah, Malaysia. Forest Ecology and Management 130, 213-225.

Poffenberger, M. (2006) People in the forest: community forestry experiences from Southeast Asia. International Journal of Environment and Sustainable Development 5, 57-69.

Possingham, H.P., Bode, M., Klein, C.J. (2015) Optimal Conservation Outcomes Require Both Restoration and Protection. PLoS Biology 13, e1002052.

Potter, L.M. (1996) The dynamics of Imperata: historical overview and current farmer perspectives, with special reference to South Kalimantan, Indonesia. Agroforestry Systems 36, 31-51.

137

References

Pouzols, F.M., Burgman, M.A., Moilanen, A. (2012) Methods for allocation of habitat management, maintenance, restoration and offsetting, when conservation actions have uncertain consequences. Biological Conservation 153, 41-50.

Pouzols, F.M., Moilanen, A. (2013) RobOff: software for analysis of alternative land-use options and conservation actions. Methods in Ecology and Evolution 4, 426-432.

President of the Republic of Indonesia, (1999) Law No.41 on Forestry. President of Republic of Indonesia, Jakarta, Indonesia.

Pressey, R.L., Bottrill, M.C. (2009) Approaches to landscape- and seascape-scale conservation planning: convergence, contrasts and challenges. Oryx 43, 464-475.

Proctor, S., McClean, C., Hill, J. (2011) Protected areas of Borneo fail to protect forest landscapes with high habitat connectivity. Biodiversity and Conservation 20, 2693-2704.

PT. Restorasi Ekosistem Indonesia, (2015) Harapan Rainforest. PT. Restorasi Ekosistem Indonesia, Bogor, Indonesia. http://harapanrainforest.org/.

Putz, F.E., Pinard, M.A. (1993) Reduced-Impact Logging as a Carbon-Offset Method. Conservation Biology 7, 755-757.

Putz, F.E., Zuidema, P.A., Synnott, T., Pena-Claros, M., Pinard, M.A., Sheil, D., Vanclay, J.K., Sist, P., Gourlet-Fleury, S., Griscom, B., Palmer, J., Zagt, R. (2012) Sustaining conservation values in selectively logged tropical forests: the attained and the attainable. Conservation Letters 5, 296-303.

R Development Core Team (2013) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. http://www.R-project.org/.

Rabinowitz, A. (1995) Helping a Species Go Extinct: The Sumatran Rhino in Borneo. Conservation Biology 9, 482-488.

Raes, N., (2009) Borneo: a quantitative analysis of botanical richness, endemicity and floristic regions based on herbarium records, Faculteit der Wiskunde en Natuurwetenschappen Leiden University, National Herbarium of the Netherlands.

Raes, N., Cannon, C.H., Hijmans, R.J., Piessens, T., Saw, L.G., van Welzen, P.C., Slik, J.W.F. (2014) Historical distribution of Sundaland’s Dipterocarp rainforests at Quaternary glacial maxima. Proceedings of the National Academy of Sciences 111, 16790-16795.

Raes, N., Roos, M.C., Slik, J.W.F., van Loon, E.E., ter Steege, H. (2009) Botanical richness and endemicity patterns of Borneo derived from species distribution models. Ecography 32, 180-192.

Rahmawati, A. (2013) An economic analysis of ecosystem restoration concession policy in Indonesia: a new strategy for sustainable forest management? International Journal of Green Economics 7, 56- 70.

138

References

Reid, V.W., (1992) How many species will there be?, in: Whitmore, T.C., Sayer, J. (Eds.), Tropical Deforestation and Species Extinction. Chapman & Hall, London, UK, pp. 55–73.

Renwick, A.R., Robinson, C.J., Martin, T.G., May, T., Polglase, P., Possingham, H.P., Carwardine, J. (2014) Biodiverse Planting for Carbon and Biodiversity on Indigenous Land. PLoS ONE 9, e91281.

Resosudarmo, I.A.P., (in press) REDD and Forest Community, in: Resosudarmo, B.P., Imansyah, M.H., Napitupulu, L. (Eds.), Development, Environment and the People of Kalimantan. Indonesia Regional Science Association (IRSA), Jakarta, Indonesia.

Ritung, S., Wahyunto, Agus, F., Hidayat, H., (2007) Land Suitability Evaluation with a case map of Aceh Barat District. Indonesian Soil Research Institute and World Agroforestry Centre, Bogor, Indonesia.

Rondinini, C., Di Marco, M., Chiozza, F., Santulli, G., Baisero, D., Visconti, P., Hoffmann, M., Schipper, J., Stuart, S.N., Tognelli, M.F., Amori, G., Falcucci, A., Maiorano, L., Boitani, L. (2011) Global habitat suitability models of terrestrial mammals. Philosophical Transactions of the Royal Society B: Biological Sciences 366, 2633-2641.

Roos, M.C., Keßler, P.J.A., Robbert Gradstein, S., Baas, P. (2004) Species diversity and endemism of five major Malesian islands: diversity–area relationships. Journal of Biogeography 31, 1893-1908.

Rosenzweig, M.L. (1968) Net Primary Productivity of Terrestrial Communities: Prediction from Climatological Data. The American Naturalist 102, 67-74.

RSPO, (2013) Principles and Criteria for the Production of Sustainable Palm Oil. The Roundtable on Sustainable Palm Oil, Kuala Lumpur, Malaysia. http://www.rspo.org/resources/key- documents/certification/rspo-principles-and-criteria.

RSPO, (2014) RSPO Remediation and Compensation Procedures. The Roundtable on Sustainable Palm Oil, Kuala Lumpur, Malaysia. http://www.rspo.org/news-and-events/announcements/rspo- remediation-and-compensation-procedures.

Rudel, T.K., Coomes, O.T., Moran, E., Achard, F., Angelsen, A., Xu, J., Lambin, E. (2005) Forest transitions: towards a global understanding of land use change. Global Environmental Change 15, 23-31.

Running, S., Thornton, P., Nemani, R., Glassy, J., (2000) Global Terrestrial Gross and Net Primary Productivity from the Earth Observing System, in: Sala, O., Jackson, R., Mooney, H., Howarth, R. (Eds.), Methods in Ecosystem Science. Springer New York, pp. 44-57.

Runting, R.K., Meijaard, E., Abram, N.K., Wells, J.A., Gaveau, D.L., Ancrenaz, M., Posssingham, H.P., Wich, S.A., Ardiansyah, F., Gumal, M.T., Ambu, L.N., Wilson, K.A. (2015) Alternative futures for Borneo show the value of integrating economic and conservation targets across borders. Nature Communications 6, 6819.

Ruslandi, Venter, O., Putz, F.E. (2011) Overestimating conservation costs in Southeast Asia. Frontiers in Ecology and the Environment 9, 542-544.

139

References

Saatchi, S.S., Harris, N.L., Brown, S., Lefsky, M., Mitchard, E.T.A., Salas, W., Zutta, B.R., Buermann, W., Lewis, S.L., Hagen, S., Petrova, S., White, L., Silman, M., Morel, A. (2011) Benchmark map of forest carbon stocks in tropical regions across three continents. Proceedings of the National Academy of Sciences 108, 9899-9904.

Saner, P., Loh, Y.Y., Ong, R.C., Hector, A. (2012) Carbon Stocks and Fluxes in Tropical Lowland Dipterocarp Rainforests in Sabah, Malaysian Borneo. PLoS ONE 7, e29642.

Santika, T., McAlpine, C.A., Lunney, D., Wilson, K.A., Rhodes, J.R. (2015a) Assessing spatiotemporal priorities for species’ recovery in broad-scale dynamic landscapes. Journal of Applied Ecology 52, 832-840.

Santika, T., Meijaard, E., Wilson, K.A. (2015b) Designing multifunctional landscapes for forest conservation. Environmental Research Letters 10, 114012.

Saragih, B., (2011) Economic value of non-timber forest products among Paser Indigenous People of East Kalimantan, Centre for Environmental Science (CML), Faculty of Social and Behavioural Sciences, Leiden University. Leiden University, Leiden, the Netherland.

Sardjono, M.A., Devung, S., Imang, N., (2013) Local Community Dimension of Indonesian Forest Policy and Customary Land Tenure in East Kalimantan, Commoners and the Changing Commons: Livelihoods, Environmental Security, and Shared Knowledge, the Fourteenth Biennial Conference of the International Association for the Study of the Commons, Mt. Fuji, Japan.

Sargis, E.J., Woodman, N., Reese, A.T., Olson, L.E. (2013) Using hand proportions to test taxonomic boundaries within the Tupaia glis species complex (Scandentia, Tupaiidae). Journal of Mammalogy 94, 183-201.

Sasaki, N., Asner, G., Knorr, W., Durst, P., Priyadi, H., Putz, F. (2011) Approaches to classifying and restoring degraded tropical forests for the anticipated REDD+ climate change mitigation mechanism. iForest - Biogeosciences and Forestry 4, 1-6.

Sasaki, N., Putz, F.E. (2009) Critical need for new definitions of “forest” and “forest degradation” in global climate change agreements. Conservation Letters 2, 226-232.

Satriastanti, F.E., (2012) SBY sees Kalimantan as the "lungs of the world". http://jakartaglobe.beritasatu.com/archive/sby-sees-kalimantan-as-the-lungs-of-the-world/, Jakarta Globe, Jakarta.

Sayer, J., (2005) Goals and Targets of Forest Landscape Restoration, Forest Restoration in Landscapes. Springer New York, pp. 101-108.

Sayer, J., Margules, C., Boedhihartono, A., Dale, A., Sunderland, T., Supriatna, J., Saryanthi, R. (2015) Landscape approaches; what are the pre-conditions for success? Sustainability Science 10, 345-355.

Sayer, J., Sunderland, T., Ghazoul, J., Pfund, J.-L., Sheil, D., Meijaard, E., Venter, M., Boedhihartono, A.K., Day, M., Garcia, C., van Oosten, C., Buck, L.E. (2013) Ten principles for a

140

References landscape approach to reconciling agriculture, conservation, and other competing land uses. Proceedings of the National Academy of Sciences 110, 8349-8356.

Schaafsma, M., Morse-Jones, S., Posen, P., Swetnam, R.D., Balmford, A., Bateman, I.J., Burgess, N.D., Chamshama, S.A.O., Fisher, B., Turner, R.K., et al. (2014) The importance of local forest benefits: Economic valuation of Non-Timber Forest Products in the Eastern Arc Mountains in Tanzania. Global Environmental Change 24, 295-305.

Schaeffner, B.C., Beveridge, I. (2012a) Cavearhynchus, a new genus of tapeworm (Cestoda: Trypanorhyncha: Pterobothriidae) from Himantura lobistoma Manjaji-Matsumoto & Last, 2006 (Rajiformes) off Borneo, including redescriptions and new records of species of Pterobothrium Diesing, 1850. Systematic parasitology 82, 147-165.

Schaeffner, B.C., Beveridge, I. (2012b) Description of a new Trypanorhynch species (Cestoda) from Indonesian Borneo, with the suppression of Oncomegoides and the erection of a new genus Hispidorhynchus. Journal of Parasitology 98, 408-414.

Schaeffner, B.C., Beveridge, I. (2012c) Prochristianella Dollfus, 1946 (Trypanorhyncha: Eutetrarhynchidae) from elasmobranchs off Borneo and Australia, including new records and the description of four new species. Zootaxa 3505, 1-25.

Scriven, S.A., Hodgson, J.A., McClean, C.J., Hill, J.K. (2015) Protected areas in Borneo may fail to conserve tropical forest biodiversity under climate change. Biological Conservation 184, 414-423.

Shearman, P., Bryan, J., Laurance, W.F. (2012) Are we approaching ‘peak timber’ in the tropics? Biological Conservation 151, 17-21.

Sheil, D., Kartawinata, K., Samsoedin, I., Priyadi, H., Afriastini, J.J. (2010) The lowland forest tree community in Malinau, Kalimantan (Indonesian Borneo): results from a one-hectare plot. Plant Ecology & Diversity 3, 59-66.

Shimada, T., Matsui, M., Yambun, P., Sudin, A. (2011) A survey of morphological variation in adult Meristogenys amoropalamus (Amphibia, Anura, Ranidae), with a description of a new cryptic species. Zootaxa 2905, 33-56.

Shoo, L.P., Scarth, P., Schmidt, S., Wilson, K.A. (2013) Reclaiming Degraded Rainforest: A Spatial Evaluation of Gains and Losses in Subtropical Eastern Australia to Inform Future Investment in Restoration. Restoration Ecology 21, 481-489.

Sihombing, J.A., (2011) Non-timber forest products (NTFPs) utilization by forest communities around forest in IUPHHK-HA PT. Ratah Timber, East Kalimantan, Faculty of Forestry, Bogor Agricultural University, Bogor, Indonesia.

Silalahi, M., Erwin, D. (2015) Collaborative Conflict Management on Ecosystem Restoration Concession: lessons Learnt from Harapan Rainforest Jambi-South Sumatra-Indonesia. Forest Research 4, 134.

141

References

Silver, W.L., Ostertag, R., Lugo, A.E. (2000) The Potential for Carbon Sequestration Through Reforestation of Abandoned Tropical Agricultural and Pasture Lands. Restoration Ecology 8, 394- 407.

Siregar, M. (2006) Species Diversity of Local Fruit Trees in Kalimantan: Problems of Conservation and Its Development. Biodiversitas 7.

Slik, J., Bernard, C., Van Beek, M., Breman, F., Eichhorn, K. (2008) Tree diversity, composition, forest structure and aboveground biomass dynamics after single and repeated fire in a Bornean rain forest. Oecologia 158, 579-588.

Slik, J.W.F., Aiba, S.I., Brearley, F.Q., Cannon, C.H., Forshed, O., Kitayama, K., Nagamasu, H., Nilus, R., Payne, J., Paoli, G., Poulsen, A.D., Raes, N., Sheil, D., Sidiyasa, K., Suzuki, E., van Valkenburg, J.L.C.H. (2010) Environmental correlates of tree biomass, basal area, wood specific gravity and stem density gradients in Borneo's tropical forests. Global Ecology and Biogeography 19, 50-60.

Slik, J.W.F., Arroyo-Rodríguez, V., Aiba, S.-I., Alvarez-Loayza, P., Alves, L.F., Ashton, P., Balvanera, P., Bastian, M.L., Bellingham, P.J., van den Berg, E., et al. (2015) An estimate of the number of tropical tree species. Proceedings of the National Academy of Sciences.

Slik, J.W.F., Paoli, G., McGuire, K., Amaral, I., Barroso, J., Bastian, M., Blanc, L., Bongers, F., Boundja, P., Clark, C., et al. (2013) Large trees drive forest aboveground biomass variation in moist lowland forests across the tropics. Global Ecology and Biogeography 22, 1261-1271.

Slik, J.W.F., Poulsen, A.D., Ashton, P.S., Cannon, C.H., Eichhorn, K.A.O., Kartawinata, K., Lanniari, I., Nagamasu, H., Nakagawa, M., Van Nieuwstadt, M.G.L., Payne, J., Saridan, A., Sidiyasa, K., Verburg, R.W., Webb, C.O., Wilkie, P. (2003) A floristic analysis of the lowland dipterocarp forests of Borneo. Journal of Biogeography 30, 1517-1531.

Slik, J.W.F., Raes, N., Aiba, S.-I., Brearley, F.Q., Cannon, C.H., Meijaard, E., Nagamasu, H., Nilus, R., Paoli, G., Poulsen, A.D., Sheil, D., Suzuki, E., Van Valkenburg, J.L.C.H., Webb, C.O., Wilkie, P., Wulffraat, S. (2009) Environmental correlates for tropical tree diversity and distribution patterns in Borneo. Diversity and Distributions 15, 523-532.

Smedley, T., (2015) If the palm oil industry waited for consumers to care, sustainability would get nowhere The Guardian. http://www.theguardian.com/sustainable-business/2015/oct/26/palm-oil- industry-consumer-understanding-sustainability-cspo-packaging-marks-spencer-boots-ecover.

Smit, H.H., Meijaard, E., van der Laan, C., Mantel, S., Budiman, A., Verweij, P. (2013) Breaking the Link between Environmental Degradation and Oil Palm Expansion: A Method for Enabling Sustainable Oil Palm Expansion. PLoS ONE 8, e68610.

Sodhi, N.S., Koh, L.P., Brook, B.W., Ng, P.K.L. (2004) Southeast Asian biodiversity: an impending disaster. Trends in Ecology & Evolution 19, 654-660.

Sodhi, N.S., Lee, T.M., Koh, L.P., Brook, B.W. (2009) A Meta-Analysis of the Impact of Anthropogenic Forest Disturbance on Southeast Asia's Biotas. Biotropica 41, 103-109.

142

References

Soehartono, T., Susilo, H.D., Andayani, N., Atmoko, S.S.U., Sihite, J., Saleh, C., Sutrisno, A., (2008) Orangutan Indonesia Conservation Strategies and Action Plan. Minstry of Forestry, Jakarta.

Soejono, Budiharta, S., Arisoesilaningsih, A. (2013) Proposing local trees diversity for rehabilitation of degraded lowland areas surrounding springs. Biodiversitas 14, 37-42.

Soekotjo (2009) Teknik Silvikultur Intensif (SILIN). Gadjah Mada University Press, Yogyakarta.

Soepadmo, E., Wong, K.M. (1995) Tree Flora of Sabah and Sarawak: Volume One. Forest Research Institute (FRIM), Sabah Forestry Department and Sarawak Forestry Department, Kuala Lumpur, Malaysia.

Soerianegara, I., Lemmens, R. (1994) Plant Genetic Resources of South East Asia 5 (1): Timber trees: Major commercial timbers. PROSEA Bogor, Indonesia.

Solheim, E., Natalegawa, R.M.M.M., (2010) Letter of intent between the Government of the Kingdom of Norway and the Government of the Republic of Indonesia on cooperation on reducing greenhouse gas emissions from deforestation and forest degradation. Oslo, Norway, 26th of May 2010., Oslo, Norway. http://www.regjeringen.no/upload/SMK/Vedlegg/2010/Indonesia_avtale.pdf.

Sonter, L.J., Barrett, D.J., Soares-Filho, B.S. (2014) Offsetting the Impacts of Mining to Achieve No Net Loss of Native Vegetation. Conservation Biology, n/a-n/a.

Stanturf, J.A., Palik, B.J., Williams, M.I., Dumroese, R.K., Madsen, P. (2014) Forest Restoration Paradigms. Journal of Sustainable Forestry 33, S161-S194.

Starling Resources, (2014) Katingan Peat Forest Restoration Project.

Struebig, M., Christy, L., Pio, D., Meijaard, E. (2010) Bats of Borneo: diversity, distributions and representation in protected areas. Biodiversity and Conservation 19, 449-469.

Struebig, M.J., Fischer, M., Gaveau, D.L., Meijaard, E., Wich, S.A., Gonner, C., Sykes, R., Wilting, A., Kramer-Schadt, S. (2015a) Anticipated climate and land-cover changes reveal refuge areas for Borneo's orang-utans. Global Change Biology 21, 2891-2904.

Struebig, Matthew J., Wilting, A., Gaveau, D.L.A., Meijaard, E., Smith, Robert J., Abdullah, T., Abram, N., Alfred, R., Ancrenaz, M., Kramer-Schadt, S., et al. (2015b) Targeted Conservation to Safeguard a Biodiversity Hotspot from Climate and Land-Cover Change. Current Biology 25, 372- 378.

Suding, K.N. (2011) Toward an Era of Restoration in Ecology: Successes, Failures, and Opportunities Ahead. Annual Review of Ecology, Evolution, and Systematics 42, 465-487.

Sumarga, E., Hein, L. (2015) Benefits and costs of oil palm expansion in Central Kalimantan, Indonesia, under different policy scenarios. Regional Environmental Change, 1-11.

Sumarga, E., Hein, L., Edens, B., Suwarno, A. (2015) Mapping monetary values of ecosystem services in support of developing ecosystem accounts. Ecosystem Services 12, 71-83.

143

References

Terauchi, D., Imang, N., Nanang, M., Kawai, M., Sardjono, M.A., Pambudhi, F., Inoue, M. (2014) Implication for Designing a REDD+ Program in a Frontier of Oil Palm Plantation Development: Evidence in East Kalimantan, Indonesia. Open Journal of Forestry 4, 259-277.

The Guardian, (2013) World's largest Redd project finally approved in Indonesia, The Guardian. http://www.theguardian.com/environment/2013/may/31/worlds-largest-redd-project-indonesia.

Thompson, I.D., Guariguata, M.R., Okabe, K., Bahamondez, C., Nasi, R., Heymell, V., Sabogal, C. (2013) An Operational Framework for Defining and Monitoring Forest Degradation. Ecology and Society 18.

Thomson, J.R., Moilanen, A.J., Vesk, P.A., Bennett, A.F., Nally, R.M. (2009) Where and when to revegetate: a quantitative method for scheduling landscape reconstruction. Ecological Applications 19, 817-828.

Tilman, D., Balzer, C., Hill, J., Befort, B.L. (2011) Global food demand and the sustainable intensification of agriculture. Proceedings of the National Academy of Sciences 108, 20260-20264.

Timmins, R.J., Duckworth, J.W., Hedges, S., Steinmetz, R., Pattanavibool, A., (2012) Bos javanicus. IUCN Red List of Threatened Species, Version 2012.2. www.iucnredlist.org (accessed 8 February 2013).

Townshend, J.R.G., Carroll, M., Dimiceli, C., Sohlberg, S., Hansen, M.C., DeFries, R., (2011) Vegetation Continuous Fields MOD44B, 2010 Percent Tree Cover, Collection 5, in: University of Maryland, C.P., Maryland (Ed.), Maryland.

UNFCCC, (2007) Bali Action Plan, FCCC/CP/2007/6/Add.1. United Nations Framework Convention on Climate Change (UNFCCC), Bonn, Germany. http://unfccc.int/resource/docs/2007/cop13/eng/06a01.pdf#page=3.

UNFCCC, (2013) National Inventory Submissions 2013. United Nations Framework Convention on Climate Change (UNFCCC), Bonn, Germany. http://unfccc.int/national_reports/annex_i_ghg_inventories/national_inventories_submissions/items/ 7383.php.

United Nations, (1998) Kyoto Protocol to the United Nations Framework Convention on Climate Change. United Nations, Kyoto, Japan. http://unfccc.int/kyoto_protocol/items/2830.php.

Van der Laan, C., Verweij, P., Quiñones, M., Faaij, A. (2014) Analysis of biophysical and anthropogenic variables and their relation to the regional spatial variation of aboveground biomass illustrated for North and East Kalimantan, Borneo. Carbon balance and management 9, 1-12. van der Werf, G.R., Morton, D.C., DeFries, R.S., Olivier, J.G.J., Kasibhatla, P.S., Jackson, R.B., Collatz, G.J., Randerson, J.T. (2009) CO2 emissions from forest loss. Nature Geoscience 2, 737-738. van Dijk, A.I.J.M., Keenan, R.J. (2007) Planted forests and water in perspective. Forest Ecology and Management 251, 1-9.

144

References

Van Dijk, A.I.J.M., Van Noordwijk, M., Calder, I.R., Bruijnzeel, S.L.A., Schellekens, J., Chappell, N.A. (2009) Forest–flood relation still tenuous – comment on ‘Global evidence that deforestation amplifies flood risk and severity in the developing world’ by C. J. A. Bradshaw, N.S. Sodi, K. S.-H. Peh and B.W. Brook. Global Change Biology 15, 110-115. van Gardingen, P.R., McLeish, M.J., Phillips, P.D., Fadilah, D., Tyrie, G., Yasman, I. (2003) Financial and ecological analysis of management options for logged-over Dipterocarp forests in Indonesian Borneo. Forest Ecology and Management 183, 1-29.

Venter, M., Venter, O., Laurance, S., Bird, M., (2012) Recarbonization of the Humid Tropics, in: Lal, R., Lorenz, K., Hüttl, R.F., Schneider, B.U., Braun, J. (Eds.). Springer Netherlands, pp. 229-252.

Venter, O., Laurance, W.F., Iwamura, T., Wilson, K.A., Fuller, R.A., Possingham, H.P. (2009a) Harnessing Carbon Payments to Protect Biodiversity. Science 326, 1368.

Venter, O., Meijaard, E., Possingham, H., Dennis, R., Sheil, D., Wich, S., Hovani, L., Wilson, K. (2009b) Carbon payments as a safeguard for threatened tropical mammals. Conservation Letters 2, 123-129.

Venter, O., Possingham, H.P., Hovani, L., Dewi, S., Griscom, B., Paoli, G., Wells, P., Wilson, K.A. (2013) Using systematic conservation planning to minimize REDD+ conflict with agriculture and logging in the tropics. Conservation Letters 6, 116-124.

Vesk, P.A., Nolan, R., Thomson, J.R., Dorrough, J.W., Nally, R.M. (2008) Time lags in provision of habitat resources through revegetation. Biological Conservation 141, 174-186.

Vira, B., Agarwal, B., Jamnadass, R.H., Kleinschmit, D., McMullin, S., Mansourian, S., Neufeldt, H., Parrotta, J.A., Sunderland, T.C.H., Wildburger, C. (2015) Introduction: Forests, Trees and Landscapes for Food Security and Nutrition. International Union of Forest Research Organizations (IUFRO).

Waring, R.H., (2013) Richard Waring personal website. Oregon State University, Oregon State University.

Watson, J.E.M., Dudley, N., Segan, D.B., Hockings, M. (2014) The performance and potential of protected areas. Nature 515, 67-73.

Wells, J.A., Meijaard, E., Abram, N., Wich, S., (2013) Forests, floods, people and wildlife on Borneo. UNEP, Bangkok, Thailand.

White, J.D., Coops, N.C., Scott, N.A. (2000) Estimates of New Zealand forest and scrub biomass from the 3-PG model. Ecological Modelling 131, 175-190.

White, J.D., Scott, N.A., Hirsch, A.I., Running, S.W. (2006) 3-PG Productivity Modeling of Regenerating Amazon Forests: Climate Sensitivity and Comparison with MODIS-Derived NPP. Earth Interactions 10, 1-26.

145

References

Whitten, T., van Dijk, P., Curran, L., Meijaard, E., Supriatna, J., Ellis, S. (2004) Sundaland. Hotspots revisited: Another look at Earth’s richest and most endangered terrestrial ecoregions. Mexico: Cemex.

Wibisono, I., Siboro, L., Suryadiputra, I.N.N. (2005) Guidelines and Silviculture Techniques in Peatland Rehabilitation. Wetlands International-Indonesia Programme, Bogor, Indonesia.

Wich, S.A., Gaveau, D., Abram, N., Ancrenaz, M., Baccini, A., Brend, S., Curran, L., Delgado, R.A., Meijaard, E., et al. (2012) Understanding the Impacts of Land-Use Policies on a Threatened Species: Is There a Future for the Bornean Orang-utan? PLoS ONE 7, e49142.

Wikramanayake, E., Dinerstein, E., Loucks, C., Olson, D., Morrison, J., Lamoreux, J., McKnight, M., Hedao, P. (2002) Terrestrial ecoregions of the Indo-Pacific: a conservation assessment. Island Press, Washington, Covelo & London.

Wilson, E.O., (1988) The current state of biological diversity, in: Wilson, E.O., Peter, F.M. (Eds.), Biodiversity. National Academy Press, Washington, pp. 3-18.

Wilson, H.B., Meijaard, E., Venter, O., Ancrenaz, M., Possingham, H.P. (2014) Conservation Strategies for Orangutans: Reintroduction versus Habitat Preservation and the Benefits of Sustainably Logged Forest. PLoS ONE 9, e102174.

Wilson, K., Auerbach, N.A., Sam, K., Magini, A.G., Moss, A.S.L., Laghans, S.D., Budiharta, S., Terzano, D., Meijaard, E. (2016) Conservation research is not happening where it is most needed. PLoS Biology 14, e1002413.

Wilson, K.A., Lulow, M., Burger, J., Fang, Y.-C., Andersen, C., Olson, D., O’Connell, M., McBride, M.F. (2011) Optimal restoration: accounting for space, time and uncertainty. Journal of Applied Ecology 48, 715-725.

Wilson, K.A., Lulow, M., Burger, J., McBride, M.F., (2012) The Economics of Restoration, in: Stanturf, J., Lamb, D., Madsen, P. (Eds.), Forest Landscape Restoration. Springer Netherlands, pp. 215-231.

Wilson, K.A., Meijaard, E., Drummond, S., Grantham, H.S., Boitani, L., Catullo, G., Christie, L., Dennis, R., Dutton, I., Falcucci, A., Maiorano, L., Possingham, H.P., Rondinini, C., Turner, W.R., Venter, O., Watts, M. (2010) Conserving biodiversity in production landscapes. Ecological Applications 20, 1721-1732.

Wilting, A., Christiansen, P., Kitchener, A.C., Kemp, Y.J.M., Ambu, L., Fickel, J. (2011) Geographical variation in and evolutionary history of the Sunda clouded leopard (Neofelis diardi) (Mammalia: Carnivora: Felidae) with the description of a new subspecies from Borneo. Molecular Phylogenetics and Evolution 58, 317-328.

Wilting, A., Cord, A., Hearn, A.J., Hesse, D., Mohamed, A., Traeholdt, C., Cheyne, S.M., Sunarto, S., Jayasilan, M.-A., Ross, J., Shapiro, A.C., Sebastian, A., Dech, S., Breitenmoser, C., Sanderson, J., Duckworth, J.W., Hofer, H. (2010) Modelling the Species Distribution of Flat-Headed Cats (Prionailurus planiceps), an Endangered South-East Asian Small Felid. PLoS ONE 5, e9612.

146

References

Wong, K., (1998) Patterns of plant endemism and rarity in Borneo and the Malay Peninsula, in: Peng, C.I., Lowry II, P.P. (Eds.), Rare, threatened, and endangered floras of Asia and the Pacific rim. Institute of Botany, Academia Sinica Monograph Series, Taiwan, pp. 139-169.

Wood, J.J., Cribb, P.J. (1994) A check-list of the orchids of Borneo. Royal Botanic Gardens, Kew, UK.

Woodwell, G.M., Hobbie, J.E., Houghton, R.A., Melillo, J.M., Moore, B., Peterson, B.J., Shaver, G.R. (1983) Global Deforestation: Contribution to Atmospheric Carbon Dioxide. Science 222, 1081- 1086.

World Resources Institute (2002) State of the forest: Indonesia. WRI, Washington, DC.

Wuethrich, B. (2007) Reconstructing Brazil's Atlantic Rainforest. Science 315, 1070-1072.

Wunder, S., Angelsen, A., Belcher, B. (2014) Forests, Livelihoods, and Conservation: Broadening the Empirical Base. World Development 64, S1-S11.

Zahawi, R.A., Holl, K.D. (2009) Comparing the Performance of Tree Stakes and Seedlings to Restore Abandoned Tropical Pastures. Restoration Ecology 17, 854-864.

Zahawi, R.A., Reid, J.L., Holl, K.D. (2014) Hidden Costs of Passive Restoration. Restoration Ecology 22, 284-287.

Ziegler, A.D., Phelps, J., Yuen, J.Q.I., Webb, E.L., Lawrence, D., Fox, J.M., Bruun, T.B., Leisz, S.J., Ryan, C.M., Dressler, W., Mertz, O., Pascual, U., Padoch, C., Koh, L.P. (2012) Carbon outcomes of major land-cover transitions in SE Asia: great uncertainties and REDD+ policy implications. Global Change Biology 18, 3087-3099.

147

Appendices

148

Appendices

Appendices for Chapter 3

Supplementary Information

Appendix S3.1. The process equations of 3-PG model

ENVIRONMENTAL MODIFIERS SOIL-WATER BALANCE

Temperature modifier fT (Ta) = Water balance ∆휽푺 = (1 − 푖푅)푅푃 + 푅퐼 − 퐸푇

(푇 −푇 )⁄(푇 −푇 ) 푇 −푇 푇 −푇 푚푎푥 표푝푡 표푝푡 푚푖푛 Fraction rainfall intercepted 풊 = 푖 min{퐿⁄퐿 , 1} ( 푎 푚푖푛 ) ( 푚푎푥 푎 ) 푹 푅푝 푔퐶 푇표푝푡−푇푚푖푛 푇푚푎푥−푇표푝푡

Evapotranspiration (푬푻) using the Penman-Monteith 1 Soil-water modifier fSW (θ) = 푛 equation with stomatal conductance 품푪 = 1+ [(1−휃⁄휃푥)/푐휃] 휃 푔퐶푥휑 min{퐿⁄퐿푔퐶 , 1} −푘 퐷 VPD modifier fVPD (D) = 푒 퐷

1 Age modifier fage (t) = 푛푎푔푒 1+(푡⁄푟푎푔푒푡푥) BIOMASS PRODUCTION

−푘퐿 Intercepted radiation Qint = (1 − 푒 )푄0

Canopy quantum efficiency αC = fT min {푓 , 푓 }푓 훼 푉푃퐷 푆푊 푎푔푒 퐶푥 INPUTS

Canopy conductance φ = min {푓푉푃퐷, 푓푆푊}푓푎푔푒 Weather data: temperature, solar radiation, −푘퐿 vapour pressure deficit (VPD), precipitation Gross primary production Pg = 훼퐶(1 − 푒 )푄0

Stand initialisation data: foliage, stem and Net primary production Pn = 0.47 Pg root biomass

Site and stand factors: max. available soil water, initial stem number, initial stand age, maximum stand age BIOMASS PARTITIONING Physiological parameters: canopy quantum 휂푅푥휂푅푛 efficiency, max. canopy conductance, Root partitioning 휼 = 푹 휂 +(휂 −휂 )푚휑 maximum litterfall rate, stomatal response to 푅푛 푅푥 푅푛

VPD, leaf area index (LAI). 1−휂 푅 Stem partitioning ηS = 1−푝퐹푆

Foliage partitioning ηS = 푝퐹푆휂푆

n Stem: foliage ratio 풑푭푺 = aB

STAND CHARACTERISTICS

훾퐹푥훾퐹0 Litter-fall rate 휸푭(풕) = −푘푡 훾퐹0+(훾퐹푥−훾퐹0)푒 STAND MORTALITY

1 훾 3/2 where 푘 = ln (1 + 퐹푥) Average stem weight 풘푺풙(푵) ≤ 푤푆푥0(1000⁄푁) 푡훾퐹 훾퐹0 (푁−푁′) 2 Mortality ∆푾풔 = −훾푆 푊푠 Basal area 푨푩 = 휋푁푆(푑퐵⁄200) 푁

2/3 푛 푛 where N’ = 1000 (wSx0/wS) DBH 푯 = 푎퐻퐵 푁

149

Appendices

Appendix S3.2. List of additional field-measured data

Englhart S., V. Keuck, and F. Siegert. 2011. Aboveground biomass retrieval in tropical forests — The potential of combined X- and L-band SAR data use. Remote Sensing of Environment 115: 1260- 1271.

Jaya A., U. J. Siregar, H. Daryono, and S. Suhartana. 2007. Biomass content of tropical peat swamp forest under various land cover conditions. Jurnal Penelitian Hutan dan Konservasi Alam 4: 12.

Kronseder K., U. Ballhorn, V. Böhm, and F. Siegert. 2012. Aboveground biomass estimation across forest types at different degradation levels in Central Kalimantan using LiDAR data. International Journal of Applied Earth Observation 18: 37-48.

Murdiyarso D., D. Donato, J.B. Kauffman, S. Kurnianto, M. Stidham, and M. Kanninen. 2009. Carbon storage in mangrove and peatland ecosystems: A preliminary account from plots in Indonesia. Center for International Forestry Research, Bogor, Indonesia.

Waldes N., and S. Page. 2002. Forest structure and tree diversity of a peat swamp forest in Central Kalimantan, Indonesia. In: International Symposium on Tropical Peatlands, Jakarta (Indonesia), 22- 23 Aug 2002. BPPT, Jakarta, Indonesia.

150

Appendices

Supplementary Tables

Table A3.1. List of parameter values included in the 3-PG simulation.

Parameter Value Source Foliage – stem partitioning ratio 1 (Clearwater et al., 1999) Stem mass v. diameter relationships 2.6 (Slik et al., 2010) Minimum temperature for growth Site specific (Hijmans et al., 2005) Maximum temperature for growth Site specific (Hijmans et al., 2005) Maximum stand age in age modifier 100 yrs (Brown and Lugo, 1990; Silver et al., 2000; White et al., 2006) Maximum litter fall rate 0.04/month (Paoli and Curran, 2007) Age at median litter fall rate 24 months (White et al., 2006) Maximum canopy conductance 0.019 m/s (Kumagai et al., 2004) Leaf area index for maximum 6.2 (Kumagai et al., 2004) conductance Stomatal response to VPD 0.04/mBar (Clearwater et al., 1999) Self-thinning rule 2 (Slik et al., 2010) Canopy quantum efficiency 0.023-0.043 mol (Eschenbach et al., 1998; Huth and C/mol photons Ditzer, 2000) Specific leaf area at age 0 6 m2/kg (Brearley et al., 2003) Specific leaf area for mature leaves 6 m2/kg (Brearley et al., 2003)

151

Appendices

Table A3.2. Summary of measured and modelled AGB in Mg per hectare after stratification according to forest type. The values in brackets are the standard deviation of the mean. Lowland Montane Heath forest Peat swamp forest forest forest Average measured AGB 477.0 (117.0) 461.9 (157.3) 342.7 (82.2) 348.7 (81.8) Average modelled AGB 484.3 (120.5) 473.6 (150.0) 352.1 (77.5) 348.5 (88.9)

152

Appendices

Supplementary Figures

Figure A3.1. Comparison between the 3-PG modelled (over 100 years) and field-measured data of AGB (Mg/ha) for each forest type: (A) lowland forest; (B) montane forest; (C) heath forest; and (D) peat swamp forest. The straight line indicates 1:1 line.

153

Appendices for Chapter 4

Supplementary Tables

Table A4.1. Summary of implementation costs for each restoration zone. Implementation costs in lowland forest are derived from the Ministry of Forestry regulation and used as the baseline (see Methods). For other forest types, we applied modifiers based on a literature review.

Lowland forest Montane forest Peat swamp forest Freshwater swamp forest Heath forest 1. Intensive square planting 1. Intensive square planting 1. Intensive square planting 1. Intensive square planting 1. Intensive square planting

500 trees per hectare 500 trees per hectare 500 trees per hectare 500 trees per hectare 500 trees per hectare Variables Cost Variables Cost Variables Cost Variables Cost Variables Cost Planning 55.01 Planning 55.01 Planning 55.01 Planning 55.01 Planning 55.01 Planting cost 674.03 Planting cost 792.36 Planting cost 844.73 Planting cost 823.50 Planting cost 1219.27 Maintenance cost (year 1) 188.54 Maintenance cost (year 1) 216.57 Maintenance cost (year 1) 224.21 Maintenance cost (year 1) 207.27 Maintenance cost (year 1) 259.88 Maintenance cost (year 2) 188.54 Maintenance cost (year 2) 216.57 Maintenance cost (year 2) 224.21 Maintenance cost (year 2) 207.27 Maintenance cost (year 2) 259.88

154 Maintenance cost (year 3) 121.60 Maintenance cost (year 3) 121.60 Maintenance cost (year 3) 121.60 Maintenance cost (year 3) 121.60 Maintenance cost (year 3) 121.60

Maintenance cost (year 4) 48.26 Maintenance cost (year 4) 48.26 Maintenance cost (year 4) 48.26 Maintenance cost (year 4) 48.26 Maintenance cost (year 4) 48.26 Total cost (US$/ha) 1,275.98 Total cost (US$/ha) 1,450.37 Total cost (US$/ha) 1,518.03 Total cost (US$/ha) 1,462.83 Total cost (US$/ha) 1,963.91

2. Strip planting 2. Strip planting 2. Strip planting 2. Strip planting 2. Strip planting 200 trees per hectare 200 trees per hectare 200 trees per hectare 200 trees per hectare 200 trees per hectare Variables Cost Variables Cost Variables Cost Variables Cost Variables Cost Planning 55.01 Planning 55.01 Planning 55.01 Planning 55.01 Planning 55.01 Planting cost 334.25 Planting cost 387.19 Planting cost 405.03 Planting cost 396.54 Planting cost 557.33 Maintenance cost (year 1) 93.25 Maintenance cost (year 1) 107.27 Maintenance cost (year 1) 109.39 Maintenance cost (year 1) 102.59 Maintenance cost (year 1) 125.52 Maintenance cost (year 2) 93.25 Maintenance cost (year 2) 107.27 Maintenance cost (year 2) 109.39 Maintenance cost (year 2) 102.59 Maintenance cost (year 2) 125.53 Maintenance cost (year 3) 48.64 Maintenance cost (year 3) 48.64 Maintenance cost (year 3) 48.64 Maintenance cost (year 3) 48.64 Maintenance cost (year 3) 48.64 Maintenance cost (year 4) 19.30 Maintenance cost (year 4) 19.30 Maintenance cost (year 4) 19.30 Maintenance cost (year 4) 19.30 Maintenance cost (year 4) 19.30

Appendices Total cost (US$/ha) 643.72 Total cost (US$/ha) 724.68 Total cost (US$/ha) 746.76 Total cost (US$/ha) 724.68 Total cost (US$/ha) 931.34

Table A4.1. Summary of implementation costs (continued)

Lowland forest Montane forest Peat swamp forest Freshwater swamp forest Heath forest

3. Gap planting 3. Gap planting 3. Gap planting 3. Gap planting 3. Gap planting 100 trees per hectare 100 trees per hectare 100 trees per hectare 100 trees per hectare 100 trees per hectare Variables Cost Variables Cost Variables Cost Variables Cost Variables Cost Planning 55.01 Planning 55.01 Planning 55.01 Planning 55.01 Planning 55.01 Planting cost 215.10 Planting cost 240.79 Planting cost 250.49 Planting cost 246.24 Planting cost 326.64 Maintenance cost (year 1) 56.82 Maintenance cost (year 1) 63.82 Maintenance cost (year 1) 64.89 Maintenance cost (year 1) 61.49 Maintenance cost (year 1) 72.95 Maintenance cost (year 2) 56.82 Maintenance cost (year 2) 63.82 Maintenance cost (year 2) 64.89 Maintenance cost (year 2) 61.49 Maintenance cost (year 2) 72.95 Maintenance cost (year 3) 27.10 Maintenance cost (year 3) 27.10 Maintenance cost (year 3) 27.10 Maintenance cost (year 3) 27.10 Maintenance cost (year 3) 27.10 Maintenance cost (year 4) 9.65 Maintenance cost (year 4) 9.65 Maintenance cost (year 4) 9.65 Maintenance cost (year 4) 9.65 Maintenance cost (year 4) 9.65 Total cost (US$/ha) 420.50 Total cost (US$/ha) 460.20 Total cost (US$/ha) 472.02 Total cost (US$/ha) 460.98 Total cost (US$/ha) 564.31

155

Appendices

Table A4.2. The occurrence of mammal species in prioritised zones under four scenarios and two investment levels (see Methods).

Lower investment Higher investment Scientific name Common name Scenario 1 Scenario 2 Scenario 3 Scenario 4 Scenario 1 Scenario 2 Scenario 3 Scenario 4 Aonyx cinerea Oriental Small-clawed Otter No Yes Yes Yes No Yes Yes Yes Arctictis binturong Binturong Yes Yes Yes Yes Yes Yes Yes Yes Bos javanicus Banteng Yes Yes Yes Yes Yes Yes Yes Yes Catopuma badia Bornean Bay Cat Yes Yes Yes Yes Yes Yes Yes Yes Coelops robinsoni Philippine Tailless Roundleaf Bat Yes Yes Yes Yes Yes Yes Yes Yes Cynogale bennettii Sunda Otter Civet Yes Yes Yes Yes Yes Yes Yes Yes Diplogale hosei Hose's Civet Yes Yes Yes Yes Yes Yes Yes Yes Elephas maximus Bornean Elephant No Yes No No Yes Yes Yes No Haeromys pusillus Lesser Ranee Mouse Yes Yes Yes Yes Yes Yes Yes Yes Helarctos malayanus Sun Bear Yes Yes Yes Yes Yes Yes Yes Yes Hemigalus derbyanus Banded Palm Civet Yes Yes Yes Yes Yes Yes Yes Yes

156 Hylobates muelleri Muller's Bornean Gibbon Yes Yes Yes Yes Yes Yes Yes Yes

Lutra sumatrana Hairy-nosed Otter No Yes Yes Yes No Yes Yes Yes Lutrogale perspicillata Smooth-coated Otter No Yes Yes Yes No Yes Yes Yes Macaca nemestrina Pig-tailed Macaque Yes Yes Yes Yes Yes Yes Yes Yes Manis javanica Pangolin Yes Yes Yes Yes Yes Yes Yes Yes Maxomys rajah Rajah Spiny Rat Yes Yes Yes Yes Yes Yes Yes Yes Maxomys whiteheadi Whitehead's Spiny Rat Yes Yes Yes Yes Yes Yes Yes Yes Megaerops wetmorei Mindanao Fruit Bat Yes Yes Yes Yes Yes Yes Yes Yes Nasalis larvatus Proboscis Monkey Yes Yes Yes Yes Yes Yes Yes Yes Neofelis diardi Clouded Leopard Yes Yes Yes Yes Yes Yes Yes Yes Niviventer cremoriventer Dark-tailed Tree Rat Yes Yes Yes Yes Yes Yes Yes Yes Nycticebus menagensis Slow Loris Yes Yes Yes Yes Yes Yes Yes Yes Pardofelis marmorata Marbled Cat Yes Yes Yes Yes Yes Yes Yes Yes Petinomys genibarbis Whiskered Flying Squirrel Yes Yes Yes Yes Yes Yes Yes Yes Appendices Petinomys setosus Temminck's Flying Squirrel Yes Yes Yes Yes Yes Yes Yes Yes

Table A4.2. The occurrence of mammal species in prioritised zones (continued)

Lower investment Higher investment Scientific name Common name Scenario 1 Scenario 2 Scenario 3 Scenario 4 Scenario 1 Scenario 2 Scenario 3 Scenario 4 Petinomys vordermanni Vordermann's Flying Squirrel Yes Yes Yes Yes Yes Yes Yes Yes Pongo pygmaeus Bornean Orang-utan Yes Yes Yes Yes Yes Yes Yes Yes Presbytis frontata White-fronted Leaf Monkey Yes Yes Yes Yes Yes Yes Yes Yes Presbytis hosei Hose's Leaf Monkey Yes Yes Yes Yes Yes Yes Yes Yes Prionailurus planiceps Flat-headed Cat Yes Yes Yes Yes Yes Yes Yes Yes Rheithrosciurus macrotis Tufted Ground Squirrel Yes Yes Yes Yes Yes Yes Yes Yes Rusa unicolor Sambar Deer Yes Yes Yes Yes Yes Yes Yes Yes Sus barbatus Bearded Pig Yes Yes Yes Yes Yes Yes Yes Yes Tarsius bancanus Western Tarsier Yes Yes Yes Yes Yes Yes Yes Yes

157

Appendices

Table A4.3. Regression analysis of the relationship between forest cover (FC) and above-ground biomass (AGB) for each forest type. We randomly sampled 100 points with a minimum distance of 10 km to extract the values of percent forest cover and AGB from Townshend’s MODIS and Saatchi’s AGB maps respectively (See Methods). Forest type Equation R2 P value Lowland forest FC = 0.1343*AGB + 22.976 0.28 < 0.001 Montane forest FC = 0.1122*AGB + 32.867 0.09 < 0.001 Heath forest FC = 0.1723*AGB + 11.326 0.43 < 0.001 Peat swamp forest FC = 0.1795*AGB + 18.85 0.29 < 0.001 Freshwater swamp forest FC = 0.1795*AGB + 18.85 0.18 < 0.001

158

Appendices

Appendices for Chapter 5

Supplementary Figures

a b 6 10

5 8

4 6 3 4

Extent (Mha) Extent 2 Cost (US$ billion) (US$ Cost

2 1

0 0 C C&F C&M C C&F C&M Scenario Scenario

Figure A5.1. Resources required to offset the impact of oil-palm plantations in Kalimantan assuming that the restoration would fail to prevent peat fires. a, Extent of landscape selected for restoration offsetting. b, Total offsetting cost accounting for opportunity and implementation costs. Only scenarios that included carbon offsets were considered in these analyses: carbon (scenario C); carbon and floristic eco-region (scenario C&F); carbon and mammal’s habitat (scenario C&M). Error bars represent the range of results accounting for lower and higher estimates of total carbon emissions.

159

a b c

Figure A5.2. Priority areas for offsetting the impacts of oil-palm development via restoration in Kalimantan assuming that the restoration would fail to prevent peat fires. Only scenarios that included carbon offsets were considered on these analyses: (a) carbon; (b) carbon and floristic ecoregion; (c)

160 carbon and mammal habitat.

Appendices

Supplementary Tables

Table A5.1. Opportunity cost from alternative forms of land management.

We employed Net Present Value (NPV) of oil-palm plantations as the baseline for opportunity cost. We assumed that highly suitable areas would generate full yield, equating to 100% NPV, while moderately and marginally suitable land would produce 75% and 50% of the full NPV, respectively. Additional revenues could be generated from timber harvesting during land clearing. We assumed that intact forest had a timber value of US$2,268 per hectare (Ruslandi et al., 2011) and logged forest retains 57% of harvestable timber of intact forest (Putz et al., 2012). For burned forest we assumed that only small-scale timber extraction was feasible due to limited commercial timber remaining, in which case a conservative value was assigned. For agroforest, we assigned the average NPV of traditional agroforestry systems.

Alternative land management Opportunity cost Sources (US$ per hectare) Oil-palm plantation (i.e. NPV) (Irawan et al., 2013; Ruslandi Highly suitable 6,355 et al., 2011) Moderately suitable 4,766 Marginally suitable 3,177 Not suitable 0

Timber (Budiharta et al., 2014a; Logged forest 1,292 Ruslandi et al., 2011) Burned forest 300 Non-forest 0

Agroforest 991 (Terauchi et al., 2014)

161

Table A5.2. Biophysical variables employed to determine land suitability for oil palm. Value ranges for each suitability class were from Indonesia’s Ministry of Agriculture (Ministry of Agriculture, 2012b; http://bbsdlp.litbang.pertanian.go.id/kriteria/kelapa%20sawit).

Land suitability class Biophysical variables Spatial data High Moderate Marginal Not suitable Annual temperature (oC) Hijmans et al. (2005) 25-28 22-25 and 28-32 20-22 and 32-35 <20 and >35 Annual rainfall (mm/yr) Hijmans et al. (2005) 1700-2500 1450-1700 and 1250-1450 and 3500- <1250 and >4000 2500-3500 4000 Slope (%) Jarvis et al. (2008) <8 8-16 16-30 >30 Peat thickness/depth (cm) Gingold et al. (2012) <60 60-140 140-200 >200 Soil depth (cm) Gingold et al. (2012) >100 75-100 50-75 <50 Drainage (class) FAO/IIASA/ISRIC/ISSCAS/JRC (2012) Good, moderate Moderate poor Poor, moderate rapid Very poor, rapid Coarse material (% gravel content) ISRIC – World Soil Information (2013) <15 15-35 35-55 >55 CEC-clay (cmol/kg) ISRIC – World Soil Information (2013) >16 <16 - - Base saturation (%) FAO/IIASA/ISRIC/ISSCAS/JRC (2012) <20 <20 - -

162 Acidity (pH) ISRIC – World Soil Information (2013) 5.0-6.5 4.2-5.0 and 6.5 <4.2 and >7.0 -

Carbon organic content (%) FAO/IIASA/ISRIC/ISSCAS/JRC (2012) >0.8 <0.8 - -

Appendices

Table A5.3. Extent of suitable habitat of mammal species impacted by oil-palm plantations in Kalimantan

Kalimantan Total habitat loss by Habitat loss relative to ID Scientific Name Common Name IUCN Status habitat (000 ha) oil palm (000 ha) Kalimantan habitat (%)

CARNIVORA SP01 Aonyx cinereus Asian small-clawed otter Vulnerable 26,550 1,842 6.94 SP02 Arctictis binturong Binturong Vulnerable 17,432 742 4.26 SP03 Arctogalidia trivirgata Small-toothed palm civet Least Concern 31,048 2,575 8.3 SP04 Cynogale bennettii Otter civet Endangered 20,786 2,252 10.83 SP05 Diplogale hosei Hose's civet Vulnerable 1,791 0 0 SP06 Hemigalus derbyanus Banded civet Vulnerable 25,244 1,357 5.38 SP07 Herpestes brachyurus Short-tailed mongoose Least Concern 22,230 1,293 5.82 SP08 Herpestes semitorquatus Collared mongoose Data Deficient 23,163 838 3.62 SP09 Lutra sumatrana Hairy-nosed otter Endangered 46,154 4,313 9.35 SP10 Lutrogale perspicillata Smooth-coated otter Vulnerable 27,534 2,606 9.47 SP11 Martes flavigula Yellow-throated marten Least Concern 21,395 1,467 6.86

163 SP12 Melogale everretti Borneo ferret badger Data Deficient 6 0 0

SP13 Mustela nudipes Malay weasel Least Concern 40,748 3,340 8.2 SP14 Mydaus javanensis Sunda Stink-badger Least Concern 37 0.854 2.26 SP15 Neofelis diardi Sunda clouded leopard Vulnerable 40,558 3,583 8.83 SP16 Paguma larvata Masked palm civet Least Concern 14,343 437 3.05 SP17 Paradoxurus hermaphroditus Common palm civet Least Concern 22,019 1,893 8.6 SP18 Catopuma bardia Bay cat Endangered 23,971 895 3.73 SP19 Pardofelis marmorata Marbled cat Vulnerable 26,503 2,086 7.87 SP20 Prionailurus bengalensis Leopard cat Least Concern 40,841 3,987 9.76 SP21 Prionailurus planiceps Flat-headed cat Endangered 18,542 2,382 12.85 SP22 Prionodon linsang Banded linsang Least Concern 25,322 2,044 8.07 SP23 Viverra tangalunga Malay civet Least Concern 24,143 1,385 5.74

Appendices

Table A5.3. Extent of suitable habitat of mammal species impacted by oil-palm plantations in Kalimantan (continued)

Kalimantan Total habitat loss by Habitat loss relative to ID Scientific Name Common Name IUCN Status habitat (000 ha) oil palm (000 ha) Kalimantan habitat (%)

PRIMATES SP24 Hylobates albibarbis Bornean white-bearded gibbon Endangered 9,971 1,390 13.95 SP25 Hylobates muelleri Muller's bornean gibbon Endangered 18,369 1,461 7.96 SP26 Macaca fascicularis Long-tailed macaque Least Concern 25,185 2,811 11.16 SP27 Macaca nemestrina Pig-tailed macaque Vulnerable 20,692 1,974 9.54 SP28 Nycticebus coucang Slow loris Vulnerable 23,850 2,825 11.85 SP29 Nasalis larvatus Proboscis monkey Endangered 45,246 4,578 10.12 SP30 Presbytis frontata White-fronted langur Vulnerable 11,447 602 5.26 SP31 Presbytis hosei Hose's langur Vulnerable 21,822 1,277 5.85 Critically SP32 Presbytis chrysomelas Bornean banded langur Endangered 7,356 832 11.32 SP33 Pongo pygmaeus Orangutan Endangered 35,559 3,621 10.18

1 SP34 Presbytis rubicunda Maroon langur Least Concern 27,383 2,877 10.51

64 SP35 Tarsius bancanus Western tarsier Vulnerable

12,472 753 6.04 SP36 Trachypithecus cristatus Silvered langur Near Threatened 42,802 4,466 10.43

CHIROPTERA SP37 Aethalops aequalis Bornean pygmy fruit bat Least Concern 1,294 0 0 SP38 Balionycteris maculata Spotted-winged fruit bat Least Concern 18,841 881 4.68 SP39 Chironax melanocephalus Black capped fruit bat Least Concern 29,887 2,143 7.17 SP40 Cynopterus brachyotis Common Short-nosed Fruit Bat Least Concern 21,902 1,334 6.09 SP41 Cynopterus horsfieldi Horsfield's fruit bat, Least Concern 9,688 728 7.52 SP42 Dyacopterus spadiceus Dayak fruit bat Near Threatened 21,334 2,556 11.98 SP43 Eonycteris major Greater nectar bat Data Deficient 29,187 2,214 7.59 SP44 Eonycteris spelaea Lesser nectar bat, Least Concern 25,201 1,825 7.24 SP45 Macroglossus minimus Long-tongued nectar bat Least Concern 18,188 890 4.9

SP46 Megaerops ecaudatus Tail-less fruit bat Least Concern 19,830 652 3.29 Appendices SP47 Megaerops wetmorei White-collared fruit bat Vulnerable 7,293 1,218 16.71

Table A5.3. Extent of suitable habitat of mammal species impacted by oil-palm plantations in Kalimantan (continued)

Kalimantan Total habitat loss by Habitat loss relative to ID Scientific Name Common Name IUCN Status habitat (000 ha) oil palm (000 ha) Kalimantan habitat (%)

CHIROPTERA SP48 Penthetor lucasi Lucas's Short-nosed fruit bat Least Concern 17,686 704 3.99 SP49 Pteropus vampyrus Large flying-fox Near Threatened 29,442 3,497 11.88 SP50 Rousettus amplexicaudatus Geoffroy's rousette Least Concern 13,859 1,522 10.98 SP51 Rousettus spinalatus Bare-backed rousette Vulnerable 14,374 830 5.78 SP52 Rhinolophus acuminatus Acuminate horseshoe bat Least Concern 32,250 2,990 9.27 SP53 Rhinolophus affinis Intermediate horseshoe bat Least Concern 29,024 1,934 6.67 SP54 Rhinolophus borneensis Bornean horseshoe bat Least Concern 17,920 644 3.59 SP55 Rhinolophus creaghi Creagh's horseshoe bat Least Concern 13,395 1,028 7.68 SP56 Rhinolophus luctus Great woolly horseshoe bat Least Concern 22,242 905 4.07 SP57 Rhinolophus philippinensis Philippine horseshoe bat Least Concern 48,206 4,218 8.75 SP58 Rhinolophus sedulus Lesser woolly horseshoe bat Near Threatened 20,045 751 3.75

165 SP59 Rhinolophus trifoliatus Trefoil horseshoe bat Least Concern 16,159 821 5.08

SP60 Hipposideros ater Dusky roundleaf bat Least Concern 3,860 108 2.82 SP61 Hipposideros bicolor Bicolored roundleaf bat Least Concern 17,946 1,508 8.41 SP62 Hipposideros cervinus Fawn-colored roundleaf bat Least Concern 22,787 915 4.02 SP63 Hipposideros cineraceus Least roundleaf bat Least Concern 22,920 2,013 8.79 SP64 Hipposideros diadema Diadem roundleaf bat Least Concern 11,341 613 5.41 SP65 Hipposideros doriae Borneo roundleaf bat Near Threatened 46,805 4,412 9.43 SP66 Hipposideros dyacorum Dayak roundleaf bat Least Concern 8,650 323 3.74 SP67 Hipposideros galeritus Cantor's roundleaf bat Least Concern 20,900 1,092 5.23 SP68 Hipposideros larvatus Intermediate roundleaf bat Least Concern 35,492 3,860 10.88 SP69 Hipposideros ridleyi Ridley's roundleaf bat Vulnerable 37,701 3,890 10.32 SP70 Megaderma spasma Lesser false vampire Least Concern 36,802 3,288 8.94 SP71 Nycteris tragata Hollow-faced bat Near Threatened 32,800 2,476 7.55

Appendices SP72 Murina aenea Bronzed tube-nosed bat Vulnerable 17,123 996 5.82

Table A5.3. Extent of suitable habitat of mammal species impacted by oil-palm plantations in Kalimantan (continued)

Kalimantan Total habitat loss by Habitat loss relative to ID Scientific Name Common Name IUCN Status habitat (000 ha) oil palm (000 ha) Kalimantan habitat (%)

CHIROPTERA SP73 Murina cyclotis Round-eared tube-nosed bat Least Concern 18,914 723 3.82 SP74 Murina rozendaali Gilded tube-nosed bat Vulnerable 33,622 3,707 11.03 SP75 Murina suilla Lesser tube-nosed bat Least Concern 5,471 422 7.72 SP76 Kerivoula hardwickii Hardwicke’s woolly bat Least Concern 13,190 893 6.78 SP77 Kerivoula intermedia Small woolly bat Near Threatened 15,834 921 5.82 SP78 Kerivoula minuta Least woolly bat Near Threatened 20,652 2,199 10.65 SP79 Kerivoula papillosa Papillose woolly bat Least Concern 13,666 801 5.87 SP80 Kerivoula pellucida Clear-winged woolly bat Near Threatened 14,852 1,077 7.25 SP81 Phoniscus atrox Gilded groove toothed bat Near Threatened 28,350 2,636 9.3

166

Appendices

Appendices for Chapter 6

Supplementary Tables

Table A6.1. Ontology of social-ecological characteristics of indigenous tree-based agroforestry system (i.e. simpukng and kebotn) of Dayak Paser (and the related Dayak Benuaq) communities in East Kalimantan. A social-ecological systems diagnostic approach (Ostrom, 2009) is used to unpack the system into subsystems and variables.

Resource System (RS) Subsystem Description Reference a. Sector (RS1) RS1-a. Fruit gardens (simpukng/awa pangeramu/lembo/lepu’un) planted with edible fruit trees including (Gönner, 2000; Crevello, durian, mango, jackfruit, rambutan (Nephelium spp.) and langsat (Lansium spp.). 2003; Joshi et al., 2004; RS1-b. Perennial crop agroforests (kebotn/awa pangekulo) enriched with commercial crops of perennial tree Saragih, 2011) species including coffee, cocoa and jungle rubber. b. Clarity of system The traditional boundary of agroforestry system is unclear and rarely recognised formally by the state system. (Gönner, 2000; Nanang and boundaries (RS2) Although not clearly marked, both resource units are informally recognised indicating individual Inoue, 2000; Sardjono et al., management. 2013)

167 c. Size of resource system The average combined area of fruit gardens (simpukng) and perennial crop agroforest (kebotn) per household (Gönner, 2000; Gönner and

(RS3) is four hectares. Seeland, 2002; Saragih, 2011) d. Human constructed Very limited human-constructed facilities support the agroforestry system (e.g. the absence of irrigation (Crevello, 2003; Gönner, facilities (RS4) systems and processing factories). Commercial products (e.g. dried rubber latex) are prepared locally and 2000; Saragih, 2011) sold to traders who regularly visit the villages or the nearby local markets. e. Productivity of system Productivity is low compared with the alternative, more intensive land use management. (Gönner and Seeland, 2002; (RS5) RS5-a. Mean annual productivity from 1.25 hectares of simpukng is 200 kg of rambutan, 357 kg of langsat, Crevello, 2003) 135 kg of mango, 204 of durian and 557 of jackfruit. RS5-b. Mean annual productivity from 2.62 hectares of kebotn is 800 kg of rubber and 1,966 kg of rattan. f. Equilibrium properties Simpukng and kebotn require low input of capital and labour. (Gönner, 2000, 2010; Belcher, (RS6) Simpukng and kebotn have been developed to provide a variety of forest goods and to respond ecological 2004; Saragih, 2011) (e.g. prolonged drought) and socio-economic uncertainties (e.g. change in commodity prices). The agroforestry system is perceived as a sovereign way of life, which is part of local culture and identity. g. Predictability of system Simpukng and kebotn are highly predictable due to the ease of establishment through enrichment planting, (Gönner, 2000; Joshi et al., dynamics (RS7) simple maintenance and high resilience to ecological disturbances (e.g. drought). 2004; Pambudhi et al., 2004) h. Storage characteristics RS8-a. Fruits in the simpukng are generally harvested seasonally following mass fruiting event. (Crevello, 2003; Gönner,

(RS8) RS8-b. Perennial crops in the kebotn are either harvested regularly throughout the year (e.g. rubber latex, 2000) Appendices coconut) or harvested seasonally (e.g. coffee). i. Location (RS9) Paser District, East Kalimantan Province, Indonesia.

Resource Units (RU) Subsystem Description Reference a. Resource unit mobility All units (i.e. various fruit and perennial crop tree species) are immobile (however harvested products are (Saragih, 2011) (RU1) generally transportable). b. Growth or replacement rate Growth of tree species planted on simpukng and kebotn is influenced by land suitability and has strong (Gönner, 2000; Joshi et al., (RU2) relationships with biophysical factors such as soil, climate and topography. 2004; García-Fernández and Simpukng and kebotn are established mainly through enrichment planting on degraded forests or secondary Casado, 2005; Saragih, 2011) forests, which retain some extent of native vegetation cover. These land cover types are suitable for growing both shade-tolerant (e.g. coffee and cocoa) and intolerant (e.g. coconut and durian) species. c. Interaction among resource Simpukng and kebotn provide suitable habitat for wildlife and could provide a conservation benefit. However, (Gönner, 2000) units (RU3) colonisation and movement through agroforests by wildlife could also facilitate killing and hunting for bush meat, either cancelling or negating this conservation benefit. d. Economic value (RU4) Economic value is highly influenced by external factors (e.g. market and policy), and this results in price (Crevello, 2003; Meijaard et fluctuations and lower predictability of economic returns for local farmers. al., 2014; Terauchi et al., RU4-a. Annual income generated from simpukng is IDR 6.34 million (US$526). 2014) RU4-b. Annual income generated from rubber and rattan in kebotn (assuming current farm gate price) is IDR 3.2 million (US$267) and IDR 2.99 million (US$250) per household respectively. In better market conditions, income per household from rubber and rattan could reach US$2,000 and US$1,376 respectively. 168 e. Number of units (RU5) RU5-a. One hectare of simpukng contains up to 704 fruit trees (DBH>10cm) belonging to 93 species. (Gönner, 2000; García-

RU5-b. One hectare of kebotn of rubber contains an average of 400 rubber trees and 118 other trees. One Fernández and Casado, 2005) hectare of kebotn of rattan contains 1,750 canes intertwined with 515 trees of various species. f. Distinctive markings (RU6) RU6-a. Simpukng is highly similar to the surrounding old-growth forest due to high floristic diversity (i.e. (Lawrence, 1996; Gönner, abundance and richness with more than 90 species of tree per hectare) and complex structure (i.e. canopy 2000; García-Fernández and layering). Casado, 2005) RU6-b. Kebotn of rubber is easily recognised due to the dominance of rubber trees (61.1% of total basal area). Kebotn of rattan appears similar to old growth forest with slightly lower tree species richness (17.6 species/0.1 ha compared with 22 species/0.1 ha) and canopy cover (66.9% compared with 83.9%). g. Spatial and temporal Both simpukng and kebotn mainly occur near to transportation networks (e.g. roads, rivers). (Gönner, 2000, 2010; Saragih, distribution (RU7) In many cases, land management may be dynamically shifted between the two types, depending on individual 2011) farmers’ interests and market conditions, for example by intensifying rubber planting in simpukng to become kebotn, and vice versa.

Appendices

Governance System (GS) Subsystem Description Reference a. Government organisations Since the New Order Era (1967), the state, through the Ministry of Forestry, has had a full power to control (President of the Republic of (GS1) the management of the Forest Estate (i.e. areas designated as permanent forest, whether actually forested Indonesia, 1999; Nanang and or not) and its allocation to formal land use classes (e.g. logging concessions and protected areas). Inoue, 2000; Sardjono et al., Since the Reformation Era (1998), some authority relating to forestry (e.g. supervision of reforestation 2013) programs) has been delegated to provincial and district government through Forestry Services. These government institutions aim to reduce bureaucratic complexity and increase the effectiveness of government programs. b. Non-government Local Dayak people have developed their customary rules regarding forest resources management (e.g. land (Gönner, 2000; Crevello, organisations (GS2) ownership, land use classification and viable activities, labour sharing, sanction systems and conflict 2003; Sardjono et al., 2013) resolution). The governance systems are applied informally within the group. c. Network structure (GS3) Agroforestry systems involve strong horizontal networks within the group (usually at village level) but very (Nanang and Inoue, 2000; weak vertical networks connecting resource users with the government (especially at national level). Moeliono et al., 2009) Decentralisation through delegation of power to local government has reduced the distance and complexity of network structures. This, however, has had varying impact, with some areas achieving more direct connections to official decision-makers, while in other areas it has led to new forms of patronage and political favouritism. d. Property-rights systems GS4-a. Formally, agroforestry lands are owned by the state when occurring inside the Forest Estate, while (President of the Republic of 169 (GS4) those outside the Forest Estate are privately owned. Indonesia, 1999; Gönner,

GS-4b. Informally, an effective property rights system on agroforestry land has been applied for centuries, 2000; Nanang and Inoue, which is divided into two categories of ownership: private and community, depending on customary 2000; Saragih, 2011) consensus. e. Operational rules (GS5) Daily decisions in managing privately owned agroforest are carried out by individual farmers. (Crevello, 2003; Saragih, 2011) f. Collective-choice rules Decisions related to communal interests are made at village meetings, led by a traditional leader. (Gönner, 2000) (GS6) g. Constitutional rules (GS7) Through Basic Forestry Law No. 41/1999 and its generated regulations (e.g. Ministry of Forestry Decree No. (Sardjono et al., 2013) P. 37/2007 jo. P. 18/2009 jo. P. 13/ 2010 jo P. 52/2011), local communities could be given rights to manage the Forest Estate through community forest (Hutan Kemasyarakatan/HKm). Community forest involving tree planting can only be designated in parts of Production Forest on condition that no logging concession is active. h. Monitoring and sanctioning Sanctioning processes are conducted in group meetings (berinok) at village level if rule breaking only involves (Gönner, 2000; Saragih, 2011; processes (GS8) internal group members. Sardjono et al., 2013) Sanctioning processes are taken to formal institutions (e.g. administrative office) if rule breaking involves any

external parties (e.g. government, logging concession holders or oil-palm plantations). Appendices

Users (U) Subsystem Description Reference a. Number of users (U1) On average, the number of households in agroforestry villages is 185 (with total population of 749 persons). (Paser Statictics Service, 2014b) b. Socio-economic attributes Livelihoods are mainly from the agroforestry sector with the combination of subsistence and monetary (Dewi et al., 2005; Gönner, of users (U2) economy. 2000; Saragih, 2011) Educational level is generally low with six years of primary school plus three years of secondary school. c. History of use (U3) Traditional Dayak groups have been practising agroforestry systems for more than 300 years. (Gönner, 2000, 2010) d. Location (U4) Village settlements are close to forested areas (at the average distance of 2 km) and have limited access to (Saragih, 2011; Sihombing, cities. 2011) e. Leadership/ Historically, the groups have strong leadership from a traditional leader (kepala adat) but this is weakening (Gönner, 2000) entrepreneurship (U5) due to increasing authority held by officials who are appointed by the government (kepala desa). Dayak communities have generally low levels of engagement in entrepreneurship outside of traditional livelihoods, and new economic opportunities (e.g. trading sector, oil-palm plantation) are mostly dominated by outsiders. f. Norms/ social capital (U6) Strong social cohesiveness through mutual assistance and labour sharing (gotong-royong) among group (Gönner, 2000) members especially for agricultural activities. g. Knowledge of SES/ mental The elderly generally has better knowledge on technical aspects of traditional agroforestry system than the (Belcher et al., 2004; Gönner, 170 model (U7) younger generation. On the other hand, young members have better access to external information such as 2000; Saragih, 2011)

resource prices and market development. They are less attracted to swidden agriculture but remain interested in rubber and rattan cultivation. There is varied preference among forest frontier communities for continuing agroforestry practices or altering into alternative livelihood options (e.g. oil-palm plantation). h. Dependence on Forest-frontier people are highly dependent on agroforestry for both economic and cultural benefits: (Abram et al., 2014; Dewi et resource (U8) U8-a. Agroforestry sector contributes to 42.2-62.4% (mean 51.5%) of total household income. al., 2005; Saragih, 2011) U8-b. High users of agroforestry products to support their culture and way of life. Agroforestry systems in rural Paser provide between 47 and 88 plant species for food and vegetables, 38 to 56 species for medicinal purposes as well as 15 to 23 species for traditional ceremonies. i. Technology used (U9) Very limited uses of machinery and production materials (e.g. agrochemicals and fertilizers). (Gönner, 2000)

Appendices

Interactions (I) Subsystem Description Reference a. Harvesting levels of Harvesting level is decided at household level by individual farmers. (Crevello, 2003) diverse users (I1) b. Information sharing Information sharing among group members is frequent and in-depth, occurring primarily through meetings, (Gönner, 2000) among users (I2) social events and religious ceremonies. c. Deliberation processes (I3) Village meetings usually reach full consensus through long discussions. (Gönner, 2000) d. Conflict among users (I4) Conflict among group members is low. (Crevello, 2003; Gönner, Conflict with outside institutions is increasing especially with industrial timber plantation and oil-palm 2000; Nanang and Inoue, companies, while conflict with logging concessions is moderate as logging companies provide financial 2000; Sardjono et al., 2013) compensation and generally allow indigenous people to access concession forests for hunting and gathering. e. Investment activities (I5) Investment is low due to limited capital and limited access to credit facilities. (Saragih, 2011) f. Lobbying activities (I6) Lobbying activities are limited. g. Self-organising activities Self-organisation goes well for activities involving internal group members (e.g. determining land ownership (Gönner, 2000; Sardjono et (I7) and labour sharing). On the other hand, self-organisation involving external institutions is often al., 2013) challenging due to the mismatch between governance systems (i.e. state legal system vs traditional

171 customary rules). h. Networking activities (I8) Internal networking activities among group members are intense while networking with external institutions (Crevello, 2003; Gönner, (e.g. government, traders) is lacking due to limited access to transportation and information. 2000)

Outcomes (O) Subsystem Description Reference a. Social performance Simpukng and kebotn improve land productivity (compared with degraded forest or exhausted fields) and (Belcher et al., 2004; Gönner, measures (O1) provide a variety of goods for forest-dependent people 2010; Joshi et al., 2004) Many forest-dependent people still prefer agroforestry systems for their livelihood and want to expand their simpukng and kebotn through national reforestation program. Economic value of simpukng and kebotn is low compared with alternative, intensive forms of land use management (e.g. oil-palm plantation, mineral extraction and industrial timber plantation). b. Ecological performance Simpukng and kebotn can maintain forest cover with high biodiversity value and carbon stock. (García-Fernández and measures (O2) Simpukng and kebotn are highly resilient to natural disturbances (e.g. forest fire, drought). For example, Casado, 2005; Gönner, 2000) despite the severe impact of forest fires in 1997, many rubber trees recovered quickly and were tapped in the subsequent weeks.

Appendices c. Externalities to other Increasing competition with other land uses (e.g. development of oil palm and industrial timber plantations, (Belcher et al., 2004; Nanang SESs (O3) which may be promoted or involve coercion from the state or companies) has reduced the availability of and Inoue, 2000; Sardjono et land for simpukng and kebotn and resulted in social conflicts between villagers and external entities. al., 2013)

Table A6.2. List of tree species of interest for simpukng and kebotn considered in this study.

Latin name Local name Common Uses name

Artocarpus Nangka Jackfruit Edible fruit and seed, vegetable, heterophyllus medicinal, timber Artocarpus integer Cempedak - Edible fruit, vegetable, timber Cocos nucifera Kelapa Coconut Edible fruit; condiment, timber Coffea canephora Kopi Coffee Beverage Dimocarpus longan Mata kucing Longan Edible fruit Durio zibethinus Durian Durian Edible fruit, timber Hevea brasiliensis Karet Rubber Latex, timber Lansium domesticum Langsat - Edible fruit, medicinal Nephelium lappaceum Rambutan Rambutan Edible fruit, medicinal

Parkia speciosa Petai Stinky bean Vegetable, condiment, timber Theobroma cacao Kakao Cocoa Condiment

172

Table A6.3. Example of biophysical requirements to determine land suitability classes for rubber (Hevea brasiliensis). Value ranges for each class are from Ritung et al. (2007). Biophysical requirements for other 10 species are available on the website of Indonesian Ministry of Agriculture (Ministry of Agriculture, 2012) and can be downloaded through the link below: http://bbsdlp.litbang.pertanian.go.id/tamp_komoditas.php

Land suitability class Biophysical variables High Moderate Marginal Not suitable

Annual temperature (oC) 26-30 24-26 and 30-34 22-24 <22 and >34

Annual rainfall (mm/yr) 2500-3000 2000-2500 and 1500-2000 and <1500 and >4000 3000-3500 3500-4000

Slope (%) <8 8-16 16-30 >30

Drainage (class) Well Moderate Moderate poor, Very poor, rapid poor

Texture (USDA texture Fine, slightly - Slightly coarse Coarse class) fine, medium

Coarse material (% gravel <15 15-35 35-60 >60 content)

Soil depth (cm) >100 75-100 50-75 <50

CEC-clay (cmol/kg) - - - -

Base saturation (%) <35 35-50 >50 -

Acidity (pH) 5.0-6.0 4.5-5.0 and 6.0-6.5 >6.5 -

Carbon organic content (%) >0.8 <0.8 - -

173