Tropical landscapes in transition?

Tropical landscapes in transition? Widespread land-use change and measures to maintain forests, carbon stocks and biodiversity in North and East Kalimantan, Indonesia

Tropical landscapes in transition? Widespread land-use change and measures to maintain forests, carbon stocks and biodiversity in North and East Kalimantan, Indonesia Carina van der Laan, Utrecht University, Faculty of Geosciences, Department of Innovation, Environmental and Energy Sciences, Copernicus Institute of Sustainable Development, Group Energy & Resources. ISBN: 9789086720729 Cartography and figures: Ton Markus Copyright © 2016 Carina van der Laan Niets uit deze uitgave mag worden vermenigvuldigd en/of openbaar gemaakt door middel van druk, fotokopie of op welke andere wijze dan ook zonder voorafgaande schriftelijke toestemming van de uitgevers. All rights reserved. No part of this publication may be reproduced in any form, by print or photo print, microfilm or any other means, without written permission by the publishers. Printed on recycled paper by Proefschriftmaken/BOXpress.

Tropical landscapes in transition? Widespread land-use change and measures to maintain forests, carbon stocks and biodiversity in North and East Kalimantan, Indonesia

Tropische landschappen in transitie? Wijdverspreide landgebruiksverandering en maatregelen voor het behoud van bossen, koolstofvoorraden en biodiversiteit in Noord- en Oost-Kalimantan, Indonesië (met een samenvatting in het Nederlands)

Bentang alam tropis dalam masa transisi? Meningkatnya perubahan penggunaan lahan dan upaya-upaya menjaga hutan, cadangan karbon dan keanekaragaman hayati di Kalimantan Utara dan Timur, Indonesia (dengan ringkasan dalam bahasa Indonesia)

Proefschrift ter verkrijging van de graad van doctor aan de Universiteit Utrecht op gezag van de rector magnificus, prof. dr. G.J. van der Zwaan, ingevolge het besluit van het college voor promoties in het openbaar te verdedigen op vrijdag 23 september 2016 des middags te 2.30 uur

door Carina van der Laan geboren op 3 juni 1981 te Amsterdam

Promotor:

Prof.dr. A.P.C. Faaij

Copromotoren:

Dr. P.A. Verweij Dr. S.C. Dekker

Dit proefschrift werd (mede) mogelijk gemaakt met financiële steun van NWO-WOTRO (onderdeel van de Nederlandse Organisatie voor Wetenschappelijk Onderzoek) en de KNAW (de Koninklijke Nederlandse Akademie van Wetenschappen), binnen het Agriculture beyond Food onderzoeksprogramma.

Table of Contents Abbreviations 7 1 Introduction 1.1 Land development and forest loss in the tropics 1.2 The Forest Transition Theory 1.3 Land development and forest cover loss in Indonesia 1.4 Strong growth of oil palm and other cash-crop cultivation in Indonesia 1.5 Problem definition 1.6 Research questions and outline of the dissertation

11 11 12 13 14 15 17

2 2.1 2.2

21 21 23

Case study area North and East Kalimantan provinces, Indonesian Borneo West Kutai and Mahakam Ulu districts, East Kalimantan

3 Land use and land cover change trajectories in a tropical forest landscape 3.1 Introduction 3.2 Material and methods 3.4 Results 3.5 Discussion 3.6 Conclusions 4

Impacts of land development and land zoning policies on land use and forest cover projected for 2030 4.1 Introduction 4.2 Theory 4.3 Material and methods 4.4 Results 4.5 Discussion and conclusions 5

Analysis of biophysical and anthropogenic variables and their relation to the regional spatial variation of aboveground biomass 5.1 Introduction 5.2 Methods 5.3 Results 5.4 Discussion 5.5 Conclusions

27 28 29 31 39 41 43 44 46 48 52 59 65 66 68 73 78 79

5

6

Impacts of historical and projected land use change on carbon stocks and plant biodiversity in East Kalimantan, Indonesia 6.1 Introduction 6.2 Materials and methods 6.3 Results 6.4 Discussion and conclusions 7 Mitigation of unwanted direct and indirect land use change 7.1 Introduction 7.2 Materials and methods 7.3 Assessment of the four mitigation measures under the low, medium and high scenarios 7.4 Results 7.5 Discussion

81 82 83 89 98 105 106 107 111 119 121

8 Synthesis and discussion 8.1 Background 8.2 In what stage of the forest transition (curve) is the study region, what is the expected trend and what are the LULC change processes and drivers? 8.3 Can forest cover in the study region be stabilised or even increase and how? 8.4 LULC development in the nearby future 8.5 How to monitor LULC change and forest cover?

126 132 138 139

9 Appendices

141

10 References

197

English Summary

215

Ikhtisar Bahasa Indonesia

221

Nederlandse samenvatting

227

Acknowledgements / Dankwoord

233

Curriculum Vitae

237

6

125 125

Abbreviations AGB Aboveground biomass BMP Best/better management practices CO2 Carbon dioxide ha Hectares FFB Fresh Fruit Bunches HL Hutan Lindung (Watershed protection forest zone) HP Hutan Produksi (Production forest zone) HPK Hutan Produksi yang dapat di-Konversi (Production forest for Conversion zone) HPT Hutan Produksi Terbatas (Limited production forest zone) HSD Honest Significant Difference KBNK Kawasan Budidaya Non Kehutanan (Non-forest area, formerly called Areal Penggunaan Lain: APL) KSPA Kws. Suaka Alam dan Pelestarian Alam (Conservation forest zone) LR Limited-restricted (scenario) LuR Limited-unrestricted (scenario) LULC Land use and land cover MEMR Ministry of Energy and Mineral Resources MoF Ministry of Forestry Mt Million tonne(s) OER Oil Extraction Rate PKO Palm Kernel Oil OPT Oil Palm Trunk RCA Responsible Cultivation Areas t tonnes TGHK Tata Guna Hutan Kesepakatan uLR unlimited-restricted (scenario) uLuR unlimited-unrestricted (scenario) yr Year

7

8

“How can you buy or sell the sky, the warmth of the land? The idea is strange to us. If we do not own the freshness of the air and the sparkle of the water, how can you buy them? … Whatever man does to the web [of life], (s)he does to himself ”. Quote Chief Seattle, 1854

9

10

1 Introduction

1.1

Land development and forest loss in the tropics

Due to the increasing world population and overall living standard, there is a growing demand for food, feed, fibre and fuel. The increase in production of these commodities needs to be accompanied by innovations to increase yields (Gutiérrez-Vélez et al., 2011; Phalan et al., 2013), by smart use of fertilisers and improved pest management (Eshel et al., 2014), and by improved crop varieties and diversification (Massawe et al., 2016). In spite of potential innovations, it will also lead inevitably to an extension of agricultural land by land use and land cover (LULC) change. LULC change occurs particularly in tropical rural regions (Eisner et al., 2016; Gibbs et al., 2010). In these regions, climate and extensive land areas, low population densities and lack of clear land tenure systems seem to provide optimal potential for rapid land development. On the one hand, land development provides income to certain groups in society, ranging from multinational companies to local companies and smallholders, and government actors. On the other hand, land development results in the decline of forest cover, species richness, carbon stocks and local food production, and can lead to land use conflicts (Abood et al., 2014; Fargione et al., 2008; Gerber, 2011; Hooijer et al., 2010; Immerzeel et al., 2014; Koh and Wilcove, 2008; Laurance et al., 2014; Sodhi et al., 2010). These so-called unwanted LULC changes occur in many regions in the tropics (Benhin, 2006; Gibbs et al., 2010), particularly in the most forested regions, namely the Amazon region in Brazil, the Congo basin and the lowlands of Indonesia (Abood et al., 2014; Broich et al., 2011a; Carlson et al., 2012b; Ernst et al., 2013; Gaveau et al., 2014, 2013; Hansen et al., 2013; Mosnier et al., 2014; Phalan et al., 2013). This dissertation focuses on past LULC change processes and LULC change trajectories that occurred in a landscape in East Kalimantan that is under high land development pressure, namely West Kutai and Mahakam Ulu districts. Additionally, it identifies the main contributing land uses to LULC change and forest loss and applies these in a land use change model to project future LULC change. Next to this, in this dissertation, we show the projected LULC changes and impacts on aboveground carbon stocks and plant biodiversity. Finally, we assess four measures to minimise unwanted LULC change. In this introductory chapter and in the synthesis in Chapter 8, we place our findings in the context of the Forest Transition Theory.

11

1.2

The Forest Transition Theory

According to the Forest Transition Theory, a country or region first goes through (1) a stage of high forest cover and low deforestation rates, then (2) a stage with high forest cover and high deforestation rates, subsequently (3) a stage with low forest cover and high deforestation rates, (4) a stage in which forest cover stabilises and deforestation decreases, after which (5) the low forest cover may increase through reforestation processes (Figure 1.1) (Angelsen, 2009; Mather, 1992; Rudel et al., 2005). It is hypothesized that the underlying causes of the forest transition include economic development, industrialisation and urbanisation across all stages, including increasing agricultural land use in the first stages (Meyfroidt et al., 2010) and increasing agricultural adjustment0F in the later stages (Angelsen and Rudel, 2013; Mather and Needle, 1998; Mather, 1992; Rudel et al., 2010). As a result, many tropical forested regions seem to be ‘moving’ through a forest transition (Figure 1.1) (Meyfroidt et al., 2010), resulting in forest-agricultural mosaics (Angelsen and Rudel, 2013).

Forest cover

Congo Basin Brazil, Peru, Indonesia, Cameroon

Stage 1 Stage 2

Stage 3

Chile, India Stage 4

Costa Rica, Vietnam

Stage 5

Time Figure 1.1. Countries characterised according to their position along the forest transition curve (adapted from Angelsen, 2009). The estimated stages of the Congo Basin (stage 1; (Mayaux et al., 2013)), Brazil, Indonesia, Peru and Cameroon (stage 2; (Meyfroidt et al., 2010)), Chile, India (stage 4; Meyfroidt et al., 2010)) and Mato Grosso state, Brazil (stage 4; (Macedo et al., 2012)) and Costa Rica and Vietnam (stage 5; (Meyfroidt et al., 2010)), according to the Forest Transition Theory. The x-axis can be replaced by macro-economic variables (ICRAF, 2014).

The lowlands in Indonesia and the Congo basin are claimed to be in the first stages of a forest transition (Angelsen, 2007; Barbier et al., 2010; Casson et al., 2014; Margono et al., 2014; Mayaux et al., 2013; Mosnier et al., 2014; Murdiyarso et al., 2008; Rudel et al., 2009), meaning high forest cover and processes of deforestation, while the Amazon basin in Brazil seems to be moving towards decreasing deforestation rates and stabilising forest cover (Macedo et al., 2012) (Figure 1.1). In the lowlands of Indonesia, mining, logging and large-scale monocultures for palm oil, pulpwood and timber production are the land uses identified to contribute mostly to deforestation (Abood et al., 2014). In the Congo Basin, expansion of agro-industrial plantations and small-scale subsistence agriculture (Ernst et al., 2013; Tegegne et al., 2016) are the main land uses identified. In the Brazilian Amazon, pasture and agriculture have been the main land uses contributing to forest cover loss, in combination with infrastructure development (Verweij et al., 2009). On the other hand, several countries have managed to increase their forest cover and agricultural production at the same time, however, often

12

hand in hand with displacement of land uses elsewhere (Lambin and Meyfroidt, 2011). For example, forest cover in Costa Rica and Vietnam is increasing and the two countries are moving through the forest transition curve (Figure 1.1). Over the past years, however, these two countries have made a transition from net absorbers of land use (meaning high production and export of agricultural and forest products) to net displacers of land use (meaning import of agricultural and forest products) (Meyfroidt et al., 2010). Such displacement can be caused by, for example, the stagnation of production in the countries or the increase in the national demand for certain commodities due to economic growth (Meyfroidt et al., 2010). Thus, moving through the forest transition curve towards reforestation does not necessarily mean that consumption of agricultural and forest products in a country is not related to deforestation (Lambin and Meyfroidt, 2011). All tropical regions have their own specific and complex underlying socio-economic and legal causes and drivers of forest loss. While the underlying causes are widespread, the conversion to agriculture is the dominant process across the tropics (Benhin, 2006; Gibbs et al., 2010; Laurance et al., 2014). Angelsen & Rudel (2013) identified 5 categories of drivers of increasing forest cover or its stabilisation: 1) Scarcity of forest products due to shrinking forest stock and rising demand of forest products, 2) Scarcity of forest environmental services, 3) Diminishing agricultural rent from continuing forest conversion, 4) Economic development and structural changes, and 5) Policy changes. For example, institutional and policy factors have shown to be important factors in minimising deforestation and forest degradation in Cameroon and the Republic of Congo in the long term (Tegegne et al., 2016). Macedo et al. (2012) and Tegegne et al. (2016) also indicate that deforestation should be restricted by policies and that policies need to be in place that discourage the expansion into forested lands and promote and plan the use of under-utilised low-carbon stock lands. Meanwhile, the Forest Transition Theory is under scrutiny in the literature. For example, it is uncertain whether certain tropical regions or the entire tropics will reach the regeneration stage (Perz, 2007). Additionally, more focus should be on the prevention of displacement of land uses to more distant regions (Meyfroidt et al., 2010). Focus is mostly on forest cover and not on forest quality, e.g. pristine versus secondary forest and natural forest versus plantation forest, while forest degradation processes can still play a major role in LULC change and plantation forest is generally not considered as forest (Ankersen et al., 2015). Additionally, other important local factors, such as the maintenance of local communities’ lands and an acceptable level of local food production are often not accounted for. Finally, scientific articles claim that the exclusive focus on forests might undermine the importance of non-forest habitats (Melo et al., 2013). The largest challenge regarding unwanted LULC change and forest cover maintenance in the context of the Forest Transition Theory, is to find ways to stabilise a high forest cover. This means that the transitions from stages 1-2 to stages 3-4 need to be avoided and instead the transition should preferably move directly from stage 1-2 to stage 5.

1.3

Land development and forest cover loss in Indonesia

In the past years, Indonesia has reached the highest deforestation rates in the world (Margono et al., 2014) and there are no signs that deforestation will slow down in the near future. Even since the extended nationwide moratorium of 2011 on the issuance of new licences of oil palm, pulpwood and selective logging concessions on primary forests and peatlands (Presidential Instruction 10/2011; 6/2013; 8/2015) this has not been the case (Margono et al., 2014). Considering the ongoing land development and deforestation rates in particularly Sumatra (Broich et al., 2011a), Kalimantan

13

(Margono et al., 2014) and Papua (Margono et al., 2014), the country seems to be moving through the second stage of the forest transition, with large regional differences between districts and provinces (Angelsen, 2007; Barbier et al., 2010; Margono et al., 2014; Murdiyarso et al., 2008; Rudel et al., 2009). These rapid land developments are accompanied by large-scale fires that have affected forests, plant and animal species and the health and livelihoods of people, particularly in Kalimantan and Sumatra (Marlier et al., 2015). Multiple underlying causes of deforestation have been identified, such as road development, corruption, illegal logging and agricultural development (Benhin, 2006; Brun et al., 2015; Gatto et al., 2015). Additionally, the government’s transmigration program initiated by the government from the late 1970s until the early 1990s and the spontaneous migration that resulted from it has resulted in the movement of thousands of people from the inner to the outer islands of Indonesia (Benhin, 2006).

1.4

Strong growth of oil palm and other cash-crop cultivation in Indonesia

Since the early 1980s, timber and tree crop plantations in Indonesia have grown rapidly with timber being promoted by the government through regulations and subsidies (World Bank, 2000). Until 1995, the degradation and loss of forest was mostly caused by selective logging. After 1995, however, the government focused more on the development of cash-crop plantations, including rubber, oil palm and pulpwood (Müller et al., 2014). As a result, Indonesia is one of the world leaders in the production of palm oil and other cash-crop products, such as timber, plywood, pulp and paper (FAOSTAT, 2016). Related to this, the palm oil, pulpwood and mining industries have been identified as main contributors to unwanted LULC change and forest loss in Indonesia, particularly in Sumatra, Kalimantan and Papua (Abood et al., 2014; Broich et al., 2011a; Carlson et al., 2012b; Gaveau et al., 2014, 2013; Phalan et al., 2013). These activities are accompanied by logging activities and by largescale fires that can appear even more severe and frequent under El Niño (Langner et al., 2007; Marlier et al., 2015; Siegert et al., 2001; Van Nieuwstadt and Sheil, 2005). The development of oil palm plantations in Indonesia has been turbulent throughout history and is growing drastically especially on Sumatra and Kalimantan (Abood et al., 2014; Broich et al., 2011b). Since its introduction in 1848, the cultivation area of oil palm has grown from only four seedlings to approximately 100,000 ha in 1967 to 2.5 million ha in 1997 (Casson, 1999) and an estimated 10.8 million ha in 2015 (USDA, 2015a). This rapid and large-scale expansion of oil palm plantations has provided important economic benefits, however, it has also become a reason for serious concern, since oil palm plantation development is playing a key role in tropical forest cover loss, the unequal distribution of benefits among stakeholders (Obidzinski et al., 2012) and land use conflicts (Gerber, 2011). Currently, the palm oil industry is identified as the third main contributor to forest loss (Abood et al., 2014) and its contribution to LULC change in Indonesia is expected to increase substantially in the coming years (Koh and Ghazoul, 2010; OECD/FAO, 2015). This expectation is based on the strong increase of oil palm plantations in recent decades (FAOSTAT, 2015) and the growing domestic and global demand for palm oil for food and non-food products, including biodiesel (Carriquiry et al., 2010; OECD/FAO, 2015). In 2012, almost half of the global crude palm oil (CPO) was produced in Indonesia, specifically, 26 million ton (MT) CPO out of 53 Mt CPO globally (FAOSTAT, 2014). Approximately 18 Mt CPO was exported, with the three largest importing regions being China, India and the European Union. Meanwhile, the domestic demand in Indonesia could increase to a projected 10.7 million ton (= 26% of 40.8 million ton CPO in total) in 2020 (Platts, 2015); as the Ministry of

14

Energy and Mineral Resources (MEMR) has planned to double the mandatory biodiesel blending for transportation and industry in only five years from 15% in 2015 to 30% in 2020 (MEMR Regulation 12/2015) (USDA, 2015b).

1.5

Problem definition

To understand LULC change, forest loss and how Indonesia can make the transition to a stage with stabilising high forest cover and halting deforestation, it is important, firstly, to have a better understanding of current and potential future LULC change processes and its drivers, and the consequent forest loss (Part I). Secondly, it is important to analyse the impacts of (unwanted) LULC change and forest loss on two important features of tropical forested landscapes: aboveground biomass and plant species richness (Part II) which are coupled to the two main global services of carbon sequestration and biodiversity (MEA, 2005). Part I and Part II of this dissertation focus mostly on West Kutai and Mahakam Ulu districts, because LULC change analyses at the landscape scale provide the best insight into the mechanisms underlying forest loss. Thirdly, it is important to analyse the measures that can be used to mitigate unwanted LULC change to guide Indonesia to the next stages of the forest transition: with stabilising forest cover and the potential for regeneration (Part III). Part III focuses on North and East Kalimantan as a whole and comprises LULC change processes related to all main commodities, in order to analyse and mitigate displacement effects. Part I. Land use and land cover change and impacts on forest cover Studies have shown that LULC change processes vary over time and across space because of varying local ecological, biophysical, socio-economic and political conditions (Brockhaus et al., 2012; Hansen et al., 2009; Luttrell et al., 2012; Mertens and Lambin, 1997). Additionally, LULC changes do not solely occur from one LULC type to another, but instead can be characterised by a sequence of changes, defined as trajectories (Dennis and Colfer, 2006; Ekadinata and Vincent, 2011; Inoue et al., 2013; Lambin, 1997; Mertens and Lambin, 2000; Petit et al., 2001). For instance, after forest clearance, the subsequent land cover or land use is not necessarily permanent, but may change over time into one or multiple different LULC types (e.g. Lambin 1997; Lambin et al. 2003; Ekadinata and Vincent 2011; Inoue et al. 2013). LULC change trajectories have been qualitatively and quantitatively analysed at the district level in Kalimantan and Sumatra (Carlson et al., 2012b; Dennis and Colfer, 2006; Ekadinata and Vincent, 2011; Inoue et al., 2013). The main trajectories involved conversions from a) forest to small-scale mixed land uses for the production of rice, fruit or jungle rubber, or from b) forest to small-scale or large-scale monocultures for palm oil and rubber production (Carlson et al., 2012b; Ekadinata and Vincent, 2011; Inoue et al., 2013). Also, trajectories were identified with agroforests or small-scale rubber as an intermediate land use (Ekadinata and Vincent, 2011; Inoue et al., 2013). However, a full picture of LULC change trajectories in space and time is lacking and therefore the aforementioned analytical steps are integrated in this dissertation. Such an integrated trajectory analysis incorporates the quantification and the schematic and spatially-explicit presentation of LULC change trajectories and processes by using spatial analyses supported by interviews with experts (Chapter 3). Policy making and civil society can be informed by projections of potential future land use and how these impact forest cover and local food production. Two important features influencing these projections are land development (see 1.4) and land (allocation) zoning. In Indonesia, land zoning

15

policies, laws and regulations have been developed to allocate specific zones in the Forest area and the Non-Forest area to forestry, agricultural development and conservation (Brockhaus et al., 2012; Broich et al., 2011a). Broadly speaking, the forest zones in the Forest area are allocated to forest conservation, logging or the development of forest commodities, while the Non-forest area and the Production forest for Conversion zone are allocated to non-forestry activities, including the development of estate plantations such as oil palm. However, some studies found that an important share of the land clearing and agricultural development in Indonesia has occurred in land allocation zones where these are prohibited (Brockhaus et al., 2012). Additionally, recent developments in land allocation zoning policies make the reclassification from all zones of the Forest area to the Non-Forest area possible (Minister of Forestry Instruction 33/2010; 34/2010; 29/2014). As a result of these reclassification regulations, larger areas may become available for the development of cash-crop plantations, and thus more frequent and large-scale land development may become a possible scenario. To date, it is unclear how changing land zoning regulations will impact the trajectories of LULC change and forest cover, particularly in view of the growing demand for cash crop products (Chapter 4). Part II. Impacts of land development and forest loss on carbon stocks and biodiversity Tropical forests store more carbon and inhabit more species than any other terrestrial ecosystem (Gibbs et al., 2007; Sodhi et al., 2010, 2004). The degradation of tropical forests and peatlands or the land clearance for agriculture or mining can lead to a strong reduction in biomass and species richness in the tropics (Agus et al., 2009; Danielsen and Heegaard, 1995; Danielsen et al., 2009; Dekker et al., 2015; Eisner et al., 2016; Fargione et al., 2008; Germer and Sauerborn, 2008; Gibbs et al., 2007; Immerzeel et al., 2014; Kotowska et al., 2015; Pearce and Moran, 1994; Warren-Thomas et al., 2015; Wicke et al., 2008b; Wilcove et al., 2013). LULC change can also result in net carbon storage, but only if land development occurs on lands with carbon stocks that are lower than the new land use, such as previously underutilised degraded lands, grasslands or shrublands (Danielsen et al., 2009; Fairhurst and McLaughlin, 2009; Fargione et al., 2008; Smit et al., 2013; Wicke, 2011). An increase in species richness is generally challenging to accomplish. AGB is not static, but spatially and temporally highly variable, particularly in the tropics due to varying biophysical conditions present, such as terrain and soil types, and anthropogenic disturbances such as fire or logging (Baker et al., 2004; Chave et al., 2003, 2001; de Castilho et al., 2006; Houghton, 2005). This makes its quantification and the avoidance of high AGB densities or high carbon stocks challenging (considering that about half of AGB consists of carbon) (Chapter 5). Biodiversity is important for the maintenance of ecosystem services and resilience (MEA, 2005). However, less than 1% of the publications on palm oil between 1970 and 2008 was related to biodiversity and species conservation. Although attention for this subject has increased, more research is needed on the response of biodiversity to land development. Lucey et al. (2015) found positive correlations, but with different orders of magnitude, between the responses of carbon stocks and potential biodiversity to different land uses. On the other hand, Murray et al. found low carbon stocks where there was a higher potential species richness (Murray et al., 2015). It is unknown what the effects on aboveground carbon stocks and plant biodiversity will be under different stringencies of land zoning and different levels of land development and whether these effects will be similar (Chapter 6). Part III. Measures to mitigate unwanted land use change A variety of measures and methods have been developed to mitigate the adverse social and environmental impacts of LULC change for the development of mining and agriculture, and

16

particularly of oil palm expansion. However, as described above, palm oil production is not the only contributor to unwanted LULC change in Indonesia. Thus, reconciling the production of all contributing commodities while maintaining forests, peatlands and local food production is important. The aforementioned nationwide moratorium on the issuance of new licences for concessions on primary forests and peatlands is one example. Additionally, mechanisms such as Reducing Emissions from Deforestation and forest Degradation plus (REDD+) (UN-REDD, 2013) have been developed to provide an incentive for actors to avoid deforestation and forest degradation. Meanwhile, land zoning tools such as the Responsible Cultivation Area method (Smit et al., 2013) can support the maintenance of forests by shifting agriculture towards under-utilised (degraded) lands that are suitable and available and have low carbon stocks and low biodiversity levels, while community land-use is respected (Wicke et al., 2011). However, the development of suitable and available under-utilised lands and responsible land-use planning and zoning at a regional scale have not yet been integrated with land-sparing strategies, e.g. yield improvements for multiple commodities. The integration of such strategies is important to mitigate unwanted LULC change and displacement effects (Abood et al., 2014; Angelsen and Kaimowitz, 2001; DeFries and Rosenzweig, 2010; Gutiérrez-Vélez et al., 2011; Laurance et al., 2014; Lee et al., 2014c) (Chapter 7).

1.6

Research questions and outline of the dissertation

Based on the above, we defined the following general research questions to assess whether it is possible for a tropical region to move through the forest transition curve, while maintaining a high forest cover: How to analyse LULC change processes and drivers in tropical landscapes in transition, the consequent impacts on forest cover, carbon stocks and biodiversity, and the mitigation of these impacts? As shown in the sections above, Indonesia is a very important example of a tropical country that is going through widespread LULC change and the associated forest losses. The provinces of North and East Kalimantan, and the districts of West Kutai and Mahakam Ulu therein, are regions that are representative of processes that are occurring in the outer islands of Indonesia. These regions are at the forefront of strong agricultural developments, widespread logging, large-scale fires and high forest losses (Fuller et al., 2004, 2010; Müller et al., 2014). We assessed LULC changes, processes and drivers, quantitatively and spatially-explicitly at the landscape level which enabled us to analyse the data with a sufficiently high resolution. The research question applied to the study area is as follows: In what stage of the forest transition curve is the region of West Kutai and Mahakam Ulu districts in the province of East Kalimantan, Indonesia, what are the processes and drivers of LULC change and forest loss and (how) can the region move towards a stabilised forest cover and potential reforestation, while bridging the forest transition? These research questions are divided in the following 5 research questions in Part I to III: Part I. Land use and land cover change and impacts on forest cover RQ 1. Which LULC change processes and trajectories at the local to landscape scale contribute to tropical forest loss and degradation?

17

RQ 2. What are the impacts of four contrasting LULC change scenarios with varying levels of land development and varying stringencies of land allocation zoning on LULC, and particularly on forest cover and local food production until 2030? Part II. Impacts of land development and forest loss on carbon stocks and biodiversity RQ 3. Which of a preselected set of biophysical and anthropogenic variables contribute significantly to the spatial variation of AGB at the regional level? RQ4. What are the impacts of four contrasting LULC change scenarios on aboveground carbon and plant biodiversity at the landscape level? Part III. Measures to mitigate unwanted land use change RQ 5. How can the production of palm oil be reconciled with the production of a set of other main commodities while mitigating unwanted direct and indirect LULC change?

The five research questions are addressed in Part I, II and III (see Figure 1.2). In Chapter 2, first, the case study area is described in more detail. In Chapter 3 and 4, LULC change processes and trajectories at the landscape level are analysed for 1990 and 2009 (RQ 1) and projected for 2009-2030 (RQ 2). In Chapter 5, the spatial variation of AGB and the explanatory variables of AGB are analysed (RQ

Part I

Part II

Part III

Chapter 3 Land use change 1990-2009 KuBar & MahUlu

Chapter 4 Land use change Projected for 2009-2030 KuBar & MahUlu

Chapter 5 Aboveground biomass and predictor variables 2008 North & East Kalimantan

Chapter 6 Impacts on aboveground biomass and species richness projected for 1990-2030 KuBar & MahUlu

Chapter 7 Mitigation unwanted land use change projected for 2008-2020 North & East Kalimantan

Figure 1.2. Outline of thesis chapters under Part I, II and III.

18

3). In Chapter 6, the impacts of LULC change on AGB or carbon and on plant species richness or biodiversity are analysed (RQ 4). In Chapter 7, four measures are analysed to mitigate unwanted LULC change by palm oil, pulpwood, rubber and rice production in East Kalimantan (RQ 5). In Chapter 8, I provide a synthesis on the main findings and place them in a broader context.

19

20

2 Case study area

2.1

North and East Kalimantan provinces, Indonesian Borneo

The natural resource-rich provinces of North and East Kalimantan are situated in the north-eastern part of Borneo (Figure 2.1). The province of North Kalimantan was previously part of East Kalimantan and was officially established on 25 October 2012. Because most of the available data are from the period up to 2012, we included both these provinces in the analyses. Henceforth, these provinces are indicated in this dissertation as North-East Kalimantan. The terrain consists of undulating slopes and altitudes up to about 2,200 m. The two provinces are rich in natural resources, such as oil, natural gas and coal, and have large tracts of forests in the lowland and mountainous areas. These characteristics make the province attractive for exploitation and development. This landscape is highly dynamic with regard to its past, current and expected land use changes. Until the early 1970s, the original land cover in the lowlands of North-East Kalimantan consisted of extensive dipterocarp forests with high aboveground biomass and species richness (Toma et al., 2005). However, large-scale degradation, deforestation and conversion to agricultural land have taken place, driven by forest and land development policies in the 1980s, (Murdiyarso & Adiningsih, 2006). High intensity logging was the most common land-clearing activity (Murdiyarso & Adiningsih, 2006), however, also large-scale forest fires occurred that were often initiated for land clearing purposes and were strengthened by El Niño (Siegert & Hoffmann, 2000; Siegert et al., 2001; Murdiyarso & Adiningsih, 2006; Slik et al., 2008). In 1997-98, again very destructive fires related to ENSO occurred, burning more than 5 million hectares of North-East Kalimantan’s primary and secondary forests (Hoffmann et al., 1999; Siegert & Hoffmann, 2000; Page et al., 2002). Hoffmann et al. (1999) have found that in that time period approximately 75 % of the burned forested lands were allocated for logging, timber or oil palm concessions. Over the last few decades, the frequency and spatial extension of fires have increased in North-East Kalimantan due to deforestation and degradation processes associated with logging, mining and agriculture, and intensifying droughts related to ENSO events (Slik et al., 2008). The most recent widespread and destructive forest fires occurred in 2015 (WRI, 2016).

21

North Kalimantan

West Kalimantan

Mahakam Ulu

East Kalimantan

West Kutai

Central Kalimantan

South Kalimantan 0

250 km

Kalimantan

Indonesia

Figure 2.1. West Kutai and Mahakam Ulu districts and their location in the region of North and East Kalimantan provinces, Indonesian Borneo.

22

In recent years, the North and East Kalimantan provinces were the 3rd largest greenhouse gasemitting provinces in Indonesia, with approximately 255 Mt CO2 emitted per year (GCF-TF Indonesia, 2011). In 2008, approximately 60% of the greenhouse gas emissions was related to agricultural expansion, including the development of oil palm plantations (GCF-TF Indonesia, 2011).

2.2

West Kutai and Mahakam Ulu districts, East Kalimantan

West Kutai and Mahakam Ulu districts are located in the western part of the province of East Kalimantan (Figure 2.1) and have a total land area of approximately 3.3 million hectares. The region has been subject to important political and administrative changes, including the division of the former West Kutai district into these two districts in 2012. We analysed the West Kutai and Mahakam Ulu districts as a single region because the data originate from the period before the division in two districts in 2012. The higher altitudes up to ~2,200 m in the north-west of the area are still covered with forest. Traditionally, the indigenous Dayak communities practise small-scale shifting cultivation for the production of rice, vegetables and fruit for household use, and rubber and rattan through traditional rubber and rattan gardens (Inoue et al., 2013). However, since the 1980s, these practices are increasingly being integrated into or replaced by the larger-scale and commercialised production of rubber and palm oil, and the extraction of timber, coal and gold (Inoue et al., 2013). As a result, the region experienced large-scale forest degradation and deforestation due to logging and forest fires (Müller et al., 2014). Additionally, in the same time period, the transition from small-scale shifting cultivation as the dominant land use toward the intensive cultivation of cash crops, caused forest loss in the lowlands in the southeast (Inoue et al., 2013; Müller et al., 2014). These land developments have strongly affected the original forests in the region that are dominated by the Dipterocarpaceae (Slik et al., 2002; Toma et al., 2005).

23

24

Part I. Land use and land cover change and impacts on forest cover

25

26

3

Land use and land cover change trajectories in a tropical forest landscape

Illustrated for the West Kutai and Mahakam Ulu districts, East Kalimantan, Indonesia Carina van der Laan, Arif Budiman, Stefan C. Dekker, Wiwin Effendy, André P.C. Faaij, Arif Data Kusuma, Pita A. Verweij This chapter has been submitted to the journal PLoS ONE

Abstract In Indonesia, land use and land cover (LULC) change for the palm oil, mining, timber and pulpwood industries is threatening tropical forests, biodiversity and ecosystem services. However, LULC change is a highly dynamic and complex process at the local level that varies over time and space, and that can undergo characteristic trajectories. We applied quantitative and spatially-explicit analyses using Landsat-based LULC maps, supported by field information and expert knowledge, to analyse LULC change trajectories and processes in the West Kutai and Mahakam Ulu districts in East Kalimantan from 1990-2009. Over two decades, forest cover declined substantially by forest degradation and forest conversion into smallholder rubber, pulpwood plantations, mixed cropland and oil palm plantations. More importantly, these developments followed characteristic trajectories, and became more dynamic in space and over time. Accounting for such trajectories is essential to improve future projections of LULC change and to support spatial planning policies.

27

3.1 Introduction Widespread land use and land cover (LULC) change has a strong negative impact on biodiversity and ecosystem services for tropical forest landscapes, where it causes a decline in carbon stocks and in the provision of food and livelihoods to local communities (Laurance et al., 2014; Phalan et al., 2013; van der Werf et al., 2009; Willemen et al., 2013). This also holds true for Indonesia, which is considered to be in the first stages of a large-scale forest transition (Rudel et al. 2009; Barbier et al. 2010; Lambin and Meyfroidt, 2010), meaning overall high deforestation rates, however with large regional differences between districts and provinces (Ekadinata et al., 2008; Hansen et al., 2009). Main contributors to LULC change and forest loss have been identified for Indonesia, including large-scale fires and logging activities (Langner et al., 2007; Siegert et al., 2001; Van Nieuwstadt and Sheil, 2005), and the pulpwood, palm oil and mining industries (e.g. Benhin 2006; Carlson et al. 2012; Gaveau et al. 2013; Phalan et al. 2013; Abood et al. 2014; Gaveau et al. 2014). The aforementioned studies have analysed and visualised the larger-scale LULC change patterns and processes, including for example deforestation, land cover degradation and conversion. LULC change is, however, a very diverse phenomenon at the local scale, because of the high geographical variability in biophysical environments, socio-economic activities and cultural contexts (Hansen et al., 2009; Mertens and Lambin, 1997). Additionally, the complex political and institutional context in Indonesia contributes strongly to local variations in LULC change (Brockhaus et al., 2012; Luttrell et al., 2012). For example, the governments at the district, provincial and national levels can have different interests, but are all involved in the decision-making process with regard to land use allocation, spatial planning and handing out of concession permits for land development (Brockhaus et al., 2012). Studies have shown that because of these varying local conditions and decisions, LULC change does not only vary spatially, but also over time, and is characterised by a sequence of changes, defined as trajectories (Dennis and Colfer, 2006; Ekadinata and Vincent, 2011; Inoue et al., 2013; Lambin, 1997; Mertens and Lambin, 2000; Petit et al., 2001). For instance, after forest clearance, the subsequent land cover or land use is not necessarily permanent, but may change over time into one or multiple different LULC types (e.g. Lambin 1997; Lambin et al. 2003; Ekadinata and Vincent 2011; Inoue et al. 2013). To account for the local spatial and temporal variability of LULC change in the tropics, LULC change processes and trajectories need to be analysed at the local to landscape scale (e.g. Lambin 1997; Petit et al. 2001; Lambin et al. 2003). LULC change trajectories have been analysed at the district level in Kalimantan and Sumatra, Indonesia (Carlson et al., 2012b; Dennis and Colfer, 2006; Ekadinata and Vincent, 2011; Inoue et al., 2013). The main trajectories involved conversions from a) forest to land uses that were small-scale mixed cultures for the production of rice, fruit or jungle rubber, or from b) forest to small-scale or large-scale monocultures for palm oil and rubber production (Carlson et al., 2012b; Ekadinata and Vincent, 2011; Inoue et al., 2013). Also, trajectories were identified with agroforests or smallscale rubber as an intermediate land use (Ekadinata and Vincent, 2011; Inoue et al., 2013). Inoue et al. (2013) qualitatively described LULC change trajectories in the former West Kutai district and presented these in a schematic overview that showed the complexity of LULC change in this area. Other studies in Kalimantan and Sumatra quantified LULC change trajectories based on pairs of LULC maps covering 20-30 years, thereby showing the relative importance of each of the trajectories (Carlson et al., 2012b; Dennis and Colfer, 2006; Ekadinata and Vincent, 2011). Anomaly maps have been developed at the regional scale for Sumatra and Kalimantan to visualise processes such as deforestation or agricultural expansion (Broich et al., 2011a, 2011b; Gaveau et al., 2014). Carlson et al. (2012) presented the LULC changes into oil palm in Ketapang district, West Kalimantan, in a spatially-

28

explicit way, highlighting the impact on other LULC types. However, to gain more insights into the temporal and spatial variation of LULC change trajectories in tropical landscapes, an integration of the abovementioned analytical steps is needed. Such an integrated trajectory analysis would incorporate the quantification and the schematic and spatially-explicit presentation of LULC change trajectories and processes. Furthermore, in most of the aforementioned studies, LULC types were aggregated into major categories, such as ‘cropland’ or ‘forest’. Consequently, processes such as land use intensification, forest degradation and regeneration could not be identified. In our study, we aim to characterise, quantify and visualise the LULC change processes and trajectories at the local to landscape scale that contributed to tropical forest loss and degradation by using spatially explicit analyses supported by interviews with experts. We selected the resource-rich districts of West Kutai and Mahakam Ulu, in the Indonesian province of East Kalimantan, because we expected highly dynamic LULC change processes there, driven by land development, the exploitation of natural resources and large-scale fires (Casson, 2001; Hoffmann et al., 1999; Inoue et al., 2013; Van Nieuwstadt, 2001). We conducted pixel-to-pixel analyses of LULC change processes and trajectories using Landsat data of the years 1990, 2000 and 2009. LULC change studies at pixel level have proven useful to define which LULC types were replaced by which other types over a certain timescale (e.g. Benhin 2006; Carlson et al. 2012; Gaveau et al. 2013; Phalan et al. 2013; Abood et al. 2014; Gaveau et al. 2014). Additionally, we quantified, schematically presented, and spatially-explicitly visualised the stepwise LULC changes, thereby clearly operationalising the LULC change trajectories that were previously only qualitatively defined for the study area by Inoue et al. (2013). To understand the LULC change processes and trajectories, we used detailed LULC classes, in order to be able to disentangle processes such as forest degradation, forest regeneration and land use intensification. The outcomes of our study can assist in the modelling of future land use and land cover change at the local to landscape scale and guide spatial planning of agriculture. The relative importance of the landscape-scale trajectories, their spatial and temporal variation and identification of the actors involved can support interventions for the maintenance of forests and local small-scale food production.

3.2

Material and methods

3.2.1 Case study region We analysed the West Kutai and Mahakam Ulu districts as a single area because the data originate from the period before the division in two districts in 2012. The districts have a total land area of ~33,000 km2 and are located in the western part of East Kalimantan (Figure 2.1 and Figure A 2 in Appendix), which is a quite remote and isolated area due to minimal infrastructure. Nonetheless, West Kutai and Mahakam Ulu are very attractive areas for widespread agricultural expansion and the extraction of timber, coal and gold (BPS, 2012). The indigenous Dayak communities used to practise small-scale shifting cultivation (Inoue et al., 2013). The original rotation cropping and the regeneration of forest made LULC change in the past cyclical (Mertens and Lambin, 2000). The production of rice, vegetables and fruit for household use, and rubber and rattan through traditional rubber and rattan gardens is now increasingly being integrated into or replaced by the larger-scale and commercialised production of rubber and palm oil, and the extraction of timber, coal and gold (Inoue et al., 2013).

29

3.2.1.1 Social, economic and political context Indonesia has a turbulent social, economic and political history. Government sponsored and spontaneous permanent transmigration from the inner to the outer islands of Indonesia has resulted in widespread agricultural development in Sumatra and Kalimantan (Benhin, 2006). After the economic crisis and resignation of Suharto in 1998, Indonesia has been in a transition period from a centralised system of governance towards a more open and liberal socio-political environment, known as Reformasi (Casson, 2001). This involved rapid formal and informal decentralisation processes that were driven mostly by decisions made by provincial and district level actors who found that the benefits from natural resources and agriculture flowed away from their regions to the national government and the private sector (Casson, 2001). Prior to Reformasi, forests were under the control of the central government, and Indonesia had an overcapacity in the wood processing industries, resulting in high deforestation rates. During Reformasi, the Indonesian government reformed the forestry sector, with attempts of more just and equitable benefits of forest resources management between governments, the private sector and local communities (Casson and Obidzinski, 2007). Meanwhile, district and provincial governments have received greater benefits from the development of agriculture in their jurisdictions. 3.2.1.2 Land allocation zoning In Indonesia, land is divided into Forest area and Non-forest area. According to the 1983 Forest Land Use Planning by Consensus (Tata Guna Hutan Kesepakatan: TGHK), Forest area has been categorised as the Production forest zone (Hutan Produksi: HP), Limited production forest zone (Hutan Produksi Terbatas: HPT), Watershed protection forest zone (Hutan Lindung: HL), Conservation forest zone (Kws. Suaka Alam dan Pelestarian Alam: KSPA) and Production forest for Conversion zone (Hutan Produksi yang dapat di-Konversi: HPK) (Broich et al., 2011a). These zones are under the jurisdiction of the Indonesian Ministry of Environment and Forestry. In the Forest area zones, only forest-related activities are allowed, such as forest conservation or logging. The development of estate plantations, such as oil palm plantations, is prohibited in Forest area (see Figure A 3A in Appendix for a map of the land allocation zones in the study area). Non-forest area (Kawasan Budidaya Non Kehutanan: KBNK, formerly called Areal Penggunaan Lain: APL) is designated for the development of agriculture. These land allocation zones do not necessarily indicate actual forest cover (Broich et al. 2011) (see Figure A 3A). 3.2.2 Data LULC maps were developed for the years 1990, 2000 and 2009 by automatic unsupervised classification combined with manual digitation of a time series of 30m resolution Landsat TM/ ETM-7 satellite imageries (see further Budiman et al. 2014). The LULC maps are shown in Figure A 2 (in Appendix). Digitising took place at a scale of 1:90,000 with a minimum mapping unit of approximately 8 mm x 8 mm on the map. To support the classification and validation of the LULC maps, we used a variety of data sources. For instance, field observations, topographic maps, floristic zone maps, soil maps, infrastructure data, survey data, focus groups discussions and expert knowledge. The validation of the LULC maps by the 55 field observation points resulted in a map accuracy of 89%. For more information about the classification process, see Appendix 1. For the analyses, the polygon LULC maps were geometrically rectified to ‘Equal Area’ and rasterised to 100 m. The LULC classes included three forest types representing different degradation levels, namely closed canopy, medium open canopy and very open canopy forest. Other vegetation types were shrubland and grassland. For the land uses for agriculture we distinguished between small-scale and large-scale

30

agriculture, including mixed cropland and smallholder rubber as small-scale land uses, and rubber plantation, pulpwood plantation and oil palm plantation as large-scale land uses. Furthermore, we classified cleared land, gold mining, coal mining, settlements and water (see further Appendix 1 and Table A 1). A semi-structured expert group interview was conducted with eight local field experts to analyse which LULC changes have occurred in the West Kutai and Mahakam Ulu districts between 2000 and 2009. We selected a semi-structured interview as this is an open procedure, allowing new ideas to arise. Additionally, we selected a group-consensus approach instead of an individual knowledge-elicitation approach (see Perera et al. 2011). The experts were asked what the main land uses and the main LULC change processes were in the study area in the past 10 years. We also asked what the main initial land uses were on (previously) forest land, if the initial land use in a certain area remained permanent or was changed into other land uses, and which land uses these were. Also, we asked why farmers choose certain land uses and decided to convert to others. The group-consensus discussion resulted in a short list of main LULC change trajectories. 3.2.3 Analyses We conducted three analytical steps to quantify and visualise the LULC change processes and trajectories (for more details on the processes and trajectories, see Appendix 2). First, we quantified the LULC for 1990, 2000 and 2009 based on the LULC maps and conducted pixel-to-pixel cross tabulations in ArcGIS 10.1 to define what LULC types increased and what LULC types decreased in land area between 1990-2000 and 2000-2009 and during the whole time period 1990-2009. Second, trajectory analyses (see for example Ekadinata and Vincent 2011) were conducted by analysing the remote-sensing based LULC maps and the interviews with the experts. The trajectory analyses consisted of the following steps. The pixel-to-pixel cross tabulations from the first step resulted in matrices with on the vertical axis the initial LULC types and on the horizontal axis the LULC types of the subsequent time step. These LULC change matrices reflect the area (ha) in which a one-step trajectory occurred within the selected time period, namely 1990-2000, 2000-2009 and 1990-2009 (see Table A 2 to Table A 5 in the Appendix). Subsequently, the resulting trajectories were translated into a schematic diagram, so that the trajectories were clearly visualised. To acquire more detailed information about the LULC changes, processes and trajectories that have occurred, including which trajectories were dominant, and the time step between each of the LULC changes, we consulted LULC change field experts. Finally, as the cross tabulations were conducted on a pixel-to-pixel basis, we were able to visualise the processes and trajectories spatially explicitly, by linking the outcomes of the pixelto-pixel tables to the pixels on the maps. Third, we overlaid the LULC maps for 2009 and the LULC change processes map for 2000-2009 with concession and land allocation zoning data for the year 2009 (the last two data sets were only available for 2009).

3.4 Results 3.4.1 Area change of land cover and land uses The LULC maps showed that in 1990 most of the land (~88%) in the West Kutai and Mahakam Ulu districts was covered with forests that varied from closed to very open canopy (Figure 3.1, Figure A 2 in Appendix, Table 3.1). The remaining ~12% of the land was mostly covered with shrublands, grasslands and pulpwood plantations, or was utilised for the small-scale production of rubber or for mixed cropland (Figure 3.1 A, Figure A 2, Table 3.1). The pixel-to-pixel cross tabulations of the LULC

31

Land area (million ha)

3.5

Closed forest Medium open forest Very open forest Smallholder rubber Pulpwood plantation Oil palm plantation Settlement Mixed cropland Coal mining Gold mining Rubber plantation Shrubland Grassland/ cleared land Water

3.0 2.5 2.0 1.5 1.0 0.5 0 1990

2000

2009


3.5 3.0 2.5 Total forest cover Closed forest Medium open forest Very open forest Shrubland

2.0 1.5 1.0 0.5 0

1990

2000

2009

0.20 Rubber plantation Coal mining Mixed cropland Oil palm plantation Pulpwood plantation Smallholder rubber Total land use

0.15 0.10 0.05 0

1990

2000

2009

9002


0.25

Figure 3.1. LULC change in the West Kutai and Mahakam Ulu districts between 1990 and 2009, based on quantitative analyses of the LULC maps (Budiman et al., 2014): top,A) trend of LULC change linearly extrapolated from 1990 to 2000 and from 2000 to 2009; middle,B) change in natural land cover types (excluding grasslands); bottom,C) LULC change in agricultural lands (excluding gold mining and settlements).

maps showed, however, that between 1990 and 2009 about one third (~1 million hectares or Mha) of the LULC in the study area changed (Table 3.2). According to the pixel-to-pixel cross tabulations, between 1990 and 2000, total forest cover declined with ~9% (Table A 7) and the land used for agriculture, mining and settlements increased more than threefold to ~229,100 Mha (Figure 3.1. A, B and C, Table 3.1). The land used for smallholder rubber cultivation, pulpwood plantations, mixed cropland and oil palm plantations expanded greatly, particularly between 2000 and 2009, when forest cover declined with 5% (Table A 7). Also, the grassland area doubled from 1990 to 2009 to 103,000 ha (Table 3.1). The expansion of coal mining, settlements and gold mining was substantially lower than of the agricultural LULC types (Figure 3.1A). Not only the total forest land area decreased, but also the quality of the forest in terms of canopy cover declined. Closed canopy forests declined from 1990 to 2009 by ~29% to ~1.1 Mha in 2009 (Figure 2.2B, Table 3.1 and S7). In contrast, medium open and very open canopy forests increased with

32

Table 3.1. Absolute land area (ha) per land use and land cover type and relative land use and land cover changes (%) between 1990, 2000 and 2009. 1990 Natural land cover types 3,200,400 Total forested land 2,902,300 1. Closed canopy forest 1,579,000 2. Medium open canopy forest 957,200 3. Open canopy forest 366,000 4. Shrubland 248,900 5. Grassland/cleared land 49,200 Land use types 68,600 6. Mixed cropland 11,600 7. Smallholder Rubber 44,000 8. Rubber plantation 1,000 9. Pulpwood plantation 8,100 10. Oil palm plantation 100 11. Gold mining 100 12. Coal mining 600 13. Settlement 3,100 14. Water 24,800 Total

2000

2009

% of total land in 1990



3,155,500 2,797,600 1,264,400 1,114,300 418,900 309,500 48,400 108,300 15,600 66,100 1,400 16,400 4,200 600 800 3,200 29,800

3,033,500 2,651,900 1,114,400 1,094,600 442,800 278,600 103,000 229,100 26,200 122,500 3,000 31,700 31,200 100 7,800 6,700 31,100

97 88 48 29 11 8 2 2 0.4 1 0 0 0 0 0 0.1 1

96 85 38 34 13 9 2 3 0.5 2 0 1 0.1 0 0 0.1 1

92 81 34 33 13 9 3 7 1 4 0.1 1 1 0 0.2 0.2 1

3.3 Mha

3.3 Mha

3.3 Mha

The sum and percentages are rounded.

Table 3.2. Quantified LULC change processes based on remote sensing data (Budiman et al., 2014) that occurred in the periods 1990-2000, 2000-2009 and 1990-2009. See for a definition of the LULC change processes Appendix 2. Area (ha)

No LULC change Total LULC change Degradation Deforestation Land clearance Regeneration Conversion to agricultural land Abandonment and regeneration Water Total land area

% of total LULC change

1990-2000

2000-2009

1990- 2009

2,759,000 530,000 365,000 92,000 1,600 28,000 42,000 1,500 5,000

2,675,000 617,000 259,000 114,000 49,000 61,000 130,000 4,500 1,000

2,268,000 1,019,000 569,000 178,000 33,000 68,000 169,000 3,000 6,000

3.3 Mha

3.3 Mha

3.3 Mha

1990-2000

2000-2009

1990-2009

69 17 0 5 8 0

42 18 8 10 21 1

56 17 3 7 17 0

The sum and percentages are rounded.

33

~14 and ~21% between 1990 and 2009 (Table A 7) to ~1.1 and ~0.44 Mha, respectively (Figure 3.2B, Table 3.1). From the pixel-to-pixel cross tabulations, we were able to characterise and quantify six LULC change processes that occurred in the West Kutai and Mahakam Ulu districts. These were degradation, deforestation, regeneration, abandonment and regeneration, conversion to agricultural lands or mining sites, and land clearance of non-forest LULC types (see Appendix 2 and Table A 6). Compared to the period 1990-2000, in the period 2000-2009 the area with LULC change increased by ~17% (Table 3.2). Over the ~20 years studied, degradation was the most dominant LULC change process in terms of land area (Table 3.2). Deforestation and conversion to agricultural land or mining sites were the second and third most contributing LULC processes (Table 3.2). Deforestation continued at a comparable rate between 1990 and 2009 (~17-18% over 10 years) (Table 3.2). However, degradation decreased from 69% of the total LULC change in the period 1990-2000 to 42% of the total LULC change in 2000-2009, while conversion to agriculture or mining increased from 8% to 21% for these two time periods (Table 3.2). 3.4.2 Quantified and visualised LULC change processes and trajectories The main LULC change trajectories and processes were quantified and visualised based on the remotesensing based LULC maps and the interviews with the experts. 3.4.3 LULC change processes and trajectories identified by pixel-to-pixel cross tabulations Figure 3.2 shows the quantification and schematic visualisation of the LULC change processes and trajectories based on the pixel-to-pixel cross tabulations of the LULC maps. The boxes on the left indicate the naturally-occurring LULC types and the boxes on the right the agricultural land uses, mining sites or settlements. Each arrow in Figure 3.2 shows the change from one LULC type to another that occurred from 1990 to 2000 (Figure 3.2A) and from 2000 to 2009 (Figure 3.2B). The thickness of each arrow indicates the area of change for each of the trajectories within the given time period. The LULC changes identified for the entire period of 1990-2009 are shown in Figure A 4 (in Appendix). Figure 3.2 A and B show that most of the forest was lost by degradation and deforestation. The main land uses to which these natural lands were converted to during the period analysed included rubber, pulpwood plantations, mixed cropland and oil palm plantations. For the time periods analysed, intermediate LULC types were forest, shrubland, grassland and mixed cropland; this means that these were the result of conversions from and to other LULC types. Permanent LULC types were oil palm plantations, pulpwood plantations and settlements; these were replacing other LULC types, but were not replaced by other land uses during the time periods analysed. Interestingly, smallholder rubber was a permanent LULC type in the period 1990-2000; however, it was both an intermediate and a permanent LULC type in the period 2000-2009, when this LULC type was also converted into rubber, oil palm and pulpwood plantations, coal mining sites and settlements. The increase in the number of arrows in Figure 3.2B compared to Figure 3.2A clearly shows that LULC change increased and became more dynamic in the period 2000-2009, compared to 1990-2000. Additionally, in the period 1990-2000, more degradation and deforestation occurred, and conversions to small-scale agriculture occurred more frequently than conversions to large-scale monocultures (Figure 3.2A). After 2000, however, the number of conversions to large-scale monocultures, for example oil palm and pulpwood plantations, increased substantially (Figure 3.2B). 3.4.4 LULC change processes and trajectories identified by field experts We consulted local field experts to acquire more detailed information about the LULC changes, processes and trajectories that have occurred in the study area in the past decade. The main LULC

34

≤2,500 5,000 10,000 20,000 Closed canopy forest

50,000

Mixed cropland

100,000 200,000

Smallholder rubber Medium open canopy forest

250,000 ha Rubber plantation

Settlement Very open canopy forest Oil palm plantation Shrubland

Coal mining

Gold mining Grassland/ Cleared land Pulpwood plantation

Closed canopy forest

≤2,500 5,000 10,000 20,000 50,000

Mixed cropland

100,000 200,000



Settlement

Very open canopy forest

Oil palm plantation Shrubland

Coal mining

Gold mining

Pulpwood plantation

9002

Grassland/ Cleared land

Figure 3.2. The LULC trajectories (single arrows) identified in the West Kutai and Mahakam Ulu districts: A, top figure) in the period 1990-2000; B, bottom figure) in the period 2000-2009. Each arrow shows a trajectory. The thickness of the lines (see legend) indicates the number of hectares involved in the LULC trajectory.

35

change processes identified by the experts (thick arrows in Figure A 5 in Appendix) were large-scale deforestation and conversion by concession holders, and small-scale conversions of forest lands, shrublands or grasslands to mixed cropland and/or smallholder rubber. On a much smaller scale, regeneration also occurred from grassland to shrubland and forest types. According to the experts, LULC change has been highly dynamic in the region with LULC types rapidly succeeding each other. However, the time steps between the LULC conversions varied. The following trajectories were identified based on the group-consensus discussion with the experts (Figure A 5) in order of complexity: A. One-step trajectories of degradation and/or deforestation to grasslands or shrublands; B. Multiple-step trajectories of (i) deforestation to grasslands, and (ii) conversion from grasslands to large-scale plantations; C. Multiple-step trajectories of (i) degradation and/or deforestation and conversion to small-scale mixed land uses, (ii) conversion after 1-2 years to smallholder rubber, and (iii) in certain cases, further to monocultures, mostly oil palm. The time period to convert smallholder rubber to oil palm varied locally. According to the experts, the choice of the smallholders between maintaining their small-scale rubber production or converting their lands into oil palm plantations (trajectory C in Figure A 5) depended on the price competitiveness of rubber versus palm oil and on the presence of success stories in the village regarding oil palm cultivation. More importantly, when small-scale rubber fields were converted into oil palm plantations, some smallholders established new mixed cropland areas in newly obtained forested lands, and LULC change trajectory C was initiated again. According to the experts, forest, cleared lands, mixed cropland and smallholder rubber were intermediate land uses, and oil palm plantations, pulpwood plantations and coal mining were permanent land uses between 2000 and 2009. 3.4.5 Spatially-explicit analyses of LULC change processes and trajectories We mapped LULC change processes for 1990-2000 (Figure 3.3A) and 2000-2009 (Figure 3.3B) to indicate the spatial variation of the LULC change. The matrix of coloured pixels is more complex in Figure 3.3B than in Figure 3.3A, which indicates that LULC change became more dynamic over time. Forest degradation (yellow pixels) and deforestation (red pixels) were more clustered in the period 1990-2000 than in 2000-2009 (Figure 3.3A and B). In Figure 3.4, we mapped the LULC trajectories related to smallholder rubber and oil palm in 1990-2009. The yellow pixels show that most of the smallholder rubber remained unchanged (92% of total smallholder rubber). In turquoise, we plotted the share of the land area where smallholder rubber changed to other land uses (~7% of total smallholder rubber in 1990), such as settlements and oil palm plantations, which shows the loss of smallholder rubber. In blue, we showed the land that was converted from any LULC type to smallholder rubber, which shows the further expansion of this land use. In red, the expansion of oil palm is shown. 3.4.6

Spatially-explicit analyses of LULC change in concessions and land allocation zones

LULC change in concessions The results of the overlay of the LULC types and LULC change processes that occurred in the period 2000-2009 with concession types are presented in Table A 8 andTable A 9, respectively. Table A 8 shows that 51% of the study area was allocated to timber, logging and oil palm concessions. Almost half of the remaining forests in 2009 were found in concessions, mostly in logging concessions (~40%) and oil palm concessions (~5%) (Table A 9). Most of the closed canopy forest was located outside 36

A

B

0

100 km

9002

Abandonment and regeneration Clearance Conversion Deforestation Degradation Regeneration No change

Figure 3.3. Spatially explicit visualisation of the LULC change processes that occurred in the West Kutai and Mahakam Ulu districts: A) in the period 1990-2000; B) in the period 2000-2009 (1 raster cell is 100 x 100 m).

concessions (~72%). In the logging and oil palm concessions, also most of the forest degradation and deforestation occurred between 2000 and 2009, with 51% and 38% in logging concessions and 11% and 20% in oil palm concessions, respectively (Table A 9). Regeneration occurred mostly outside concessions (49-60%, Table A 9).

37

0

50 km

9002

From other land use/cover type to oil palm From other land use/cover type to smallholder rubber Oil palm Smallholder rubber From smallholder rubber to other land use/cover type

Figure 3.4. Spatially-explicit representation of the land use and land cover (LULC) trajectories related to smallholder rubber and oil palm between 1990 and 2009 (1 raster cell is 100 x 100 m).

About 64% of the oil palm in 2009 was planted within oil palm concessions (~16,600 ha). Additionally, in 2009, only less than ~6% of the oil palm concessions was actually planted with oil palm (Table A 8). The remainder area in oil palm concessions was covered with forest (~48%), grass or shrubs (~33%) and smallholder rubber (9%,Table A 8). Smallholder rubber was mostly planted in oil palm (~46%) and logging concessions (~39%), while mixed cropland was evenly spread throughout the three concession types (Table A 8). In summary, most of the remaining forest in 2009 was found in logging and oil palm concessions, where also most of the forest loss and degradation occurred between 2000 and 2009. Within these

38

concessions, most of the conversions occurred to smallholder rubber (in all concessions), oil palm plantations (in oil palm concessions) and pulpwood plantations (timber concessions). More details are included in Table A 8.

LULC change in land allocation zones Overlaying the 2009 LULC map with the land allocation zoning map demonstrated that the LULC types were generally found in the zones that were designated for these LULC types (Table A 10). In 2009, most of the natural forests were located in the Limited production forest zone and Watershed protection forest zone ~50%, Table A 10). As expected, most oil palm plantations (82%), smallholder rubber (73%), mixed cropland (69%), coal mining (87%) and rubber plantations (88%) were present in the Non-forest area (Table A 10). Concession types in land allocation zones However, there is some discrepancy between the concession and land allocation zoning data (Table A 11). For example, only ~75% of the oil palm concessions was located in the Non-forest area, while ~21% was located in the Production forest zone (Table A 11).

3.5 Discussion In this paper, we have shown that an integrated analysis at the local to landscape level allows a more comprehensive estimate of the main land uses and the main LULC change processes and trajectories that have contributed to forest cover loss. The cross tabulations and schematic visualisation of the LULC change processes and trajectories, the information acquired from the experts and the spatiallyexplicit analyses were complementary to each other and resulted in the following findings. First, forest degradation and deforestation were the dominant processes that led to the decline of closed canopy forests over the ~20 year period. Forest regeneration also occurred, but on a much smaller scale. Over the time period analysed, a shift occurred from mainly forest degradation and deforestation and the development of small-scale agriculture in 1990-2000 to an increased development of large-scale monoculture agriculture in 2000-2009. These findings are in line with the findings of Ekadinata and Vincent (2011), Inoue et al. (2013) and Müller et al. (2014) for the Bungo district in Sumatra and the former West Kutai district. This shift coincided with the Reformasi phase in Indonesia, in which before 1999, centralised policies were more focused on the establishment of logging concessions and pulpwood plantations, while after 1999, the decentralised policies had a stronger focus on agricultural expansion (Casson, 2001). Besides, this is also supported by the study of Inoue et al. (2013), who found that in the study area, villagers show an increased interest in developing commercialised rubber plantations next to swidden agriculture and traditional rubber gardens. The analysed districts seem to be in a transition from mostly forest cover and small-scale mixed land uses to dominantly monocultures (Rudel et al., 2009), with small-scale mixed land use as an intermediate LULC type (Ekadinata and Vincent, 2011). This transition is driven by multiple factors, including the governments’ interests for profitable cash crop plantations for the production of rubber and palm oil, for example, and people’s responses to local economic opportunities (Carlson et al., 2012b; Casson, 1999; Lambin et al., 2001; Warren-Thomas et al., 2015). However, such a transition from small-scale mixed land uses to monocultures can indicate the increased interest of small-scale farmers to convert their mixed land use into monocultures, either independently or through plasma-nucleus schemes (Feintrenie et al., 2010; Rist et al., 2010), or it can indicate that small-scale farmers are being displaced,

39

which can negatively impact their livelihood (Mertz et al., 2009). In turn, to fulfil local food demands, food production may shift elsewhere, risking an additional conversion of forests. Second, we found that LULC change in the study area occurred in characteristic trajectories and became more complex and dynamic over time and space. By the cross tabulations, we were able to create a shortlist of the main land uses that contributed to the decline or degradation of forest over timescales of approximately ten years, between 1990 and 2009. These included smallholder rubber, pulpwood plantations, mixed cropland and oil palm plantations. By the combined analyses of the remote sensing data and the expert knowledge, we identified forest, shrubland, grassland and mixed cropland as intermediate land uses, smallholder rubber as both an intermediate and permanent land use, and oil palm plantations, pulpwood plantations and settlements as permanent land uses. In the Bungo district in Jambi, Ekadinata and Vincent (2011) found that rubber agroforest was an intermediate type of land use with a high likelihood of conversion, indicating the occurrence of these trajectories in other areas as well. The aforementioned shortlist is similar to the main contributors to forest loss that have been identified in previous studies (e.g. Gaveau et al. 2013; Abood et al. 2014). However, instead of ranking these as LULC ‘contributors’ (Abood et al., 2014), we thus found that these land uses had been part of characteristic LULC change processes and trajectories that had contributed to LULC change, in particular forest cover loss. As a result, if the intermediate land uses in these trajectories were not observed during a certain time period, one of the land uses involved can be misinterpreted as a main contributor. In addition, by the cross tabulations, we were able to quantify and visualise the intermediate and permanent land uses and the one-step trajectories, which was complemented by the expert knowledge that enabled the identification of more steps in the trajectories and estimation of the time step involved in each conversion. The quantification and visualisation of LULC change processes and trajectories with detailed LULC classes in this study adds to earlier studies on LULC change trajectories (e.g. Ekadinata and Vincent 2011; Carlson et al. 2012; Inoue et al. 2013). Third, our spatially explicit analyses showed that in the period 2000-2009, most of the forest loss and degradation occurred in logging and oil palm concessions (~60 %), which is similar to the estimates for entire Kalimantan (Abood et al., 2014). In 2009, also most of the remaining forest was found in these concessions. Furthermore, most of the monoculture plantations and concessions in our study area were developed in the land allocation zones as designated by the government, which is thus in line with the government’s spatial planning. Nonetheless, about one third of the forest degradation and deforestation, and almost half of the conversions to agriculture occurred outside concessions, and partly outside the designated land allocation zones. Our results thus demonstrate that spatial planning by the government played an important role in the expansion of monoculture agriculture into forest areas, as most of this expansion and forest loss occurred inside of concession areas. Spatial planning can thus play a substantial role in the sustainable development of agriculture which involves the maintenance of forest cover, particularly the high forest cover that still exists in the current concession areas. A method that can support the sustainable development of agriculture and maintenance of forests is the so-called Responsible Cultivation Area method (Smit et al., 2013). If supported by thorough ground checks, this method can be used to identify suitable and available areas such as under-utilised grasslands or ‘degraded lands’ for the expansion of oil palm and other monoculture crops. However, the focus of policy making and spatial planning with regard to agricultural expansion should not be exclusively on the responsible expansion of monoculture plantations. As we found in our study that smallholders played an important role in LULC change, the inclusion of small-scale agricultural development is also important. These abovementioned results are essential to understand and project which LULC change processes may occur in certain areas. Our results can be used to parameterise the type and sequence

40

of LULC conversions in LULC change modelling. Additionally, the use of annual remote sensing data could have supported the identification of more intermediate steps in the trajectories and the quantification of their spatial and temporal dimensions (Broich et al., 2011a). This information is particularly important for unravelling what is exactly contributing to forest loss and degradation in the tropics, where land development and LULC change are very dynamic in space and over time (Brockhaus et al., 2012; Hansen et al., 2009; Luttrell et al., 2012; Mertens and Lambin, 1997). Unfortunately, no annual cloud-free remote sensing data were available at a moderate spatial resolution for the study area in the time period analysed, although more recently such data have become available (e.g. by SarVision and the World Resources Institute). Therefore, we used the interviews with the experts to understand the LULC change processes and trajectories that have occurred between the remote sensing observation years, which makes the use of these data very relevant for our questions. Because LULC and ecological, social and political conditions are very dynamic in the study area, we recommend more regular and more intensive field data collection in subsequent trajectory analyses, for example on farmers’ choices, illegal logging practices, forest fires and immigration. Using such data, also the land uses mixed cropland and smallholder rubber, land uses that can constitute a mixture of small-scale farming practices, can be refined. Our study shows that trajectory analyses can provide a more comprehensive view on which land uses are involved in LULC change and forest decline, which is particularly important for spatial planning in tropical landscapes where agriculture develops rapidly (Dennis and Colfer, 2006; Ekadinata and Vincent, 2011; Inoue et al., 2013; Lambin, 1997; Mertens and Lambin, 2000; Petit et al., 2001). With the growing global demand for rubber and palm oil, and the higher returns of these cash crops for small-scale farmers, more conversions from small-scale mixed land uses to monocultures are expected. Such land intensifications can change the socio-cultural character of the landscape, impact local food production and result in declined levels of biodiversity (Ahrends et al., 2015; Ekadinata and Vincent, 2011; Fitzherbert et al., 2008) and carbon stocks (Lee et al., 2014a). As West Kutai and Mahakam Ulu are remote districts, it is important that communities, and particularly smallholders, are informed about these adverse effects of land intensification so that they can make well-informed decisions about using their lands for forest maintenance, mixed cropland and local food production or for the monoculture-based production of export commodities. Besides, involving communities in the spatial planning process is recommended, as they may play an important role in LULC change and may be affected by it. If forest landscapes are to be maintained, maintenance of forests and small-scale mixed land uses needs to be incentivised, for example by REDD+ or through subsidies for local food production.

3.6 Conclusions This study aimed to quantify and visualise the LULC change processes and trajectories that have contributed to forest loss and degradation in a tropical forest landscape in a spatially-explicit way. Pixel-to-pixel cross tabulations, spatial analyses and expert knowledge enabled a comprehensive estimate of LULC types and of the LULC change trajectories and processes that contribute to forest cover loss and degradation in West Kutai and Mahakam Ulu districts in the Indonesian province of East Kalimantan. This approach is useful for application in other tropical forest areas where strong agricultural developments are taking place. Particularly in areas where no annual remote sensing data are available, the integration of expert knowledge is of great importance.

41

Our results show that the LULC of about one third of the study area changed in the period 19902009, with a decreasing forest cover of approximately 9%. Based on our quantitative and spatial analyses, forest degradation and deforestation were found to be the dominant LULC change processes causing a declining forest cover and quality. Additionally, we found that LULC change became more dynamic and complex over time and in space. Our study also shows that LULC change, agricultural expansion and forest loss do not occur at random, but follow characteristic trajectories and (at least partly) concession and land allocation zoning policies, involving multiple LULC types. We compiled a shortlist of LULC types that replaced forest cover over timescales of approximately ten years, namely smallholder rubber, pulpwood plantations, mixed cropland and oil palm plantations. The integrated trajectory analyses were useful in showing that LULC change occurred in mostly one-step trajectories of degradation and deforestation; multiple-step trajectories from forest to grassland and further to large-scale monocultures; and multiple-step trajectories from forest to small-scale mixed land uses, to smallholder rubber, and sometimes further to monocultures, mostly oil palm. By these trajectories, we have shown that not only forests are vulnerable for conversion, but also small-scale mixed land uses, resulting in land use intensification. By this study, we have thus shown the relative importance of each of the trajectories, where these occur in the landscape, how these change over time and which actors are involved. The analyses of LULC change processes and trajectories may be further refined by using annual remote sensing data. The type and sequence of LULC conversions that result from trajectory analyses can be used for the parameterisation of LULC changes in order to improve LULC change modelling. We recommend involving industries and local communities, and particularly smallholders, in the spatial planning process, as some could play an important role in agricultural development and/or may be affected by LULC change. Because most of the monocultures and concessions were developed in the designated land allocation zones and thus in accordance with the government’s spatial planning, the government can and should play an important role in guiding the expansion of agriculture in more sustainable directions that involve the maintenance of forests and local food production.

42

4 Impacts of land development and land zoning policies on land use and forest cover projected for 2030 Illustrated for the West Kutai and Mahakam Ulu districts, Indonesia Carina van der Laan, Judith A. Verstegen, Stefan C. Dekker, Laurens Bakker, Pita A. Verweij, Arif Budiman, André P.C. Faaij This chapter has been submitted to the journal Land Use Policy

Abstract With the growing demand for cash crop products, such as palm oil, rubber, and pulpwood and also coal, the extent of forest loss in tropical regions such as Indonesia is increasing rapidly. Recent developments in the land allocation zoning policies in Indonesia make the reclassification from Forest area, i.e. the area allocated for forest conservation, logging or the development of forest commodities, to Non-forest area possible. As a result, larger areas may become available for the development of cash-crop plantations and thus large-scale land development becomes a possible scenario. This study aims to explore the impacts of the land allocation zoning policies and of different levels of land development on forest cover and local food production. To this purpose, we modelled and visualised the impacts of LULC change on forest cover and local food production in the West Kutai and Mahakam Ulu districts under four contrasting scenarios between 2009 and 2030 with the spatio-temporal PCRaster Land Use Change (PLUC) model. The four ‘what-if ’ scenarios included i) limited and unlimited land development for oil palm, pulpwood, rubber, mixed cropland, mining and settlements, and ii) restricted and unrestricted land allocation zoning. We found strong varying effects on forest cover and food production under the scenarios. With limited development, LULC change occurs mostly in the lowlands and forest loss between 2009 and 2030 is small (~ 0.1 Mha, ~4%). Under the unlimited-restricted scenario, however, forest cover declines stronger with 0.4 Mha (~17%), and under the unlimited-unrestricted scenario, forest cover decline is even stronger (1.6 Mha, ~60%) in 2009-2030, where LULC change shifts into the higher altitudes. With restricted zoning, we found more

43

displacement of land uses than with unrestricted zoning. The displaced land uses include smallholder rubber, mixed cropland and pulpwood plantations. The loss of forests and the relocation of food production to the higher altitudes that will result from this, is expected to have negative impacts on local food production, ecosystem services and biodiversity.

4.1 Introduction With the growing demand for cash crop products, such as palm oil, rubber, timber, pulpwood and coal, the extent of forest loss and peatland conversion in tropical regions such as Indonesia is increasing rapidly (Abood et al., 2014; Gaveau et al., 2013; Hansen et al., 2009; Koh et al., 2011). Natural forest cover in Kalimantan, the Indonesian part of the island of Borneo, declined with approximately 4.7% between 2000 and 2010, from 30.4 Mha in 2000 to 28.9 Mha in 2010 (Gaveau et al., 2013). Such largescale land use and land cover (LULC) changes by means of logging and fire lead to large quantities of greenhouse gas emissions and biodiversity losses (Abood et al., 2014; Carlson et al., 2012b; Fitzherbert et al., 2008; Slik et al., 2010b, 2008; van der Werf et al., 2009). Additionally, the conversion of rice fields and mixed croplands into monoculture rubber and oil palm plantations through characteristic trajectories of LULC change is expected to lead to a decline in local food production (Susanti and Burgers, 2012; see also Chapter 3). To prevent further deforestation and as a result of international collaborations, Indonesia has extended the nationwide moratorium of 2011 on the postponement of issuance of new licences of oil palm, pulpwood and selective logging concessions in primary forests and peatlands (Presidential Instruction 10/2011; 6/2013; 8/2015). Such developments are in line with the national Conservation Law (Law 5/1990) under which a series of National Parks has been established. In addition, they are consistent with the national emphasis on REDD+ (i.e. Reducing Emissions from Deforestation and Forest Degradation +) schemes which seek profit from forests through conservation over deforestation (Bakker and Fristikawati, 2014). The effectiveness of the moratorium and REDD+ initiatives for the conservation of forests and peatlands is, however, threatened by contrasting political and economic interests and the subsequent decision to use land for agricultural over forest products (Benhin, 2006) (Brockhaus et al., 2012; Butler et al., 2009; Luttrell et al., 2012; Murray et al., 2015; Wibowo and Giessen, 2015). Because of the recent moratorium, much attention is currently focussed on the implementation and revision of policies, laws and regulations that allocate land to forestry, agricultural development and conservation (Brockhaus et al., 2012; Broich et al., 2011a). Under the current land allocation zoning policies, land is subdivided into Forest area and Non-forest area. Broadly speaking, the forest zones in the Forest area are allocated to forest conservation, logging or the development of forest commodities, while the Non-forest area and the Production forest for Conversion zone (i.e. a Forest area zone that can be converted to Non-forest area) are allocated to non-forestry activities, including the development of estate plantations such as oil palm. During decentralisation in the past 15 years, the land use administration in Indonesia has become highly fragmented with more authority to deliver land administration services concentrated at the provincial and district government levels, besides the national government. While the criteria for land allocation zoning are established by the Ministry of Environment and Forestry1F and the National Development Planning Board, various ministries, different levels of government administration and different agencies have discretional space in interpreting and applying these criteria. Consequently, overlaps and contradicting usage are possible, based on various legislations within the same national legal system. Land allocation thus takes place in

44

a complex landscape of institutions and legal frameworks at multiple decision-making levels (Bakker and Reerink, 2015; Brockhaus et al., 2012) and is subject to multiple sectoral policies and interests. Studies have focused on the effects of land development on forest cover and the discrepancies between LULC, concessions and land allocation zoning. Some studies found that an important share of the land clearing and agricultural development in Indonesia has occurred in land allocation zones where these are prohibited (Brockhaus et al., 2012; Broich et al., 2011b; Chapter 3). For example, Broich et al., (2011a) found that between 2000 and 2008, approximately 80% of the forest loss in Kalimantan and Sumatra occurred in the zones where land clearing is allowed, while approximately 20% of the forest loss occurred in the zones where land clearing is restricted or even prohibited. Additionally, discrepancies were found between oil palm, timber and logging concessions and the designated land allocation zones (Chapter 3). For example in West Kutai and Mahakam Ulu districts, approximately 21% of the oil palm concessions in 2009 were located in the zones where oil palm cultivation is prohibited (Chapter 3). Recent developments in land allocation zoning policies make the reclassification from all zones of the Forest area to Non-forest area possible, given certain criteria, e.g. that a minimum of 30% of province land remains Forest area (Minister of Forestry Instruction 33/2010; 34/2010; 29/2014). As a result of these conversion regulations, larger areas may become available for the development of cash-crop plantations, and thus more frequent and large-scale land development becomes a possible scenario. To date, it is unclear how these land zoning regulations will impact the trajectories of LULC change and forest cover, particularly in view of the growing demand for cash crop products. This study therefore aims to explore the impacts of the land allocation zoning policies and of different levels of land development on LULC, and particularly on forest cover and local food production. Because land use regulations and spatial planning tend to change continuously and sometimes rapidly, any prediction of future LULC as a result of agricultural development and other land use changes is surrounded by large uncertainty (Müller et al., 2014). What land use change models can contribute, however, are assessments of given ‘what-if ’ scenarios, i.e. providing a projection instead of a prediction. Therefore, we develop four contrasting what-if scenarios with i) limited and unlimited land development of the main land uses in the study area, and ii) restricted and unrestricted application of the land zoning policies. We selected the West Kutai and Mahakam Ulu districts in the East Kalimantan province as the study area because this region is experiencing a transformation from small-scale mixed land uses to large-scale monoculture plantations and ongoing deforestation and forest degradation (Inoue et al., 2013; Müller et al., 2014; Chapter 3). The impacts of LULC change on forest cover and local food production under the four scenarios are modelled and visualised between 2009 and 2030 with the spatio-temporal PCRaster Land Use Change (PLUC) model (Verstegen et al., 2012). We selected the PLUC model as it can integrate uncertainties of the LULC change processes in dynamic simulations, thus enabling end-users to account for the uncertainties that are inherent to spatial planning (Verstegen et al., 2012). This is essential to LULC change modelling in a region with such dynamic LULC change processes as the West Kutai and Mahakam Ulu districts. In section 2, we elaborate on the land allocation zoning policies in Indonesia. In section 3, we describe the materials and methods, and in section 4 we present the results, followed by the discussion and conclusions in section 5.

45

4.2 Theory 4.2.1 Land allocation zoning in Indonesia The Indonesian Constitution (Article 33(3), 1945) states that land, water and natural resources are controlled by the state and must be used for the greatest prosperity of the people. Although the state does not have the property rights of the land, it allows the government to manage virtually everything related to land. Land in Indonesia is divided into Forest area (Kawasan Hutan) and Non-forest area (Kawasan Budidaya Non Kehutanan: KBNK, formerly called Areal Penggunaan Lain: APL) (see, for instance, Minister of Forestry (MoF) Decision SK.352/Menhut-II/2004, Article 1). Forest area is under the jurisdiction of the Ministry of Environment and Forestry, following the Forestry Law no. 41/1999. Private property rights do not exist for Forest area2F, but the Ministry of Environment and Forestry can give out a variety of usage rights to parties. Under MoF Instruction 34/2010, eight land allocation zones are classified, yet in this study, we follow the Forest area categorisation used by Broich et al. (2011a), who used the classification of existing MoF maps. This categorisation distinguishes five zones ranging from Conservation forest to Production forest (Broich et al., 2011a) and is a simpler version of the zones distinguished in MoF Instruction 34/2010. Formally, only forest-related activities such as forest conservation or logging are allowed, and only forest commodities listed by the Ministry of Environment and Forestry can be cultivated in Forest area. Small-scale traditional activities such as the extraction or cultivation of Non-Timber Forest Products (NTFPs), ecotourism and education related to the forest also allowed in Forest area zones. The definitions according to Broich et al. (2011a) following Ministry of Forestry Indonesia (2008) for the five zones are: • The Conservation forest zone (Kawasan Suaka Alam dan Pelestarian Alam: KSPA) is defined as a forest with the principal function of preserving the ecosystems and biodiversity. • The (Watershed) Protection forest zone (Hutan Lindung: HL) has the principal function of protection of life support systems to regulate water flow, preventing floods, controlling erosion, preventing sea water intrusion and maintaining soil fertility. • The Production forest zone (Hutan Produksi: HP) has a principal function of producing forest products, including pulpwood, as defined by the Ministry of Environment and Forestry. • The Limited production forest zone (Hutan Produksi Terbatas: HPT) is the area with a limited production function and with a specific range of slopes, soil types, and rainfall intensity, outside the Protected forest area, forest preserves, nature conservation forest and hunting parks. • The Production forest for Conversion zone (Hutan Produksi yang dapat di-Konversi: HPK) is a Forest area zone that is reserved to be converted to Non-forest area. Non-forest area is managed by the regional, provincial and national government according to the Basic Agrarian Law (Law no. 5/1960). Non-forest area can be held as private property by Indonesian citizens and by companies through right of use (Hak Guna Usaha). 4.2.2

Development in and reclassification of land allocation zones

Settlements, agriculture, forestry and mining In principle, the development of mixed cropland and estate plantations such as oil palm and rubber is only allowed in the Non-forest area (Broich et al., 2011a). Mixed cropland can be developed in Forest area, but only under Adat (customary rights) and population usage rights. Forestry activities such as Hutan Tanaman Industri (HTI) pulpwood plantations are allowed in the Production forest zone. In theory, mining can be conducted in all zones. Settlements cannot be established in the Conservation 46

forest or the Protection forest zone, and are usually not established in the Production forest or the Limited production forest zone. Table A 12 in Appendix 3 shows the areas of the different LULC types in 2009 for each land allocation zone.

Change of Forest area zone or use The head of the district or municipality can propose to reclassify land allocation zones within their jurisdiction from Forest area to Non-forest area under specific conditions and requirements, while the provincial Governor can do this for border areas between districts and municipalities (MoF Instruction 34/2010, Article 10). In principle, all zones of the Forest area can be reclassified into the Production Forest for Conversion zone and further to Non-forest area (MoF Instruction 34/2010, Article). Figure 4.1 provides a simplified representation of possible reclassifications from Forest area zones to Non-Forest area, based on this Ministerial Instruction. To assess the impacts of these land allocation zoning policies on LULC, we made the following assumptions. The status of the Production forest for Conversion zone in the Forest area can be reclassified to Non-forest area (Step c in Figure 4.1). Additionally, the status of the Production forest zone and Limited production forest zone can be reclassified into the Production Forest for Conversion zone

Restricted land zoning: zones remain unchanged Forest area Conservation forest

Protection forest

a

Production forest Limited production forest

b Unrestricted land zoning: zones can be reclassified

Production forest for Conversion

Non-forest area

9002

c

Figure 4.1. Simplified schematic representation of the land allocation zones and the possible reclassification of Forest area zones to Non-forest area as assumed under the scenarios based on the MoF Instruction 34/2010 (Article 2 and Article 4). Under the restricted zoning scenarios, the land allocation zones remain like they were in 2009 (see the left grey column and Figure A 3). Under the unrestricted zoning scenarios, reclassification from Forest area to Non-Forest area is possible as indicated by steps b and c in the figure. We assume that the reclassifications in Step a will not occur, because of its complexity due to the compliance to extensive legal conditions and requirements.

47

and then further to Non-forest area (Steps b and c in Figure 4.1). In this paper, we assume that the reclassification from or to the Conservation forest zone and Protection forest (MoF Instruction 34/2010, Article 7 and 6; Step a in Figure 4.1 and vice-versa) will not occur, because of its complexity due to the compliance to extensive legal conditions and requirements. As a result of these legislations towards unrestricted land zoning, theoretically, a large land area can be converted into Non-forest area. An important criterion is that reclassification to Non-Forest area is possible only if the total Forest area in the province, whether or not forested, remains at least 30% of that province’s total land area (MoF Instruction 34/2010, Article 3). In this article, the application of the land allocation zoning policies for the reclassification of Forest area zones to Non-Forest area (MoF Instruction 34/2010) is called unrestricted land zoning. Under restricted land zoning, we assume that reclassification under the zoning policies is not applied and the land allocation zones remain as they were in 2009.

4.3

Material and methods

4.3.1 Case study West Kutai and Mahakam Ulu districts are located in the western part of the province of East Kalimantan and have a total land area of 3.3 million hectares. Figure A 2 (in Appendix) presents the LULC of the study area in 1990, 2000 and 2009 and Figure A 3 shows a map of the land allocation zones in the study area in 2009. The region has been subject to important political and administrative changes, including the division of the former West Kutai district into these two districts in 2012. From the 1980s onward, the region experienced large-scale forest degradation and deforestation due to logging and forest fires, and due to the strong transformation from shifting cultivation as the dominant land use toward the intensive cultivation of cash crops (Inoue et al., 2013; Müller et al., 2014; see also Chapter 3). Until 1995, the degradation and loss of forest was mostly caused by selective logging, but after 1995, the government focused more on the development of cash-crop plantations, including rubber, oil palm and pulpwood (Müller et al., 2014). Earlier studies showed that LULC change in the study area follows characteristic trajectories, for example, in one-stage trajectories from forest to small-scale mixed land uses or from forest to monoculture plantations, and in multi-stage trajectories from forest to small-scale mixed land uses and further to monoculture plantations (Inoue et al., 2013; see also Chapter 3). Rubber and oil palm are currently the most important land uses in the study region in terms of expansion. Oil palm is mostly cultivated by companies as large-scale monoculture plantations. In contrast, rubber is mostly produced by smallholders in monoculture plantations of 4-6 ha, and, to a lower extent, in agroforestry systems where rubber trees occur mixed with secondary vegetation. Mixed cropland refers to paddy fields, shifting cultivation, seasonal crops and orchards. Although land clearing and forest degradation through logging and forest fires are important LULC change processes in the region (Hoffmann et al., 1999; NASA, 2016; Siegert et al., 2001; see also Chapter 3), we aimed to focus solely on the expansion of agriculture, forestry and mining, since we assume that the demands for these land uses and land zoning largely determine the LULC change trajectories and displacement. We modelled the development of the following land use types: mixed cropland, rubber (mostly smallholder), oil palm, pulpwood, mining (mostly coal) and settlements, since these have expanded rapidly since the 1990s in the study area, driven by the increased demands for the commodities related to these land uses (BPS, 2014, 2009; Budiman et al., 2014; see also Chapter 3). In this paper, we further define these land uses as active land use types (or shortly: land use types) since these expand

48

based according to the increase of future demands. We distinguished three forest types with varying degradation levels: closed, medium open and very open canopy forest. The forest types and shrubs, grasslands, cleared lands and gold mining were considered as passive land use types since these remain unchanged or decrease in area when these are replaced by an active land use type.

4.3.2 Policy-derived scenarios We defined four contrasting what-if policy-derived scenarios to model the effects of the land allocation zoning policies and different levels of land-use demand on future LULC and forest cover in the West Kutai and Mahakam Ulu districts. Our projections are not predictions of future LULC, but rather simulations under the defined what-if scenarios. The scenarios were built as a quadrant (Figure 4.2), with two dimensions: the projected land allocation zoning policies and the projected level of land-use demand of the land use types. The moratorium on concessions on primary forests and peatlands is implicitly included in these scenarios, and is in line with limited development.

limited restricted development

limited unrestricted development

unlimited restricted development

unlimited unrestricted development

Unrestricted land zoning

Restricted land zoning

Limited land development

Unlimited land development Figure 4.2. Scenario matrix with on the horizontal axis the land allocation zoning policy projections (i.e. assuming restricted or unrestricted land allocation zoning), and on the vertical axis land-use demand projections (i.e. assuming limited and unlimited land development).

Land allocation zoning dimension We defined two land allocation zoning scenarios, namely restricted and unrestricted zoning (see Table A 13). Under the restricted zoning scenarios, we assumed that development of the land uses is restricted to the land allocation zones in 2009 and that the potential reclassifications under MoF Instruction 34/2010 are not applied. Therefore, we assumed that all Forest area zones remain as Forest area in 2009 (Figure A 3). The Production forest for Conversion zone is not present in the study area. As a result, we assumed that in the restricted zoning scenarios, the development of agriculture is restricted to the Non-forest area in 2009 (Figure A 3, Table A 13). Additionally, we assumed that mining is restricted to the Non-forest area because of the negative impacts of and negative international attention for mining in forest areas (Abood et al., 2014). We also assumed that under 49

the restricted land zoning scenarios these land uses will not be undertaken in peatlands because of its unsuitability and high carbon emissions related to these forms of peatland use (Hooijer et al., 2010). Under the unrestricted zoning scenarios, on the other hand, we assumed that reclassifications under MoF Instruction 34/2010 are applied and that, consequently, all zones, except for the Protection forest zone and the Conservation zone, may be reclassified into Non-forest area. These zones thus become available for agriculture and mining (Figure A 3; Table A 13). Under the unrestricted zoning scenarios, land development can also be undertaken in peatland areas (Figure A 6).

Land development dimension For the land-use demands of the land use types, we constructed a limited development and an unlimited development projection. Between 1990 and 2009, the selected land use types have shown a nearly exponential growth in terms of land area (see Chapter 3). Therefore, for the unlimited development projections, we extrapolated the exponential fit for every land use type over 1990-2009, to 2009-2030, following Equation 6 in Appendix 6. The projected exponential-growth curves are shown for each of the selected land use types in Figure 4.3A. For the limited development projection, we assumed that the expansion of agriculture, plantations and mining will slow down from 2014 until 2030 due to the implementation of the moratorium extension of the Governor on mining, pulpwood and oil palm plantation permits on forested lands and peatland (Presidential Instruction 6/2013) (Hardjanto, 2015; Tempo.co, 2015). To this purpose, we assumed that the demand between 2009 and 2014 will follow the exponential-growth curve, while after 2014, the land-use demand will decrease with 50% annually, assuming that the number of permits handed out will increasingly slow down. This resulted in an S-shaped curve of which the equation (Equation 7) is included in Appendix 6. The projected limited development (i.e. S-shaped-growth) curves of the land use types are shown in Figure 4.3B. Integrating the two scenario dimensions resulted in the four scenarios as presented in Figure 4.2.

2,0 Oil palm plantation

Land area (Mha)

Rubber (mostly smallholder) 1,5

Pulpwood plantation Mining (mostly coal)

1,0

0,5

0,0 1990

Settlement Mixed cropland Maximum land area available for development

2000

2010

2020

2030 1990

2000

2010

2020

2030

Figure 4.3. Unlimited (exponential curve) (a) and limited (S-shaped) (b) land-use demand curves for the land use types, extrapolated between the observed LULC areas of 1990, 2000 and 2009 and further until 2030. The green line shows the maximum available land area as capped by the zoning restrictions under the unrestricted zoning scenarios.

50

The limited development projections resulted in a total land-use demand of the land use types of 0.38 Mha for 2020 and 2030 (Figure 4.3, Table A 15, Table A 16). Oil palm is the land use type with the largest expansion area under all scenarios, with a cover of approximately 115,000 ha towards 2030 in the limited development scenarios and 0.33-1.47 Mha in the unlimited development scenarios (Table A 16). The government designated a total oil palm area of 0.5 Mha for this region (Budiman and Smit, 2011), which falls within the range of the limited and unlimited development projections. Rubber and pulpwood plantations have the next largest projected land areas towards 2030, namely ~167,000333,000 ha for rubber and ~45,000-106,000 ha for pulpwood plantations (Table A 16). The unlimited development projections resulted in a much higher land-use demand, namely ~0.89 Mha under restricted zoning and ~2.14 Mha under unrestricted zoning in 2030 (Figure 4.3, Table A 15, Table A 16). 4.3.3 Exploring the scenarios with the PCRaster Land Use Change model We selected the PCRaster Land Use Change model (PLUC) to explore the LULC change scenarios in the study area. The PLUC model is implemented in the PCRaster Python framework that combines geospatial analysis, spatiotemporal modelling and Monte Carlo simulation to assess uncertainty (Karssenberg et al., 2010). The PLUC model simulates LULC change for a selected number of land use types as a function of a set of user-specified spatio-temporal suitability factors and no-go areas per land use type (see Figure A 7 in Appendix 8). This makes the model suitable for the modelling of the effects of land zoning policies. Additionally, PLUC is a demand-driven model, which means that it spatially allocates areas for the different land use types in the study area. This creates the possibility to allocate user-specified future land-use demands for the selected land use types. Details about the model are included in Appendix 8. In the first step of the analyses (step 1 in Figure 4.4), we parameterised the PLUC model for this study, based on the selected land use types (see Appendix 8). To this purpose, we defined and weighted a set of suitability factors, namely altitude, slope, distances to primary roads, secondary roads, rivers and towns, travel time to the nearest palm oil mill (only for oil palm), travel time to the nearest town, current LULC and the area of equivalent LULC type in the direct neighbourhood for the spatial allocation of the land use types (see Appendix 8). For the selection and weighing of the suitability factors, we conducted a forward (Wald) stepwise binary logistic regression in SPSS for each active land use type, namely for mixed cropland, rubber (mostly smallholder), oil palm, pulpwood, mining (mostly coal) and settlements. The suitability factors that were significantly correlated with conversion into a specific active land use type between 2000 and 2009 were included in the PLUC model. For a more detailed description and overview see the text and Table A 18 in Appendix 8. Additionally, we defined a set of no-go areas for each active land use type, which are areas where these land use types cannot be developed because of socio-economic or biophysical restrictions (see Table A 19 in Appendix 8). To evaluate the performance of the parameterised PLUC model, we conducted a model run with 100 Monte Carlo simulations between 2000 and 2009, using the LULC map of 2000 as the initial land use map, accounting for i) the observed 2000-2009 land-use demand of each active land use type, and ii) the unrestricted zoning projection 2000-2009. The evaluation was conducted by defining the standard error (SE) of the area of each active land use type in 10x10 km blocks that resulted from a comparison between the observed LULC in 2009 and the projected LULC in 2009 over the 100 Monte Carlo simulations (see further Appendix 9). The higher the standard error, the higher the error between the observed LULC in 2009 and the projected LULC in 2009 based on 100 Monte Carlo simulations and thus the higher the uncertainty of the projection in that particular location.

51

Step 1. Parameterisation of the PLUC model

Step 2. Run PLUC model

Step 3. Analyses and evaluation of PLUC output

Based on land use types

100 Monte Carlo runs

GIS analyses Statistical analyses

Temporal patterns LULC change

Spatial patterns LULC change

Based on scenarios • land-use demand • land allocation zoning

Based on spatial allocation of land use • suitability factors • no-go areas

Figure 4.4. Analytical framework showing the three steps in the analyses. First, we parameterised the PLUC model for the study area, followed by an evaluation of the amodel; second, we ran the model under the four scenarios; third, we analysed the output.

In the second step (step 2 in Figure 4.4), we projected the LULC for 2009-2030 in time steps of one year in the study area under the four scenarios. Because PLUC generates annual LULC maps, this enables the analysis of the LULC change trajectories over time as a result of the specified land allocation zoning policies and levels of land-use demand. As pixel-by-pixel attributions of LULC change are highly uncertain, we calculated the probability that a land use type would expand at a certain location by conducting a Monte Carlo analysis of 100 model realisations for each of the four scenarios (Figure 4.4), in which a random noise map is added to the list of suitability factors for each land use type as the stochastic variable (Appendix 8). In the third step (step 3 in Figure 4.4), we analysed the model output. First, we quantified and mapped the land area of all LULC types that were projected for 2020/2030 under the four scenarios and compared these to the LULC in 1990, 2000 and 2009. Second, we quantified the occurrence of the LULC change trajectories showing the displacement of land use types and their relocation elsewhere. To this purpose, we calculated the number and type of conversions of each raster cell that have been projected between 2009 and 2030, thereby deriving LULC change trajectories of one, two, three or more stages. Finally, we calculated and mapped the probability of occurrence of each land use type in a certain area under each of the scenarios (see further Appendix 8).

4.4 Results 4.4.1 Evaluation of uncertainties in the PLUC outcomes In Figure A 8 (in Appendix), the standard error (SE) of the area of each active land use type in 10x10 km blocks is shown. Oil palm shows the largest average standard error (SE = 9.4) across the landscape and thus the highest variation in its spatial allocation. Pulpwood plantations show the second-largest average standard error (SE = 8.3), followed by smallholder rubber (SE = 6.2) and mixed cropland (SE =

52

3.5

LR

LuR

3.0

Land area (Mha)

2.5 2.0 1.5 1.0

0

9002

0.5

1990

2000

2009

2020

3.5

2030

1990

2000

2009

2020

uLR

2030

uLuR

3.0

Land area (Mha)

2.5 2.0 1.5 1.0 0.5 0

1990

2000

2009

2020

2030

1990

2000

2009

2020

2030

Settlement

Mixed cropland

Mining (mostly coal)

Shrubs/grassland

Oil palm plantation

Forest - very open

Rubber (mostly smallholder)

Forest - medium open

Pulpwood plantation

Forest - closed

Figure 4.5. Observed LULC area (ha) in 1990, 2000 and 2009 (based on the LULC maps adapted from Budiman et al., 2014; see further Chapter 3) and projected LULC area for 2020 and 2030 by the PLUC model under the four scenarios (LR, limited restricted; LuR, limited unrestricted; uLR, unlimited restricted; and uLuR, unlimited unrestricted.).

4.4). The simulated spatial expansion of oil palm and pulpwood plantations over space is thus the most uncertain.

53

4.4.2 Temporal patterns of LULC change Figure 4.5 shows the LULC change between 1990 and 2009 and the LULC change projected until 2020 and 2030 under all scenarios (see also Table A 15, Table A 16 and Table A 20). With limited land development, the declines in total forest land between 2009 and 2030 under restricted and unrestricted zoning are similar (~ 0.1 Mha, ~4%; Table A 20). With unlimited land development of the main land uses between 2009 and 2030, however, total forest land declined substantially stronger under the unrestricted zoning scenario (approx. 60%), compared to the restricted zoning scenario (approx. 17%). The application of the land zoning policies that enable the reclassification of the Production forest zone and the Limited production forest zone, where most of the forest in 2009 was present (Table A 12), to Non-forest area can thus result in substantial additional forest loss under a scenario of unlimited land development (Figure 4.5). With limited land development, the largest declines in the period 2009-2020-2030 were found in very open canopy forest (approximately 18%) and shrub/grasslands (approximately 13%). With unlimited development, however, we found a larger decline in closed and medium open forest cover, namely 5-6% and 15-21%, respectively, in 2020, and 5-34% and 17-75%, respectively, in 2030. Figure 4.6 shows the 15 most frequently projected LULC change trajectories for the study area as expressed in the land area (ha) in which these occurred between 2009 and 2030 under each scenario. For simplicity reasons, we grouped all forest types. The one-stage, two-stage and three-stage trajectories are indicated by two, three and four circles, respectively. Our projections show that most of the LULC change between 2009 and 2030 occurred in one-stage trajectories (Figure 4.6A), which means that conversions occurred mostly from only one LULC type to another. The main one-stage conversions between 2009 and 2030 are those from forest and shrub/grasslands to oil palm (~60,000195,000 ha) or to rubber (~49,000-98,000 ha). The area in which multi-stage trajectories occurred is about 10% of the total area where LULC change occurred (Figure 4.6), indicating modest displacement of land use. Unlimited development resulted in a larger total area with multi-stage trajectories than limited development (Figure 4.6B). Restricted zoning resulted in a larger area with multi-stage trajectories than unrestricted zoning. Furthermore, with restricted zoning, mostly smallholder rubber and mixed cropland are being displaced, while with the unrestricted zoning, mostly smallholder rubber and pulpwood plantations are being displaced during these multi-stage trajectories (Figure 4.6B). 4.4.3 Spatial patterns of LULC change The unlimited development projections in Figure 4.7 show that land development starts in 2020 in the lowlands in the south-east, and over time expands further to the higher altitudes in the north-west. The expansion of LULC change under the unlimited-unrestricted scenario into the higher altitudes results in the decline of more medium open and closed canopy forest compared to the unlimited-

Figure 4.6. (see next page) Top 10 of most occurring LULC change trajectories (uppermost four panels) and top 10 of most occurring multi-stage trajectories (bottom four panels) under the four scenarios between 2009 and 2030 (expressed in the mean total area covered by the trajectories over the 100 Monte Carlo simulations). The scenarios include: LR, limited restricted; LuR, limited unrestricted; uLR, unlimited restricted; and uLuR, unlimited unrestricted. Each circle constitutes a LULC state. A trajectory of two, three or four circles implies that the initial LULC has changed one, two or three times in the period 20092030, respectively.

54

Top 10 all trajectories Total area trajectory (ha)

LR

LuR

52,476

49,055

30,130

41,636

29,476

28,042

15,044

9,007

9.746

8,502

7,703

7,825

7,610

7,001

4,518

5,881

4,507

1,766

4,273

1,649

Total area trajectory (ha)

uLR

uLuR

84,805

121,538

82,024

90,386

71,488

73,844

39,803

26,604

27,691

17,965

22,993

16,668

17,658

14,161

17,533

11,425

14,220

10,853

8,886

7,691 Forest (closed, medium and open)

Pulpwood plantation

Shrubs/grassland

Oil palm plantation

Mixed cropland


Rubber (mostly smallholder)

Settlement

Top 10 multi-stage trajectories LuR

4,518

1,649

494

212

102

53

93

48

77

30

11

10

8

9

5

7

3

3

2

3


uLR

uLuR

27,691

11,425

5,815

1,306

700

655

421

218

399

177

289

167

261

141

132

79

68

65

29

50

9002


LR

55

2009

0

100 km

LR

2030

2020

uLR

2030

9002

2020

Figure 4.7. Land use and land cover of one Monte Carlo projection for 2020 and 2030 under the four scenarios, compared to the actual LULC map in 2009 (top). The scenarios are: LR, limited restricted; LuR, limited unrestricted; uLR, unlimited restricted; and uLuR, unlimited unrestricted (1 raster cell is 250 x 250 m).

56

Forest - Closed Forest - Medium open Forest - Very open Shrub/grassland Mixed cropland Rubber (mostly smallholder) Pulpwood plantation Oil palm plantation Mining (mostly coal) Settlement Water

2020

LuR

2030

2020

uLuR

2030

57

Mixed cropland

A

0

LR

LuR

uLR

uLuR

LR

LuR

uLR

uLuR

LR

LuR

uLR

uLuR

100 km

Rubber (smallholder)

B

1: High

Figure 4.8. Probability maps for the occurrence of A) mixed cropland, B) rubber (mostly smallholder), and C) oil palm, in the study region between 2009 and 2030 under the four scenarios, namely LR, limited restricted; LuR, limited unrestricted; uLR, unlimited restricted; and uLuR, unlimited unrestricted. Black raster cells indicate a high probability (value = 1) and white raster cells a low probability (value = 0) (1 raster cell is 250 x 250 m).

58

Oil palm

C

9002

0: Low

restricted scenario in 2030. The unrestricted zoning scenario simulations show more scattered LULC change, compared to the restricted-scenario simulations (Figure 4.7). The probability maps of mixed cropland, smallholder rubber and oil palm presence are shown in Figure 4.8 for the four scenarios. Under the limited development scenarios, mixed cropland expands mainly in the centre of the region, which is the lowland-midland area (Figure 4.8A). In the unlimitedrestricted scenario, mixed cropland develops scattered throughout the study region, while in the unlimited-unrestricted scenario it develops only in the higher altitudes. Figure 4.8B shows that under the limited development scenarios, smallholder rubber expands in the lowlands in the south-east and in the highlands in the north-west. Under the unlimited development scenarios, however, smallholder rubber expands throughout the whole landscape. Oil palm development is limited to the lowlands in the south-east under the limited development scenarios, but also expands in the north-west with unlimited development (Figure 4.8C).

4.5

Discussion and conclusions

This study aimed to assess the effects of the land allocation zoning policies and different levels of land development on LULC and particularly on forest cover and local food production in Kalimantan. The four scenarios in our study showed strongly varying effects on forest cover and food production in terms of magnitude of LULC change, the location of LULC change in the landscape and the type of LULC change trajectories and displacement. According to our findings, most of the LULC change between 2009 and 2030 occurs at the cost of forests and shrub/grasslands under all scenarios. With limited development of the active land use types, forest loss between 2009 and 2030 is low (~ 0.1 Mha, ~4%) compared to unlimited development. These active land use types remain mostly in the lowlands in the south-east, and the influence of land zoning is minimal in 2020 and increases slightly up to 2030. With unlimited development, however, the effect on LULC and forest cover is much stronger (~0.41.6 Mha in 2009-2030) and the land zoning policies can make a large difference. Namely, under the unlimited-restricted scenario, total forest cover declines with 0.4 Mha (~17% in 2009-2030). Under the unlimited-unrestricted scenario, however, the decline in forest cover is much stronger with 1.6 Mha (~60% in 2009-2030), and with a stronger decline of medium open and closed canopy forests in the higher altitudes. Additionally, under the unlimited development scenarios, land development in 2009 starts in the lowlands in the south-east and over time expands further into the higher altitudes in the north-west. Meanwhile, mixed cropland and thus food production in the lowlands are displaced and relocated to the higher altitudes where it results in forest loss. Lowland areas are generally, and were assumed in the model as, more suitable and accessible for agriculture and mining and are thus more susceptible to LULC change than the highlands (Fuller et al., 2004). The relocation of food production to the higher altitudes is not desirable given the remoteness of particularly the midlands and mountains areas in this region (Obidzinski et al., 2012). The difference in terms of the LULC and forest cover change area between the limited and unlimited development scenarios is larger than the difference between the restricted and unrestricted zoning scenarios. Therefore, we conclude that the level of land development has a larger impact on LULC and forest cover than land zoning restrictions. The relatively small area in which multi-stage trajectories occurred compared to the total area of LULC change (~10% in 2009-2030) shows that displacement represents a modest share of total LULC change. With restricted land zoning, we found more displacement in terms of land area than with unrestricted land zoning. Additionally, under restricted zoning, the multi-stage trajectories showed

59

more displacement of smallholder rubber and mixed cropland, while under unrestricted zoning, we found more displacement of smallholder rubber and pulpwood plantations. Based on our input data, we assume that this is caused by the relatively lower availability of suitable land for agriculture under the restricted zoning scenarios, which causes the continuous displacement of these intermediate land use types. Whether LULC change and displacement is wanted or unwanted depends on the actual or projected socio-economic and ecological impacts and how this influences the stakeholders involved (Obidzinski et al., 2012). The decrease in forest cover can have negative impacts on ecosystem services and biodiversity (Murray et al., 2015; Paoli et al., 2010; Slik et al., 2008). The decrease or relocation of local food production can impact the accessibility to food, particularly in remote rural areas (Obidzinski et al., 2012). In our study, we assumed that mixed cropland, or the cultivation of food crops, in the region would increase, following historical trends. It can be expected though that the development of monoculture plantations results in the displacement of food production either inside or outside of the region, where it can cause further LULC change and forest loss (Inoue et al., 2013; Müller et al., 2014). We applied the PLUC model as a tool to project LULC change under ‘what-if ’ land-use demand and policy-derived scenarios that exist of predefined parameters, relationships and system boundaries. Therefore, our projections are not predictions of future LULC, but instead are simulations under policy-derived scenarios (van der Hilst et al., 2012; Verstegen et al., 2012). Our unlimited development projections of the land uses, particularly of oil palm and rubber, show a very steep curve and are thus rather extreme, but not impossible, for this timescale. A continued strong expansion of the selected land use types can be expected, since the rubber, mining, palm oil, and pulp and paper industries are an important part of the regional economy (Budiman and Smit, 2011). Additionally, the region has strategic plans to expand oil palm plantations (Budiman and Smit, 2011), a long history of mixed cropland and rubber production (Inoue et al., 2013), and large coal reserves. We assumed a strong increase of pulpwood plantations under the unlimited development scenarios, because of the plantation potential in the study region, the growing global demand for wood and the goals of Indonesia to revive the country’s pulp and paper industry and to increase its wood supply (Obidzinski and Chaudhury, 2009; Obidzinski and Dermawan, 2012). We based the LULC change modelling on suitability factors and their weights derived from statistical analyses of historical data. We kept these spatial allocation variables constant over time. In reality, however, the spatial allocation of agriculture and mining is influenced by all kinds of drivers, policies and investments that are under continuous change (Müller et al., 2014). For example, land zoning policies affect the environmental system and these environmental impacts can in turn motivate a decision to change land use policies (Ostrom et al., 2007). Such recurrent effects are not accounted for in the study. Particularly in a region with such a dynamic political and socio-economic history, this creates uncertainties in the projections over a long model run time (Veldkamp and Verburg, 2004; Verstegen et al., 2016). Our results showed that the simulated spatial expansion of oil palm and pulpwood plantations over space is on average the most uncertain, given the relatively high average standard errors compared to the other lands use types. Including socio-economic, political and biophysical factors that drive LULC change could have improved the validity of the projections (Veldkamp and Lambin, 2001), such as policy changes, demand and price fluctuations of commodities, yield optimisation or soil type. We also recommend to incorporate a method to allow for systemic changes in the model structure (Verstegen et al., 2016). With unlimited development of the main land uses, independent on the land zoning policies, future claims on land for forest maintenance cannot be reconciled with food production, mining and

60

the development of cash crops. We expect an increasing pressure over time on the closed canopy, often primary, forests, which are mostly located in the higher altitudes. The loss of forests and the relocation of food production that will result from this, is expected to have negative impacts on local food production, ecosystem services and biodiversity. We recommend further research on these subsequent impacts and how this affects local stakeholder benefits from ecosystem services and access to food. Our findings thus show that in order to maintain primary forests, peatlands and local food production, the production of palm oil, rubber, pulpwood and coal needs to be slowed down and needs to take place under restricted land zoning and responsible spatial planning. This includes the extension and the strengthening of the current moratorium, if reconciliation of all different land functions is sought. However, the implementation of the moratorium can only be effective if reclassification of Forest area to Non-forest area is no longer allowed or strongly discouraged in forested lands or peatlands, and is additionally based on land availability and suitability for the selected crop. This underlines the importance of harmonising existing forest laws and regulation with the objectives of the moratorium. Additionally, measures need to be undertaken to improve yields and chain efficiencies of the main commodities under restricted zoning (Van der Laan et al., 2016; see further Chapter 7). Conservation should also focus on medium open and very open canopy forests, because of their large regeneration potential. With limited development, sufficient grass and shrublands are present for land development, however, the application of a robust method such as the Responsible Cultivation Method (Smit et al., 2013) and thorough field validation need to define whether these lands are underutilised for responsible land development.

61

62

Part II. Impacts of land development and forest loss on carbon stocks and biodiversity

63

64

5 Analysis of biophysical and anthropogenic variables and their relation to the regional spatial variation of aboveground biomass Illustrated for North and East Kalimantan, Borneo Carina van der Laan, Pita A. Verweij, Marcela J. Quiñones, André P.C. Faaij This chapter has been published in the journal Carbon Balance and Management

Abstract Land use and land cover change occurring in tropical forest landscapes contributes substantially to carbon emissions. Better insights into the spatial variation of aboveground biomass is therefore needed. By means of multiple statistical tests, including geographically weighted regression, we analysed the effects of eight variables on the regional spatial variation of aboveground biomass. North and East Kalimantan were selected as the case study region; the third largest carbon emitting Indonesian provinces. Strong positive relationships were found between aboveground biomass and the tested variables; altitude, slope, land allocation zoning, soil type, and distance to the nearest fire, road, river and city. Furthermore, the results suggest that the regional spatial variation of aboveground biomass can be largely attributed to altitude, distance to nearest fire and land allocation zoning. Our study showed that in this landscape, aboveground biomass could not be explained by one single variable; the variables were interrelated, with altitude as the dominant variable. Spatial analyses should therefore integrate a variety of biophysical and anthropogenic variables to provide a better understanding of spatial variation in aboveground biomass. Efforts to minimise carbon emissions should incorporate the identified factors, by 1) the maintenance of lands with high AGB or carbon stocks, namely in the identified zones at the higher altitudes; and 2) regeneration or sustainable utilisation of lands with low AGB or carbon stocks, dependent on the regeneration capacity of the vegetation. Low aboveground biomass densities can be found in the lowlands in burned areas, and in Non-forest area and Production forests.

65

5.1 Introduction More insights into the spatial variation of aboveground biomass (AGB) are crucial to minimise carbon emissions and global climate change from tropical deforestation, forest degradation and agricultural expansion. According to van der Werf et al. (2009), globally, approx. 12% of anthropogenic carbon emissions in 2008 were caused by deforestation and forest degradation (van der Werf et al., 2009). During the period 1973 – 2010, Kalimantan, Indonesian Borneo, has lost ~31% of the total forest area (Gaveau et al., 2014). With regard to land use changes, according to the Governors’ Climate and Forests Task Force Indonesia (GCF-TF Indonesia, 2011), the recently merged provinces North and East Kalimantan are when combined the third largest carbon emitting provinces in Indonesia, with 255 Mt CO2e yr-1, after Central Kalimantan (324 Mt CO2e yr-1) and Riau (258 Mt CO2e yr-1). According to their ‘business as usual’ scenarios, land use change will cause carbon emissions in North and East Kalimantan to increase to 331 Mt CO2e yr-1 by 2030 (GCF-TF Indonesia, 2011). Mechanisms such as Reducing Emissions from Deforestation and forest Degradation + (REDD+) (UN-REDD, 2013) have been developed to halt such emissions by maintaining lands with high carbon stocks contained in living forest biomass, such as secondary and undisturbed forests. Meanwhile, expansion of low carbon stock agricultural lands can be instead shifted towards areas with already low carbon stocks or AGB, such as abandoned agricultural or restored degraded lands (Wicke et al., 2011) by implementation of sustainable land zoning tools (Smit et al., 2013). AGB is not static, but rather spatially and temporally highly variable, particularly in the tropics (Baker et al., 2004; Chave et al., 2003, 2001; de Castilho et al., 2006; Houghton, 2005). This makes its quantification and the avoidance of high AGB densities or high carbon stocks challenging (it is generally assumed that about half of AGB consists of carbon). As in other tropical forest landscapes, complex matrices of low to high AGB densities can be expected in North and East Kalimantan, due to varying biophysical conditions present, such as terrain and soil types, and anthropogenic disturbances such as fire or logging. For example, forest fires can cause substantial losses in carbon by the emission of large quantities of CO2 by the burning of biomass (Page et al., 2002; Slik et al., 2008), and via logging by the extraction of timber (Berry et al., 2010; Kronseder et al., 2012). However, after a fire or logging activities, regeneration can occur, resulting in an increasing sensitivity of the remaining live and dead vegetation to subsequent disturbance events (Matricardi et al., 2010; Siegert et al., 2001; Slik et al., 2008, 2002; Toma et al., 2005). Additionally, the type and severity of the disturbance and local biophysical conditions, such as altitude, soil type and the presence of pioneer species, influence the carbon accumulation potential (Hashimotio et al., 2000; Toma et al., 2005). Therefore, in this paper we address the question of how such biophysical and anthropogenic variables are related to AGB, and contribute to the spatial variation of AGB in a disturbed tropical forest landscape. AGB can be estimated at forest stand to landscape scale by plot-based measurements (Brown, 1997; Feldpausch et al., 2012). Several existing plot-based studies in tropical forest landscapes have statistically analysed the relationships between AGB and multiple biophysical variables including soil factors (Chave et al., 2001; de Castilho et al., 2006; Laurance et al., 1999; Paoli et al., 2008), altitude (Alves et al., 2010; Asner et al., 2008; de Castilho et al., 2006; Whittaker and Bormann, 1974) and slope (de Castilho et al., 2006; Ferry et al., 2010). Anthropogenic variables, however, are usually not considered, while specifically in highly disturbed tropical areas like North and East Kalimantan, these factors are expected to strongly affect AGB. Additionally, anthropogenic variables are important and useful to support the management of, and decision-making on, maintaining carbon stocks in disturbed areas. For these reasons, our analyses include both biophysical and anthropogenic variables.

66

Other plot-based studies have focused on the impacts of e.g. logging (Berry et al., 2010; Kronseder et al., 2012) and fire (Slik et al., 2008; Van Nieuwstadt and Sheil, 2005), by comparing AGB between undisturbed and disturbed land classes. The relationship between forest cover change and anthropogenic variables has also been analysed (Broich et al., 2011a; Dennis and Colfer, 2006; Siegert et al., 2001). These studies have instead focused on discrete land use and forest classes and therefore have not accounted for local scale AGB variation. Furthermore, the reviewed studies were not spatially explicit or conducted over larger spatial scales, thereby limiting a landscape scale view on the factors that influence the spatial variation in AGB or forest cover. A variety of spatially explicit data and methods exist to map and monitor land with high and low AGB or carbon over large spatial scales, such as extrapolating plot-based field AGB estimates to vegetation types with remotely-sensed reflectance data and spatial data of biophysical variables (Brown et al., 1993; Gibbs et al., 2007). For example, optical data can be used for mapping forest cover, such as Landsat (Yang et al., 2011). However, in areas with frequent cloud cover such as the tropics, radar technologies such as ALOS (Advanced Land Observing Satellite) PALSAR (Phase Arrayed L-band SAR) are more suitable (Mitchard et al., 2011b; Morel et al., 2011). Additionally, the integration of optical and/or radar technologies, including LiDAR (Light Detection And Ranging), has the potential to improve AGB estimates because it may reduce data saturation and mixed pixel problems (Englhart et al., 2011; Mitchard et al., 2011b; Quiñones et al., 2011; Saatchi et al., 2011). Although the output maps of the aforementioned studies have visualised the spatial distribution of AGB at high resolutions and over large spatial scales, these did not include the effects of biophysical or anthropogenic factors on AGB. Changes in AGB or carbon stocks have also been modelled at different spatial and temporal scales and resolutions (Carlson et al., 2012b; DeFries et al., 2002; Gibbs et al., 2007; Houghton et al., 2000; Morel et al., 2012). Additionally, studies using spatial data for AGB have compared AGB between forest types with different levels of degradation or disturbances, e.g. by logging or fire (Asner et al., 2010; Langner et al., 2012; Mitchard et al., 2011a; Morel et al., 2011; Saatchi et al., 2007). However, the focus was mostly on a single anthropogenic variable, e.g. logging or fire, and interrelationships between or interaction effects amongst variables were not investigated. The aforementioned studies are useful for the mapping and monitoring of AGB and carbon stocks, for e.g. REDD+ mechanisms. To monitor and quantify AGB whilst taking into consideration the high spatial variation, and additionally to enable the modelling of carbon stocks, further analysis of the underlying biophysical and anthropogenic conditions and processes, using a multi-variable approach, is essential. An improved level of information quality, that considered a broader set of variables and their interactions, would allow decision-making to focus on manageable factors in support of land use allocation that minimises carbon emissions and maximises carbon uptake in support of climate change mitigation. The aim of this study is to define which of a preselected set of biophysical and anthropogenic variables contribute significantly to the spatial variation of AGB. To this end, statistical analyses were conducted, including analysis of variance (ANOVA), non-spatial multiple linear regression and spatial geographically weighted regression (GWR). An AGB map based on radar remote sensing data and plot-based measurements were utilised, plus landscape scale data on terrain, soil types, land allocation zoning, fires, roads, rivers and cities, covering North and East Kalimantan, Indonesian Borneo. The results are shown in the Results section and can support the quantification and maintenance of living AGB and carbon stocks. In the Discussion section, the results are discussed in terms of their scientific and societal contribution, followed by the Conclusions and an extensive description of the data and analyses in the Methods section.

67

5.2 Methods 5.2.1 Study area The natural resource-rich provinces of North and East Kalimantan (North Kalimantan was established on 25 October 2012 and was previously part of East Kalimantan) have a high spatial variation in biophysical and anthropogenic conditions and processes. For use in this study, the provinces are regarded as one case study region. Henceforth, these provinces are indicated in this chapter as NorthEast Kalimantan. The terrain consists of undulating slopes and altitudes up to about 2,200 m. Karst and peatlands occur mainly in the lowlands (respectively, ~2% and ~4). The remaining landscape consists mainly of volcanic soils and other soil types (respectively, ~7% and 87%). This landscape is highly dynamic with regard to its past, current and expected land use changes. Until the early 1970s, the original land cover in the lowlands of North-East Kalimantan consisted of extensive dipterocarp forests with high AGB and species richness (Toma et al., 2005), but driven by forest and land development policies in the 1980s, large-scale degradation, deforestation and conversion to agricultural land have taken place (Murdiyarso and Adiningsih, 2006). The main activities were high intensity logging (Murdiyarso and Adiningsih, 2006), but also large-scale forest fires occurred that were often initiated for land clearing purposes (Murdiyarso and Adiningsih, 2006; Siegert and Hoffmann, 2000; Siegert et al., 2001; Slik et al., 2008), and events associated with El Niño Southern Oscillation (ENSO) (Priadjati, 2002). In 1997-98, again very destructive fires related to ENSO occurred, burning 5.2 million ha of North-East Kalimantan’s pristine and logged forests. Hoffmann et al. (1999) (Hoffmann et al., 1999) have found that approx. 75 % of the burned forests were allocated for logging, timber or oil palm concessions. The frequency and spatial extension of fires have increased over the last few decades in North-East Kalimantan because of deforestation and degradation processes associated with logging, mining and agriculture, and intensifying droughts related to ENSO events (Slik et al., 2008). 5.2.2 Selection of variables and proxy data layers An overview of the method is given in Figure 5.1. We included multiple biophysical and anthropogenic variables in the analyses, based on data at a regional scale so that the interrelationships between the explanatory variables could be accounted for. In Figure 5.2, a landscape-scale view on the data layers is presented in which the landscape-scale pattern for each variable is visible. The use of spatial data enabled the analyses of AGB and several explanatory variables on a continuous scale. The initial selection of the explanatory variables was based on a literature review, field visits, visual examination of spatial data and data availability (for Data sources, see Table A 23 and Appendix 19). The AGB map (Figure 5.2a) is based on ALOS PALSAR-LiDAR data and plot-based measurements (Quiñones et al., 2011). Disturbed tropical forest landscapes such as North-East Kalimantan are often covered by clouds or haze. Radar remote sensing is not affected by clouds and has proven to be a remote sensing system responsive to AGB (Mitchard et al., 2009; Morel et al., 2011). Saturation of the radar signal at medium AGB levels (150 t ha-1) restricts the use of radar remote sensing for a direct radar image inversion into AGB maps. A radar-based forest type map is used in combination with estimated vegetation heights per land cover type, derived from Geoscience Laser Altimeter System (GLAS) LiDAR data, to overcome such saturation effects. This resulted in an AGB map with a resolution of 50 m. Available vegetation height-AGB allometric equations were used to invert heights into AGB values per pixel, overcoming the effect of radar saturation. An accuracy assessment of the AGB map was conducted using field measurements over 54 plots of 0.2 ha over a range of degraded forest types in the study area. AGB values were estimated using the allometric equation developed by Saatchi et al. 2011 (Quiñones et al., 2011; Saatchi et al., 2011). The accuracy of the AGB map is

68

Selection of variables and proxy data layers Literature review, examination spatial data, expert knowledge Preprocessing data in ArcGIS - Prepare Rasters for all data layers and combine - Export attribute data to database in SPSS 20

Produce Raster in ArcGIS 10.0 for each variable

+

+

Aboveground biomass Altitude Slope Soil type Land allocation Distance to nearest: - Fire - Road - River - City

Combine all layers and select 500 pixels randomly in the landscape Export all data to SPSS for analysis

• Define the significance of differences in AGB between different variable categories

Multiple linear regression • to develop a non-spatial model to predict AGB at the landscape level

Geographically Weighted Regression • to develop a local and spatial model to predict AGB at the landscape level

• to define the relative contribution of each explanatory variable

• to define the relative contribution of each explanatory variable 9002

Correlation analysis and ANOVA • Define the strength and direction of the correlations between theindividual explanatory variables and AGB

Figure 5.1. Schematic overview of the methodological steps in the analysis.

estimated as 10 t ha-1, using the root mean squared error between the field-estimated AGB and the AGB from the radar map for the same location. For more information see Quiñones et al. (2011) (Quiñones et al., 2011). Altitude (Figure 5.2b) varied from -90 m – 2,230 m in the landscape and was selected because a relationship with AGB is expected (Alves et al., 2010; Asner et al., 2008; de Castilho et al., 2006; Whittaker and Bormann, 1974). Slope (Figure 5.2c) was found to have a positive relationship with AGB (Chave et al., 2003). Altitude and slope were derived from the Digital Elevation Model (90 m) by the Shuttle Radar Topography Mission (SRTM-DEM) (NASA, 2012). For the multiple linear regression and GWR, altitude was included as a continuous variable. For the ANOVA, altitude was categorised into several altitude ranges (Lowlands 1,500 m) (see Appendix 20). Logging and land conversion decreases AGB substantially by the harvesting and loss of especially rare tree species (Slik et al., 2008). Therefore a relationship is expected between AGB and logging intensity, and thus differences in AGB between protected forested areas, areas allocated for timber or forest concessions, and Non-forest area. The data source selected is the land (use) allocation zoning data (Figure 5.2) classified by WRI and originally produced by the Ministry of Forestry of Indonesia within the Tata Guna Hutan Kesepakatan (TGHK) mapping program (Ministry of Forestry Indonesia, year unknown) as a proxy for logging and land clearance intensity. The classes designated within this

69

(a) 0

Figure 5.2. Data layers for aboveground biomass (AGB) and the biophysical and anthropogenic variables used in this study for North-East Kalimantan: (a) AGB (9 x 9 cells focal means) (t ha-1) (Quiñones et al., 2011); (b) altitude (m above sea level) (NASA, 2012); (c) slope (%) (NASA, 2012); (d) land (use) allocation zone (WPF, Watershed protection forest; PF, Production forest; NFL, Non-forest area; CF, Conservation forest; FLP, Limited production forest) (Ministry of Forestry Indonesia, year unknown); (e) soil type (KA, karst; OT, other; PE, peat; VO, volcanic)(Consortium to Revise the HCV Toolkit for Indonesia, 2008); (f) MODIS hotspots with polygons of 500m radius (NASA/ LANCE – FIRMS, 2011); (g) main roads (Bakosurtanal, 2009); (h) main rivers (Bakosurtanal, 2009); (i) main cities (CIESIN, 2012).

70

100 km

Aboveground biomass 0

(f)

714 t ha-1

water

(g)

Fire hotspots

Main roads

(b)

(c)

0

100 km

Altitude -90

Slope 2230 m

74 %

(e)

(h)

VO

T PE

KA

Soil type

PF N FL CF FL P

W

PF

Land allocation zone

O

(d)

0

(i)

Main rivers

Main cities

71

data layer and present in the study area are: the Limited production forest zone, where logging is accompanied by measures to reduce impacts on soil erosion; The Conservation forest zone, which is conservation forest for protected areas; The Watershed protection forest zone, which is intended for watershed protection; Non-forest area has the status of non-forest use; and The Production forest zone, which is intended for commercial logging (Brockhaus et al., 2012; Broich et al., 2011a). Relationships were reported between AGB and soil drainage (Asner et al., 2008), soil texture (de Castilho et al., 2006), and soil fertility (Laurance et al., 1999; Paoli et al., 2008). Soil type (Figure 5.2) was selected as a proxy, and was included by reclassifying the improved reproduction of the RePPProT land systems map (Consortium to Revise the HCV Toolkit for Indonesia, 2008) into the categories ‘karst’, ‘peat’, ‘volcanic’, and ‘other’ (for details see Appendices S8 and S9). Forest fires (Figure 5.2f) can occur multiple times at the same spot and in this way can cause substantial losses in AGB (Cochrane, 2003; Page et al., 2002; Slik et al., 2008). Also the rate of post-fire regeneration depends, amongst others, on the frequency and age of the fire (Slik et al., 2008; Toma et al., 2005). The proxy for fire included in the multiple linear regression and GWR was ‘distance to the nearest fire’. For the fire data MODIS hotspot data from 2000 to 2008 were used (NASA/LANCE – FIRMS, 2011), because of its high accuracy recording. According to NASA, each MODIS fire hotspot represents the centre point of a ~ 1 km pixel that contains one or more fires, rather than the exact location of a fire. To overcome this uncertainty, a buffer of 500 m radius surrounding each location was created to define fire hotspot polygons. Additionally, for the ANOVA the fire variable was categorised as burned (≤ 500 m from a hotspot; i.e. the area within a fire polygon), and non-burned areas (> 500 m from a hotspot, i.e. the area outside a fire polygon) (see Appendix 20). Both roads (Figure 5.2g) and rivers (Figure 5.2h) are the primary means of transportation in North-East Kalimantan, and improve accessibility from cities (Figure 5.2i) to forest frontier areas. Therefore, a relationship is expected between AGB and the proxy distance to nearest main road (Bakosurtanal, 2009), the nearest main river (Bakosurtanal, 2009) and the nearest main city (CIESIN, 2012).

Data pre-processing In order to reduce the high local-scale variation in AGB caused by natural local variation and by the effect of speckle noise, 9 x 9 cell focal mean statistics was applied to the AGB map in ArcGIS, according to the results generated by Hoekman and Quiñones (2000) (Hoekman and Quiñones, 2000). All shapefiles were rasterised, and the proximity variables were individually processed by means of the Euclidean Distance tool of the ArcGIS Spatial analyst. For optimal processing, a sample of 500 data points was selected randomly in the data layers with a minimum distance of 1,000 m from one another to minimise the effects spatial autocorrelation (Koenig, 1999). The data layers were combined and all data queries were exported to a database in SPSS 20. Rows with missing values were deleted, resulting in a dataset of 465 data points. The continuous explanatory variables showing a skewed distribution were transformed to attain normality. Natural logarithmic (ln) data transformation was in all cases the most suitable of a series of transformations tested for attaining a linear relationship between AGB and the explanatory variables. Statistical analyses Using the Pearson’s correlation coefficient, the strength and direction of the predictive relationship between AGB and each of the continuous explanatory variables were defined. We conducted One-way ANOVA to analyse whether mean AGB among soil types and land allocation zones, among different altitudinal ranges, and between burned and unburned areas was significantly different. 72

Non-spatial backward multiple linear regression was conducted, and with every step nonsignificant (p ≥ 0.05) variables were removed one-by-one. The categorical variables land allocation zoning and soil type were included as dummy variables with, respectively, ‘Non-forest area’, and ‘other’ as the reference categories. Because land is allocated by the Ministry of Forestry of Indonesia based on climate, slope and soil type, tests for an interaction effect between land allocation zoning and altitude were carried out, by inclusion of product terms in the multiple linear regression (Allison, 1977). To verify whether the output met the assumptions underlying multiple linear regression, tests for normality and multicollinearity were carried out. To test for normality, we plotted a histogram, a normal PP plot and a normal QQ plot of the standardised residuals. We tested for the presence of significant strong multicollinearity by examining the Tolerance. Analysing ecological spatial data by multiple linear regression is challenging (e.g. (Foody, 2003; Graham, 2003), because of the possible existence of spatial autocorrelation and spatial non-stationarity, the latter being the variation in relationships and processes over space. Spatial non-stationarity was tested for by conducting the Breusch-Pagan test on random coefficients. Although often ignored, spatial autocorrelation or the spatial clustering of ecological conditions and processes is a natural, and thus widespread phenomenon (Koenig, 1999). Bini et al. (2009) (Bini et al., 2009) indicate that this can cause an unexplained shift in the regression coefficients of global or non-spatial models. To test for the presence of spatial autocorrelation, the Moran’s I Index, the z-score and the p-value for the standardised residuals were calculated. If spatial autocorrelation was present, we additionally conducted GWR in ArcGIS. GWR is a spatial and local form of multiple linear regression that considers and models the spatially varying relationships between explanatory variables and the response variable (Foody, 2003; Propastin, 2012; Wang et al., 2005). The explanatory variables that showed multicollinearity were excluded from the model.

5.3 Results 5.3.1 Relationships between AGB and the continuous explanatory variables The distribution of AGB was negatively skewed (Skewness: -0.852, st. error: 0.113, Kurtosis: 0.207, st. error: 0.226), which can be expected in a disturbed tropical forest landscape (Figure A 9 in Appendix 12). AGB varied between 2 and 480.0 t ha-1 with an overall mean of 213.6 ± 80.1 t ha-1 (for descriptive statistics, see Table A 21 in Appendix). AGB and the selected continuous explanatory variables altitude, slope, and distance to the nearest fire, road, river and city (logarithmically transformed) appeared to have a strong, positive correlation (Table A 22 in Appendix). All relationships are plotted in Figure 5.3. The Pearson’s correlation coefficients (r) indicated the strongest relationships between AGB and the terrain variables, altitude (r = 0.740, P < 0.001) and slope (r = 0.563, P < 0.001), and between AGB and distance to the nearest fire (r = 0.607, P < 0.001) and city (r = 0.478, P < 0.001). Moderately positive relationships were found between AGB and distance to the nearest river and road (r ~ 0.335, P < 0.001). Altitude and distance to the nearest city were strongly related to all other explanatory variables (r > 0.400, P < 0.001), but not to distance to the nearest river. Distance to the nearest fire was related to the distance to the nearest river and the nearest city. No strong multicollinearity was found (Tolerance>0.200). 5.3.2 Variation in AGB between altitude ranges and soil types The ANOVA on the categorised altitude variable revealed significant differences in mean AGB between the categories lowlands (< 750 m), midlands (750 – 1,500 m) and highlands (> 1,500 m), F (2,462) =

73

AGB (t ha-1)

400 300 200 100 0

1

2

3

4

5

6

Altitude (ln)

7

0

1

2

Slope (ln)

3

4

5

6

7

8

9

Distance to the nearest fire (ln)

10

AGB (t ha-1)

400 300 200

0

9002

100 9

10

11

12

Distance to the nearest city (ln)

4

5

6

7

8

9 10

Distance to the nearest road (ln)

4

6

8

10

Distance to the nearest river (ln)

Figure 5.3. Relationships between aboveground biomass (AGB, t ha-1) and all continuous variables (logarithmically transformed, ln) in the sample. The solid circles represent the 465 sample points, the grey line depicts the regression line and the black line depicts the smooth curve.

32.85, P < 0.001. The lowlands (M = 201) had significantly lower AGB than the midlands (M = 276, P < 0.001) and the highlands (M = 282, P < 0.05). A boxplot of altitude and AGB is shown in Figure 5.4a. An ANOVA was used to test for mean differences in AGB among four soil types. Means in AGB for soil types differed significantly across the four types (F(3,461) = 14.88, P < 0.001). Bonferroni’s post-hoc comparisons on the four soil types indicate that AGB on peatland (M = 142) gave significantly lower means than on karst (M = 261, P = 0.001) and volcanic soils (M = 282, P < 0.001) (Figure 5.4b). 5.3.3 Variation in AGB between burned and non-burned areas and land allocation zones AGB in non-burned areas (i.e. areas where no fire hotspots were identified by the Moderate-resolution Imaging Spectroradiometer (MODIS) between 2000 and 2008) (M = 223, P < 0.001) was significantly higher compared to burned areas (i.e. MODIS fire hotspots were identified within 500 m from the data point between 2000 and 2008) (M = 114, F (1,463) = 79.22, P < 0.001). A boxplot of fire and AGB is shown in Figure 5.4c. Fires were more common in the lowlands (98%) compared to the midlands and highlands. An ANOVA showed significant differences in the mean for AGB between the five land allocation zones (F (4,460) = 56.06, P < 0.001) (see also Figure 5.4d). After Bonferroni’s correction, pairwise comparisons showed that the mean AGB was significantly lower in the Non-forest area (M =

74

Altitude, categorised

Soil types (b)

(a) AGB (t ha-1)

400 300 200 100

Fire, categorised

ic an

rs t

Vo lc

Ka

at Pe

O

th

er

nd s

s

ig hl a

nd H

id la

M

Lo w

la

nd s

0

Land allocation zones (d)

(c) AGB (t ha-1)

400 300 200

PF

P FL

FL N

Burned

PF

Unburned

W

0

CF

9002

100

Figure 5.4. Boxplots showing the median, the upper and lower quartile for aboveground biomass (AGB, t ha-1) in the categorised (a) altitudes; (b) soil types; (c) unburned and burned areas; and (d) land allocation zones (CF, Conservation forest; WPF, Watershed protection forest; NFL, Non-forest area; FLP, Limited production forest; PF, Production forest).

152, P < 0.001) compared to the other zones, and was significantly higher in the Watershed protection forest zone (M = 272) and the Limited production forest zone (M = 253), compared to Production forest zone (M = 193) and Conservation forest zone (M = 211). 5.3.4 Multiple linear regression After removal of the non-significant explanatory variables via conducting a backward multiple linear regression, the variables altitude, distance to the nearest fire, and the categorical variables land allocation zoning and soil type significantly contributed to predicting AGB, and combined explained approx. 59% of the observed variance in AGB in the sample (Adjusted R2 = 0.589, F(9,455) = 72.46, P < 0.001). The standardised coefficients showed that in this analysis, altitude was the most important explanatory variable (Table 5.1). Altitude and distance to the nearest fire both showed a positive relation with AGB, which means that with increasing altitude and distance to the nearest fire AGB increased. Soil type also made a difference with respect to AGB. Compared to the reference category ‘other’, the categories volcanic and karst showed a higher mean AGB. Karst was the only significant coefficient compared to ‘other’. When compared to the reference category ‘Non-forest area’, all land allocation zones showed a higher mean AGB.

75

Table 5.1. Unstandardised coefficients resulting from the non-spatial multiple regression, without (Model 1) and with interaction terms (Model 2). Standardised coefficients are indicated between brackets; **P 70 %) appeared as a dark green colour with a coarse texture on the 5-4-2 bands of the Landsat images, and included pristine forests as well as post-logging regrowth with fast-growing tree species. Medium canopy forest (medium density, CC = 40-70%) had a colour and texture similar to the closed canopy forest type, but showed signs of degradation by the presence of bare soil (pink or red colour), networks of logging roads, gaps, log yards and ponds and base camps. Open canopy forest (low density, CC = 10-40%) showed more recent logging activities and the emergence of shrubland. Shrubland is a vegetation type that arises after land clearing and may include a few remaining natural trees (< 5 m high). Grassland included cleared land and occurred in dry and swampy areas. This LULC type has a grey to green colour on Landsat-5 TM bands 5-4-2 due to flooding. Plantations were classified as pulpwood plantations, oil palm plantations and rubber plantations. These plantations showed regular straight boundaries and covered relatively large areas. When the plantations are young, the colour of the vegetation is similar. Pulpwood (Acacia spp.) plantations are industrial forest plantations. Young Acacia (mangium) is light green in young stands and dark green in mature stands. Oil palm (Elaeis guineensis) plantations include large-scale estates and smallholder plantations. The colour on Landsat-5 TM bands 5-4-2 is light yellow in young stands and light green in mature stands. Rubber (Hevea brasiliensis) plantations are generally planted by plantation enterprises and are therefore homogenous in tree composition. When mature, the plantation has a brownish red colour. Smallholder rubber is a land use with rubber trees that are often mixed with secondary re-growth. Smallholder rubber has a reddish yellow colour. Mixed cropland refers to seasonal crops, orchards and paddy fields. The seasonal crops were difficult to distinguish from recently burned area as the colour on the Landsat image is similar (pink-red/reddish). We were able to identify them by assuming that the seasonal crops are commonly planted near natural forest or old secondary vegetation. Orchards were identified as these are mostly dominated by durian, coconut and mango trees, and are found behind old Dayak villages. The orchards are difficult to distinguish from old secondary vegetation as the structure is similar, however, their location is close to settlements and alongside rivers and roads. Paddy fields are commonly found on alluvial plains along the edge of big rivers. Coal mining was identified based on the purple-red colour on the Landsat images (5-4-2), with adjacent roads towards the river ports.

141

Appendix 2. Land use and land cover change processes and trajectories Land cover incorporates all biophysical attributes of the earth’s surface. Land use indicates the humaninduced utilisation of this land cover (Lambin et al., 2001). For example, forest is a certain land cover type, however, the utilisation of forests can be different; for example, forest is a land cover type that can be used as a conservation or production forest. In this paper, LULC change is the process of change or changes of one LULC type to another. In Figure A 1, a conceptual framework of LULC change processes is shown, in which the left box shows the naturally-occurring LULC types, such as grasslands, shrublands and forest types. Forest types are classified according to canopy cover; closed canopy forest (>70% canopy cover), medium open canopy forest (40-70% canopy cover) and open forest (10-40% canopy cover) (Budiman et al., 2014). The right box shows the land use types which are human-induced, e.g. agricultural lands, plantations, settlements and mining sites. The arrows show the LULC change processes that occur if one LULC type changes into another. Lambin (1997) defines three types of LULC change processes, namely land-cover conversion, land degradation or land-use intensification. Although this distinction captures the processes very well, we focus in this paper on the processes that change the land cover or the land use, defined as LULC change processes. In this study, we define land cover degradation as the process of decreasing canopy cover within the forest category, for example from closed to medium canopy forest, or a decreasing biomass density within non-forest types, such as from shrublands to grasslands. In Figure A 1we can see that forest can be degraded to another forest type and further deforested to shrubland or grassland. We define deforestation as a decrease in tree canopy cover to below a 10% threshold (“Food and Agriculture Organization of the United Nations,” 2014) or, more clearly, a change from forest to either shrubland or grassland. Regrowth or regeneration of degraded or deforested vegetation cover can occur, for instance from grassland to shrubland or forest type. Land cover and land use conversion is defined as the change from any LULC type to a LULC type that is used for the production of commodities or extraction of natural resources, as shown in Figure A 1. Such land uses can subsequently be abandoned and regenerated into natural land cover types. All non-forest types, including shrublands, grasslands and agricultural lands can be cleared through the process named land clearance. LULC change trajectories are different from LULC change processes in that these indicate the ‘exact’ sequence of LULC changes in a certain area over a specific period of time (Mertens and Lambin, 2000; Petit et al., 2001), and consequently provide information about the changes from one LULC type to another. Trajectories can involve one or more steps or transitions; can be part of LULC change processes; or can be a combination of different LULC change processes (Figure A 1). For example, a trajectory can include the LULC change processes ‘forest degradation’ and ‘conversion’ in one sequence within a specific time period. Also, forest degradation can follow the LULC change trajectory of a closed canopy forest to a medium open canopy forest, and further to an open canopy forest type.

142

Forest

sio

n

Conversion Abandonment and regeneration

d

n La

Grassland/ cleared land

er

ce

an

r ea

cl

Land use/cover type

Conversion

Degradation and deforestation

Canopy cover of the land use/cover type

Regeneration

nv

n

io

rs

e nv

Co

9002

Co

Figure A 1. Conceptual model of the LULC changes and LULC change processes in a disturbed tropical forest landscape. (Processes are indicated by a single arrow; trajectories can involve one or multiple arrows)

143

A

9002

B

144

C

Closed canopy forest Medium open canopy forest Very open canopy forest Shrubland Grassland/ cleared land Mixed cropland Smallholder rubber Rubber plantation Pulpwood plantation Oil palm plantation Gold mining Coal mining Settlement 0 Water

100 km

Figure A 2. Land use/land cover maps of West Kutai and Mahakam Ulu Districts for 1990 (A), 2000 (B) and 2009 (C) (following Budiman et al. 2014) (1 raster cell is 100 x 100 m).

145

Consessions Logging Timber Oil palm

Altitude (m) 2200

9002

1

0

100 km

Land allocation zones Non-forest area Watershed protection forest Production forest Limited production forest Conservation forest

Figure A 3. A) Land allocation zones; B) altitude (m) overlaid with concession maps for West Kutai and Mahakam Ulu Districts (source: World Resources Institute 2014).

146

Table A 1. Specification of the selected land use and land cover types (following Budiman et al. 2014). LULC type

Subclasses in LULC types

Soil type

Canopy – Crown cover

1. Closed canopy forest (>70% canopy cover)

Dry Lowland Forest rather closed canopy Forest on Sandstone rock rather closed canopy Fresh Water Swamp Forest rather closed canopy Peat Swamp Forest rather closed canopy

Dry Sandstone rock

Closed – >70% Closed – >70%

Peat swamp

Closed – >70%

Dry Lowland Forest medium open canopy Forest on Sandstone rock medium open canopy Fresh Water Swamp Forest medium open canopy Peat Swamp Forest medium open canopy

Dry Sandstone rock

Medium open – 40-70% Medium open – 40-70%

Peat swamp

Medium open – 40-70%

Dry Lowland Forest very open canopy Forest on Sandstone rock very open canopy Fresh Water Swamp Forest very open canopy Peat Swamp Forest very open canopy Forest Re-growth (Belukar) Forest Re-growth on Swampy

Dry Sandstone rock Fresh water swamp Peat swamp

Very open – 10-40% Very open – 10-40% Very open – 10-40% Very open – 10-40% Re-growth (Belukar) Re-growth

2. Medium open canopy forest (40-70% canopy cover)

3. Open canopy forest (10-40% canopy cover)

4. Shrubland

Shrubland (Semak/Belukar Muda) Shrubland on Sandstone Forest Shrubland on Swampy Overgrowing Clear cut-Shrubland

5. Grassland/cleared land

Swamp Grasses/Fernland Grassland Burnt Cleared Cleared for Industrial Pulpwood plantation Cleared for Oil palm Plantation

6. Mixed cropland

Mixed Agriculture Mixed Garden

7. Smallholder Rubber

Small Holder Rubber

8. Rubber plantation

Rubber Plantation

9. Pulpwood plantation

Acacia Plantation Paraserianthes falcataria Industrial Pulpwood plantation

10. Oil palm plantation

Oil palm Plantation Young Oil palm Plantation Young Oil Palm Plantation

11. Gold mining

Gold Mining

12. Coal mining

Coal Mining

13. Settlement

Settlement

14. Water

Water Body

Fresh water swamp Closed – >70%

Fresh water swamp Medium open – 40-70%

Swampy

Semak/Belukar Muda Sandstone rock Swampy clear cut, overgrowing Swampy

147

148

4. Shrubland

15,600

9,800

700

2,600

1,600

900

66,100

44,000

400

200

4,800

14,500

1,300

900

1,400

1,000

400

16,350

8,000

100

50

8,200

4,140

100

4,000

40

650

50

100

200

300

11. Gold mining

900

600

300

12. Coal mining

3,100

3,000

100

13. Settlement

4,600

100

300

14. Water

29,800

48,510

10

41,800

1,600

1,600

3,200

300

6. 7. 8. 9. 10. Mixed Smallholder Rubber Pulpwood Oil palm cropland Rubber plantation plantation plantation

Total 309,600

1,500

20

418,950

2,400

218,700

18,600 30

31,200

45,900

9,900

4. 5. Shrubland Grassland/ cleared land

311,300

59,800

29,200

3. Open canopy forest

24,800

1,114,300

4,700

834,000

275,600

2. Medium open canopy forest

14. Water

13. Settlement

12. Coal mining

11. Gold mining





6. Mixed cropland


1,264,400

2,600



1,261,800

1. Closed canopy forest


Table A 2. LULC change matrix (in ha) 1990-2000 for West Kutai and Mahakam Ulu districts. These matrices indicate the number of pixels change from one LULC type (vertical axis) to another (horizontal axis) within the selected time period.

149

0.50

2.00

0.04

0.03

0.01

0.50

0.20

0.30

0.10

0.10

0.01

0.01

0.03

0.02

0.01

0.10

0.10

0.14

0.01

1.00

1.50

1.30

0.01

0.01

0.20

0.40

0.04

0.03

Total (%)

9.40

0.30

0.02

0.10

0.05

0.03

14. Water

0.80 12.80

1.27

0.05

0.05

0.10

0.01

6. 7. 8. 9. 10. 11. 12. 13. Mixed Smallholder Rubber Pulpwood Oil palm Gold mining Coal mining Settlement cropland Rubber plantation plantation plantation

14. Water

13. Settlement

12. Coal mining

11. Gold mining





0.04

6.60

1.00

1.40

0.30


6. Mixed cropland

0.60

9.50

1.80

0.90


0.07

33.80

0.10

25.30

8.40



4. Shrubland

38.40

0.08



38.30



Table A 3. LULC change matrix (in %) 1990-2000 for West Kutai and Mahakam Ulu districts. These matrices indicate the number of pixels change from one LULC type (vertical axis) to another (horizontal axis) within the selected time period.

150

1,000


1,094,400

26,200

80

400

62,500

1,800

80

13,600

36,300

7,000

700

3,000

900

1,300

800

16,400

300

200

1,400

7,600

4,000

1,300

442,830

14. Water

Total

20 10

13. Settlement

278,610

102,870

122,460

31,800

400 60

400

9,900

200

5,600

4,300

4,400

1,400

12. Coal mining

60

50

300

36,500

45,800

13,200

5,200

1,700

31,130

4,200

800

30

1,400

17,600

3,000

1,700

2,400

6. 7. 8. 9. 10. Mixed Smallholder Rubber Pulpwood Oil palm cropland Rubber plantation plantation plantation

200

10

2,200

6,300

176,100

63,200

26,700

4,100


11. Gold mining




100

200


6. Mixed cropland

36,800

286,700

108,200

9,000


1,900

400

3,000

948,700

142,200



1,114,500

10,800


4. Shrubland

1,102,700



100

100

11. Gold mining

6,760

10

3,100

40

600

700

200

1,000

1,100

10

Total land area:

7,730

400

200

40

90

3,400

3,100

300

200

12. 13. Coal mining Settlement

3.3 Mha

31,050

29,700

400

200

700

10

30

10

14. Water

Table A 4. LULC change matrix (in ha) 2000-2009 for West Kutai and Mahakam Ulu districts. These matrices indicate the number of pixels change from one LULC type (vertical axis) to another (horizontal axis) within the selected time period.

151

0.03


0.80

0.01

0.30

0.01

0.20

0.10

0.10

0.04

3.70

0.01

1.90

0.06

0.40

1.10

0.20

0.02

0.09

0.03

0.04

0.02

0.90

0.01

0.50

0.01

0.04

0.20

0.10

0.04

0.90

0.10

0.03

0.04

0.50

0.09

0.05

0.07

0.00

0.20

0.01

0.01

0.10

0.09

0.01

0.01

0.20

0.09

0.00

0.02

0.02

0.03

0.03

0.01

0.01

0.02

14. Water

0.90

3.20

0.01

1.10

1.40

0.40

0.20

0.05

6. 7. 8. 9. 10. 11. 12. 13. Mixed Smallholder Rubber Pulpwood Oil palm Gold mining Coal mining Settlement cropland Rubber plantation plantation plantation

Total (%)

8.50

0.07

0.01

13.50

0.20

5.40

1.90

0.80

0.10


0.06

1.10

8.70

3.30

0.30


0.90

33.20

0.01

0.09

28.80

4.30


14. Water

13. Settlement

12. Coal mining

11. Gold mining





6. Mixed cropland


33.80

0.30


4. Shrubland

33.50



Table A 5. LULC change matrix (in %) 2000-2009 for West Kutai and Mahakam Ulu districts. These matrices indicate the number of pixels change from one LULC type (vertical axis) to another (horizontal axis) within the selected time period.

152

13. Settlement

12. Coal mining

11. Gold mining





6. Mixed cropland


4. Shrubland





Abandonment and regeneration

Regeneration

Degradation

Land clearance

Deforestation

2. 3. 4. 5. Medium Open canopy Shrubland Grassland/ open canopy forest cleared land forest

6. Mixed cropland

7. 8. Smallholder Rubber Rubber plantation


Conversion to agricultural land


Table A 6. LULC change processes defined, i.e. the conversions from LULC types in left columns to LULC types in top rows. 11. Gold mining

12. 13. Coal mining Settlement

Table A 7. Overview of net area lost in 1990 and net area gained in 2009 per LULC type (in ha) Total land lost between 1990 and 2009

Total land gained between 1990 and 2009

Ha

Ha

Ha

%

%

%

478,300 260,500

13,700 397,900

-464,600 137,400

-9 -29 14

-4 -20 16

-5 -12 -2

140,300 118,800 18,700

217,000 148,600 72,500

76,700 29,800 53,800

5,100 3,300 50 0 0 0 400 100 90

19,700 81,800 2,100 23,6000 31,000 80 7,600 3,800 6,400

14,600 78,500 2,050 236,000 31,000 80 7,200 3,700 6,310

21 12 714 235 125 178 211 293 22,982 151 1,270 119 25

15 24 6 58 34 50 43 103 3,039 1,115 49 3 20

6 -10 654 112 68 85 118 94 635 -79 821 113 4

Total ha changed

1,025,600

1,238,200

212,500

Total ha no change

2,268,000

2,055,400

3,081,100

Total forested land 1. Closed canopy forest 2. Medium open canopy forest 3. Open canopy forest 4. Shrubland 5. Grassland/cleared land Land use types 6. Mixed cropland 7. Smallholder Rubber 8. Rubber plantation 9. Pulpwood plantation 10. Oil palm plantation 11. Gold mining 12. Coal mining 13. Settlement 14. Water

Net loss or gain between 1990 and 2009

Net loss or Net loss or gain between gain between 1990-2000 2000-2009

153

≤2,500 5,000 10,000 20,000 50,000

Mixed cropland

100,000


200,000



Settlement Very open canopy forest

Oil palm plantation

Coal mining

Shrubland

Grassland/ Cleared land

Pulpwood plantation

Figure A 4. The LULC trajectories (single arrows) as identified in West Kutai and Mahakam Ulu districts between 1990 and 2009.

154

9002

Gold mining

Mixed cropland


C.1

C.2 Smallholder rubber

Medium open canopy forest

Very open canopy forest

Regeneration

A.

Rubber plantation

Settlement

Oil palm plantation

B.

Shrubs

Coal mining

Gold mining Grassland/ Cleared land Pulpwood plantation

Figure A 5. The main LULC change processes and trajectories identified for West Kutai and Mahakam Ulu based on expert knowledge. The trajectories are either dominant (thick arrows) or less dominant (thin arrows): A) degradation and/or deforestation to grasslands or shrublands; B: deforestation to grasslands, and conversion from grasslands to large-scale plantations; C1) degradation and/or deforestation to grasslands, conversion from grasslands to mixed cropland (which the experts referred to as shifting cultivation areas and/or dryland rice fields), after a few years, conversion from mixed cropland to smallholder rubber, and in certain cases; C2) further conversion from smallholder rubber to permanent LULC types, such as oil palm.

155

Table A 8. Land area (hectares) of LULC types in concessions in 2009. Total land area in 2009

Land area outside concessions*

Land area in concessions

Total

Logging concession

Oil palm concession

Timber concession

Total forested land 1. Closed canopy forest 2. Medium open canopy forest 3. Open canopy forest 4. Shrubland 5. Grassland/cleared land Land use types 6. Mixed cropland 7. Smallholder Rubber 8. Rubber plantation 9. Pulpwood plantation 10. Oil palm plantation 11. Gold mining 12. Coal mining 13. Settlement 14. Water

2,651,900 1,114,400 1,094,600

1,348,200 801,200 368,500

1,303,700 313,200 726,100

1,073,600 289,900 619,700

144,700 17,200 61,300

85,300 6,100 45,000

442,800 278,600 103,000

178,400 94,300 50,500

264,400 184,300 52,500

164,000 84,500 9,700

66,200 63,900 35,300

34,200 35,900 7,600

229,200 26,200 122,500 3,000 31,700 31,200 100 7,800 6,700 31,100

91,200 13,200 61,600 2,000 -6,300** 11,300 100 4,500 4,800 28,300

138,000 13,000 60,900 1,000 38,000 19,900

46,600 4,900 24,000 100 13,000 2,200

54,200 5,000 27,800 600 1,700 16,600

37,400 3,100 9,200 200 23,300 1,200

3,300 1,900 2,800

1,800 600 200

1,500 1,000 2,600

400 -

Total land area

3,300,000

1,618,700

1,681,300

1,214,600

300,700

166,200

49%

51%

37%

9%

5%

Total share of land in study area

* The land area outside concessions was estimated by subtracting the total land area in 2009 with the land area in concessions. ** There is a discrepancy between the total land area in 2009 and the land area in concessions, as there is overlap between the concession types.

Table A 9. Occurrence (%) of LULC change processes between 2000-2009 in concession types. Land area outside concessions LULC change processes Abandonment and regeneration Clearance Conversion Deforestation Degradation Regeneration

156

49% 35% 44% 29% 29% 60%

Land area in concessions

Total

Logging concession

Oil palm concession

Timber concession

51% 65% 56% 71% 71% 40%

6% 8% 18% 38% 51% 24%

14% 45% 26% 20% 11% 13%

31% 11% 12% 12% 9% 3%

Table A 10. Occurrence of LULC types (in 2009) in land allocation zones (in hectares). Non-forest area

Conservation forest

Limited production forest

Production forest

Watershed protection forest

Other (boundary issue)

1. Closed canopy forest 2. Medium open canopy forest 3. Open canopy forest 4. Shrubland 5. Grassland/cleared land 6. Mixed cropland 7. Smallholder Rubber 8. Rubber plantation 9. Pulpwood plantation 10. Oil palm plantation 11. Gold mining 12. Coal mining 13. Settlement 14. Water

60,100 194,300

12,500 2,200

344,100 518,400

77,900 271,100

616,400 106,200

400 1,200

204,500 148,200 84,000

900 1,700 2,400

118,400 60,000 600

104,200 66,100 5,800

13,300 2,000 100

1,500 700 1,200

18,100 89,700 2,600 5,100 25,500 – 7,100 5,800 5,400

– 50 – – – – – –

2,400 6,100 – 7,700 – 100 – 200 600

5,300 25,100 300 18,700 5,600 – 700 400 50

100 900 – – – – – 100 50

200 300 – – – – 200 25,000

Total

850,400

19,750

1,058,600

581,250

739,150

30,700

Table A 11. Occurrence of timber, logging and oil palm concessions in land allocation zones.

Overlapping pixels with other concessions Non-forest area Conservation forest Limited production forest Production forest Watershed protection forest Total

Logging concession

Oil palm concession

Timber concession

900

2,300

100

77,100 3,600 741,300 364,300 27,400

224,900 – 8,800 63,900 500

8,600 – 37,500 119,700 300

1,214,600

300,400

166,200

157

Appendix 3. Area of LULC types in the land allocation zones in 2009 Table A 12. Area of LULC types (ha) (based on LULC maps in 2009, Budiman et al., 2014) in land allocation zones (“World Resources Institute,” 2014). Non-forest area Conservation forest

Limited production forest

Production forest

Watershed protection forest

Other (boundary issue)

1. Closed canopy forest 2. Medium open canopy forest 3. Open canopy forest 4. Shrubland 5. Grassland/cleared land 6. Mixed cropland 7. Smallholder Rubber 8. Rubber plantation 9. Pulpwood plantation 10. Oil palm plantation 11. Gold mining 12. Coal mining 13. Settlement 14. Water

60,100 194,300

12,500 2,200

344,100 518,400

77,900 271,100

616,400 106,200

400 1,200

204,500 148,200 84,000

900 1,700 2,400

118,400 60,000 600

104,200 66,100 5,800

13,300 2,000 100

1,500 700 1,200

18,100 89,700 2,600 5,100 25,500 – 7,100 5,800 5,400

– 50 – – – – – –

2,400 6,100 – 7,700 – 100 – 200 600

5,300 25,100 300 18,700 5,600 – 700 400 50

100 900 – – – – – 100 50

200 300 – – – – 200 25,000

Total

850,400

19,750

1,058,600

581,250

739,150

30,700

158

Unrestricted land zoning**

no no

no no

no no

yes yes

no yes

yes no

no no

no no

no no

no no

no no

no no

no no

no no

yes no

no no

no no

yes yes

yes yes

yes yes

no no

no no

No No

yes yes

yes yes

yes yes

Yes Yes

Yes Yes

Yes Yes

no no

no no

no no

yes yes

yes yes

yes yes

Watershed Production Limited Conservation Production forest Non- Watershed Production Limited Conservation Production forest protection forest zone production forest zone for conversion Forest protection forest zone production forest zone for conversion forest zone forest zone zone* area forest zone forest zone zone*

yes yes

NonForest area

Restricted land zoning**

* not present in study area. ** Under restricted land zoning, development is not possible on peat; under unrestricted zoning, development is possible on peat.

Mixed cropland Rubber (mostly smallholder) Oil palm Mining (mostly coal) Pulpwood Settlements

Zone

Scenario

Table A 13. Development of the LULC types in the land allocation zones, assuming restricted land zoning (R) (no reclassification to the Non-Forest area) and unrestricted land zoning (UR) (reclassifications to the Non-Forest area applied)(see Figure A 3 for the land allocation zones in 2009) (‘yes’ indicates the go-areas in the PLUC model and ‘no’ indicates the no-go areas in the PLUC model).

Appendix 4. Development of LULC types in the land allocation zones under the scenarios

159

9002

Appendix 5. Peatland overlaid with land allocation zoning map

0

100 km

Non-forest area

Forest area Watershed protection forest zone Production forest zone Limited production forest zone Conservation forest zone Peat

Figure A 6. Peatland map overlaid with land allocation zoning map.

160

Appendix 6. Equations for the growth curves Equation 6 and Equation 7 show the exponential (unlimited) and S-shaped (limited) growth curves. Equation 6 yi,e (t) = ai ebi t

Equation 7 yi,s (t) = ai ebi t for t ≤ 34 (year 2014) yi,s (t) = ( ai ebi (t-1) ) + 0.5 ( ai ebi (t-1) – ( ai ebi (t-2) ))

for t > 34

Where in both equations: yi,e(t)or yi,s(t) is the demand of active LULC type i at time step t following the exponential (e) (unlimited) or S-shaped (s) (limited) curve; ai and bi are constants for each active LULC type (See Table A 14 for the fitted values). e = constant (=2.718281828) t = time step (1981 = 1) t-1 is the year before and t-2 two years before. With this equation, the land-use demand of every LULC type will decrease with 50% annually.

Table A 14. Fitted values for the constants ai and bi for the exponential-growth curve of each active land use type.

Mixed cropland Rubber (mainly smallholder) Oil palm Pulpwood plantations Settlements Coal mining

Constant ai

Constant bi

7479.866 25970 11.318 4097.272 1841 114

0.0432 0.0543 0.2732 0.0706 0.0448 0.1455

161

Appendix 7. Land use and land cover in 2009 and projected for 2020 and 2030 Table A 15. Land use and land cover area (ha) in 2009 and the projected land use and land cover area (ha) in 2020 under the four scenarios. 2009-2020 Forest – closed Forest – medium open Forest – very open Shrubs/grassland Mixed cropland Rubber (mostly smallholder rubber) Pulpwood plantation Oil palm Mining (mostly coal) Settlement

2009

Limited_ restricted

Limited_ unrestricted

Unlimited_ restricted

Unlimited_ unrestricted

1,114,442 1,094,593 442,842 381,586 26,151 125,459

1,102,838 1,080,300 361,994 328,744 32,425 166,913

1,100,856 1,070,313 365,869 336,719 32,419 167,04

1,063,381 930,019 269,688 182,388 42,113 218,475

1,045,581 864,619 177,006 163,388 42,113 229,650

31,698 31,160 7,885 6,741

44,875 115,056 15,875 8,425

44,875 115,056 15,869 8,425

68,877 438,144 34,413 9,944

67,163 618,875 38,169 10,881

Table A 16. Projected land use and land cover area (ha) in 2030 under the four scenarios. 2030

Limited_ Restricted

Limited_ unrestricted

Unlimited_ restricted

Unlimited_ unrestricted

Forest – closed Forest – medium open Forest – very open Shrubs/grassland Mixed cropland Rubber (mostly smallholder rubber) Pulpwood plantation Oil palm Mining (mostly coal) Settlement

1,102,819 1,080,300 361,856 328,625 32,431 166,975 44,900 115,225 15,888 8,425

1,100,844 1,070,244 365,731 336,663 32,431 167,100 44,900 115,225 15,881 8,425

1,055,944 907,350 246,806 158,875 43,494 217,413 139,469 328,900 150,600 8,594

730,031 271,581 63,181 56,038 64,863 332,781 105,675 1,466,425 156,788 10,081

162

Appendix 8. Analytical framework of the PCRaster Land Use Change model

Land use at t

t=t+1

Create suitability map and select next land use type

Land demand

no no

yes

Have all active land use types been allocated?

yes

Add land of this land use type

Is the demand of this land use type fulfilled?

Figure A 7. PLUC modelling framework (adapted from Verstegen et al. (2012)). The active land use types are oil palm, rubber (mostly smallholder) and pulpwood plantations, and mixed cropland, coal mining and settlements.

Parameterisation A Landsat-based LULC map for 2009 was used as the initial LULC map (Budiman et al., 2014; Chapter 3). The raster cell size of the maps was 250 x 250 m to optimise between computation time and representation of the field size of the active land use types. The model distinguishes between active dynamic, passive dynamic and static LULC types. Active dynamic land use types (further referred to as the land use types) expand based on the increase of future demands. The active land use types included the selected land uses: oil palm plantations, pulpwood plantations, rubber (mostly smallholder), mixed cropland, mining (mostly coal) and settlements. Passive dynamic LULC types remain either unchanged or decrease in their land-use demand when these are replaced by an active land use type. In our study, passive dynamic LULC types are closed canopy forest, medium open canopy forest, open canopy forest, shrubs, grasslands, cleared lands and gold mining. The static LULC type in this study is water, which is the only LULC type that remains unchanged. To spatially allocate the projected land-use demands of the active land use types (Figure 4.3) within the allocated land zones, we defined a set of suitability factors and no-go areas. The suitability factors included altitude, slope, distances to primary roads, secondary roads, rivers and towns, travel time to the nearest palm oil mill (only for oil palm), travel time to the nearest town, current LULC and 163

the area of equivalent LULC type in the direct neighbourhood. These suitability factors are normalised between 0 and 1 to be able to sum their effect.

Suitability factors of each active land use type We selected altitude and slope as suitability factors, and consequently as explanatory variables in the regression, because land in the higher altitudes and on steeper slopes is generally less suitable for the development of agriculture. Altitude and slope were derived from the Digital Elevation Model (90 m resolution) by the Shuttle Radar Topography Mission (SRTM-DEM)(NASA, 2012). Infrastructure variables were selected because agriculture and mining are generally developed close to roads, rivers and settlements for transportation of agricultural and mining products. The infrastructure suitability factors included travel time to the nearest town and to the nearest palm oil mill and distance to the nearest primary and secondary roads (Bakosurtanal, 2009), rivers (Bakosurtanal, 2009), settlements (CIESIN, 2012). Travel time to towns provides the time-distance of transportation by assuming an average speed of 70 km/h on primary roads, 40 km/h on secondary roads, 25 km/h on rivers and 10 km/h on all other land cover, i.e. no roads/no water. Oil palm is generally planted close to an palm oil mill and therefore palm oil mill was also included as a variable. The suitability factor current LULC means how the current land use type at a certain location (raster cell) affects expansion of another land use type over this location. It was selected because specific LULC change trajectories have been found in the study area (see Chapter 3). Current LULC was based on the trajectories that were identified in the study area (see Chapter 3). The suitability factor neighbourhood was selected as certain land use types are developed as larger-scale plantations. Neighbourhood was defined as the area of equivalent land use type within an extended Moore neighbourhood of 1250 m (5 x 5 raster cells). For each active land use type, we selected for the regression a sample of 500 data points randomly in the data layer, with 250 data points in raster cells where the active land use type was not present, and 250 data points in raster cells where the active land use type was present. The suitability factors were normalised between 0 and 1 and weighted (Table A 18) by forward (Wald) stepwise binary logistic regressions in SPSS. All explanatory variables that were significantly correlated to the conversion into the active land use type between 2000 and 2009 and showed no multi-collinearity (i.e. Pearson’s correlation coefficient, r >.750) were included as suitability factors in the PLUC model. We selected the time period 2000-2009 for the regression analyses since around 2000, a shift occurred from forest-products oriented land use to cash-crop land use (Müller et al., 2014). Therefore, it is expected that trends before 2000 are not representative for current processes in the region and thus the trends in 2000-2009 are most representative for the identification and weighting of the suitability factors of the active land use types for the future. The weights of the suitability factors were based on the logistic regression coefficients, which indicate the intensity and direction of the relationship between each explanatory variable and the active land use types. After the logistic regression, we added random noise as a suitability factor for each of the active land use types as the stochastic variable, in order to account for the LULC change that could not be explained by the regression models. The weight of this randomness factor was defined by 1 minus the R2 of each logistic regression. This ‘randomness’ is a raster where each cell value is a realisation from a uniform distribution between 0 and 1.

164

Table A 17. Ranges and data sources of the suitability factors. Explanatory variable

Range

Data sources

Altitude Slope Distance to river Distance to primary roads Distance to secondary roads Distance to towns Travel time to towns Travel time to oil palm mill Current LULC Neighbourhood

0-2,197 m 0 – >40% 0-27,895 m 0-135,741 m 0-79,826 m 0-56,876 m

DEM (NASA, 2012) DEM (NASA, 2012) (World Resources Institute, 2014) (World Resources Institute, 2014) (World Resources Institute, 2014) (World Resources Institute, 2014) (World Resources Institute, 2014) Based on LULC map 2009 (Budiman et al., 2014); Based on LULC map 2009 (Budiman et al., 2014);

Regression results of the suitability factors The weights of the significant suitability factors that resulted from the logistic binary regressions in terms of the regression coefficients are shown in Table A 18 for each active land use type. The goodness-of-fit or the R2 of the logistic regressions varied between 0.467 and 0.689. Oil palm and pulpwood plantations were the land uses that could be explained best by the selected variables (R2 = 0.664 and 0.689, respectively). Coal mining could be explained with an R2 of only 0.467, which shows that variables other than terrain or transportation are also important for the development of coal mining or that this land use is developed more randomly distributed throughout the study area. A strong positive correlation was found between the development of mixed cropland and altitude (Wald statistic (B) = 37.737, p = 0.000), meaning that expansion of mixed cropland is mostly expected in higher altitudes. Relatively strong positive correlations were also found between mixed cropland and secondary roads (B = 27.719, p = 0.000) and between pulpwood plantations and secondary roads (B = 45.198, p = 0.000). As a result, we can expect that mixed cropland and pulpwood plantations will be developed close to secondary roads. Pulpwood plantations and slope were also strongly correlated (B = 24.905, p = 0.000), as were rubber and slope (B = 22.561, p = 0.000), and therefore these land uses expected to be developed on steeper slopes. Pulpwood plantations and coal mining were negatively correlated to distance to rivers, while it was expected that the river network is important for transport of coal and timber. However, although significant, the weights of these variables are low in comparison with the other weights. As expected, oil palm was strongly correlated to travel time to palm oil mills (B = 20.888, p = 0.000). Oil palm was also strongly correlated to distance to secondary roads (B = 18.402, p = 0.000). The cultivation of rubber is not dependent on distance to roads or rivers, but is dependent on the cost/time-distance of transportation, as was shown by the coefficient of the travel time to towns variable. We did not find a significant relationship between distance to settlements and mixed cropland and pulpwood plantations. We expected this for pulpwood plantations, as these are usually developed by companies in the highlands, far from settlements. However, we did expect a significant relationship between settlements and mixed cropland as these are developed by communities and are thus expected close to settlements. Coal mining, rubber and pulpwood plantations show to be dependent on the current LULC, meaning that these LULC types show strong trajectories with historical LULC change.

165

166

Constant R2

Distance to palm oil mill Randomness factor

Altitude (DEM) Slope (DEM) Distance to roads 1 Distance to roads 2 Distance to rivers Distance to settlements Travel time to towns Current LULC Neighbour

Suitability factors

0.085

11.090

-63.202 0.587

0.213

27.719

0.413

0.289

Weight

37.737

Wald

Mixed agriculture

-29.941 0.613

3.788 4.335 4.562 1.482

22.561

Wald

0.387

0.063 0.072 0.076 0.025

0.377

Weight

Rubber (mainly smallholder)

-42.703 0.664

20.888

8.392

2.622 18.402

Wald

0.276 0.336

0.111

0.035 0.243

Weight

Oil palm

-67.831 0.689

7.417 4.477

45.198 -2.532

24.905

Wald

0.311

0.060 0.036

0.368 -0.021

0.203

Weight

Pulpwood plantations

Active land use types/Dependent variables

-17.068 0.503

7.971 11.580

Wald

0.497

0.205 0.298

Weight

Settlements

-1.374 0.467

6.628

-3.275 3.021

Wald

0.533

0.239

-0.118 0.109

Weight


Table A 18. The regression coefficients and weights for each significant suitability factor that resulted from the forward stepwise logistic regression. Only the Wald statistic and the weight (between 0 and 1) of the significant suitability factors are shown and these were included in the model.

No-go and go areas No-go areas are areas where the active land use types cannot be used for development because of socio-economic or biophysical restrictions (for an overview, see Table A 19). In Indonesia, land with slopes of >40% or altitudes above 2,000m cannot be converted to large-scale agricultural development because of erosion risks and therefore should remain forested, according to the Rencana Tata Ruang Wilayah Provinsi, Kalimantan Timur (PP 26/2008 art 56)(Basuki and Sheil, 2005). However, smallscale community forestry activities are legally possible in these areas if these are conducted for livelihood purposes and with explicit governmental permission. In the model, we assumed that agriculture, including mixed cropland, smallholder rubber and oil palm plantations, could not expand on slopes of >30%, as accessibility and erosion risks limit the suitability for agricultural development (Kassam et al., 1992). For similar suitability reasons, we assumed that rubber and oil palm plantations could not be developed above 750m and 500m, respectively (Verheye, 2010). Furthermore, we assumed that none of the active land use types could expand on roads and water. Earlier studies showed that LULC change in the study area follows characteristic trajectories, for example, in one-stage trajectories from forest to small-scale mixed land uses or from forest to monoculture plantations, and in multi-stage trajectories from forest to small-scale mixed land uses and further to monoculture plantations (Inoue et al., 2013; Chapter 3). To account for the LULC change trajectories, we assumed additional no-go areas and the suitability factor current land use. Based on the aforementioned studies, we assumed that none of the active land use types, except for coal mining, could expand on settlements or oil palm plantations, as such conversions usually do not occur in the study area (see Chapter 3). Also, the profitability of oil palm cultivation is higher than the other land uses and, consequently, it is unlikely that plantation owners will convert their oil palm plantation into another land use. Additionally, we assumed that mixed cropland can be converted to all other active land use types and can only be developed on all natural vegetation types and pulpwood plantations. Conversions of mining sites do not occur often (see Chapter 3). Although mining companies are, under specific circumstances, required to provide a plan and finance for reforestation or reclamation of the post-mining sites, this often does not occur. We therefore chose not to incorporate post-mining usage or conversion in the model simulations. See Table A 19 for an overview of the no-go areas.

Table A 19. No-go areas that indicate where land development is not possible due to (biophysical) restrictions. Active LULC type

No-go areas

Mixed cropland Smallholder rubber Oil palm Coal mining Pulpwood plantation Settlements

Slope >30%, roads, water, settlement, rubber, oil palm, coal mining Altitude >750m, Slope >30%, roads, water, settlement, oil palm, coal mining Altitude >500m, Slope >30%, roads, water, settlement, coal mining Roads, water Roads, water, settlement, oil palm, coal mining Roads, water, oil palm, coal mining, pulpwood plantation

167

Analyses and evaluation of the PLUC output Probability maps We calculated and mapped the probability of each active land use type to be allocated to a certain area under each of the scenarios based on the 100 Monte Carlo realisations. Each raster cell of the probability maps represents the probability of occurrence of the selected land use type, which is indicated by a number between 0 and 1. A high probability of a particular land use in a certain area in 2030, that is indicated by a number close to 1, means that the area was assigned the LULC type for almost 100% of the Monte Carlo realisations. The higher the probability, the higher the chance of occurrence of an active land use type in a certain area, which is defined by the suitability factors and by the restrictions in the land allocation zoning. The higher the probability of a land use in a certain raster cell, the lower the temporal competition with other land uses, as then only one land use was assigned to that particular raster cell over all Monte Carlo realisations.

Appendix 9. Spatial evaluation of model output To define the standard error (SE) of the area of each active land use type, we first selected 100 square blocks of 10x10 km randomly distributed throughout the landscape. Second, we compared the landuse area of each active land use type within these blocks between 2009 observed and 2009 projected. Finally, we defined and spatially visualised the standard error per square block per LULC type by the following equation:

SEi =

∆ŷi – ∆yi ∆ŷi

SEi = standard error ∆ŷi = difference between land-use area 2000 and predicted land-use area 2009 for each active LULC type within each block ∆yi = difference between land-use area 2000 and observed land-use area 2009 for each active LULC type within each block

Appendix 10. LULC change under the four scenarios Table A 20. LULC change (%) under the four scenarios. LR, limited restricted; LuR, limited unrestricted; uLR, unlimited restricted; and uLuR, unlimited unrestricted. 2020

2030

LR

LuR

uLR

uLuR

LR

LuR

uLR

uLuR

Forest – closed Forest – medium open Forest – very open Total forest

-1% -1% -18% -4%

-1% -2% -17% -4%

-5% -15% -39% -15%

-6% -21% -60% -21%

-1% -1% -18% -4%

-1% -2% -17% -4%

-5% -17% -44% -17%

-34% -75% -86% -60%

Shrubs and grassland

-14%

-12%

-52%

-57%

-14%

-12%

-58%

-85%

168

Appendix 11. Results spatial evaluation of model output

0

100 km

Standard Error 16,000 LU7x100int LU8x100int LU10x100int LU11x100int LU13x100int LU14x100int

9002

Land use Forest - Closed Forest - Medium open Forest - Very open Shrubs/grassland Mixed cropland Rubber (mostly smallholder) Pulpwood plantation Oil palm plantation Mining (mostly coal) Settlement Water

Figure A 8. Standard Error (SE, bars) of the land-use area of each active land use type between 2000 and 2009 for 100 randomly selected areas of 10x10 km for each Monte Carlo realisation, overlaid with the 2009 LULC map. In the legend, the error bar shows a standard error of 160. A high standard error bar indicates a large difference between projected and observed land use area and thus a high uncertainty at that location (1 raster cell is 250 x 250 m).

169

Appendix 12. Frequency distribution of aboveground biomass in the sample Mean = 213.56 Std. Dev. = 80.098 N = 465

100.0

Frequency

80.0

60.0

40.0

20.0

0.0

0

100

200

300

400

500

AGB (t ha-1) Figure A 9. Frequency distribution of aboveground biomass (AGB, t ha-1) (bars) in the randomly selected sample of 465 data points in East and North Kalimantan. The line shows a normal distribution.

Appendix 13. Descriptive statistics of the data Table A 21. Descriptive statistics showing the mean, standard deviation and 95% confidence interval for AGB and the continuous variables in the sample Mean

-1

AGB (t ha ) Altitude (m) Slope (%) Distance to the nearest Fire (m) Road (m) River (m) City (m)

170

Standard

95% Confidence Interval

Deviation

Lower

Upper

213.6 367.3 10.3

80.1 396.5 8.8

206.4 333.7 9.5

220.2 401.3 11.1

8248.9 6385.2 11352.0 127479.2

9658.3 9584.2 10449.6 75899.7

7408.3 5579.6 10407.7 120507.3

9061.4 7221.3 12381.5 134074.6

Appendix 14. Correlation matrix for the combination of all continuous variables Table A 22. Correlation matrix showing Pearson’s correlation coefficients (P < 0.001) for the combination of all the continuous variables (ln, logarithmically transformed). Distance to the nearest -1

AGB (t ha )

Altitude (ln)

Slope (ln)

1 0.740 0.563

1 0.745

1

Distance to the nearest Fire (ln) 0.607 Road (ln) 0.369 River (ln) 0.301 City (ln) 0.478

0.696 0.460 0.383 0.623

0.492 0.322 0.182 0.422

-1

AGB (t ha ) Altitude (ln) Slope (ln)

Fire (ln)

Road (ln)

River (ln)

City (ln)

1 0.439 0.268 0.390

1 0.103 0.436

1 0.129

1

171

Appendix 15. Interaction effects between altitude and land allocation zoning Land allocation zone

400

Forest limited production Conservation forest Non-forest land Production forest Watershed protection forest Forest limited production Conservation forest Non-forest land Production forest Watershed protection forest

AGB (t ha-1)

300

200

Forest limited production: R2 linear = 0.220 Conservation forest: R2 linear = 0.860 Non-forest land: R2 linear = 0.375 Production forest: R2 linear = 0.237 Watershed protection forest: R2 linear = 0.169

0

9002

100

2.0

4.0

Altitude (ln)

6.0

8.0

Figure A 10. Interaction effect between altitude and the land allocation zones in the multiple regression.

172

Appendix 16. Frequency distribution, PP plot and QQ plot of the standardised residuals of the multiple linear regression Mean = -3.03E-15 Std. Dev. = 0.990 N = 465

Frequency

60

40

20

0

-2

0

2

Regression Standardised Residual

4

Expected Cumulative Probability

1.0

0.8

0.6

0.4

0.2

0.0 0.0

0.2

0.4

0.6

0.8

Observed Cumulative Probability

1.0

Regression Standardised Predicted Value

2

1

0

Figure A 11. Frequency distribution (top), PP plot (middle) and QQ plot (bottom) of the standardised residuals that resulted from the non-spatial multiple regression model. The lines show a normal distribution.

-1

-2

-3

-2

0

2

4

Regression Standardised Residual

173

Appendix 17. Frequency distribution of the standardised residuals of the GWR 80.0

Mean = .0341 Std. Dev. = 0.99981 N = 465

Frequency

60.0

40.0

20.0

0.0

-4

-2

0

2

Standardised residuals

4

6

Figure A 12. Frequency distribution of the standardised residuals that resulted from the spatial GWR model. The line shows a normal distribution.

Appendix 18. Overview of the variables and the data selected Table A 23. Overview of the variables and the data selected. Data layer (proxy)

Year (resolution)

Quality data

Data and source

AGB (t ha )

2008

50m

Altitude (m) Slope (%) Soil type

2012 2012 1999

90 m 90 m Polygon

Land allocation

2009

Polygon

AGB map (Quiñones et al., 2011)and areas to be avoided for oil palm expan sion based on greenhouse gas emission criteria. Available map data either has to o low spatial detail (e.g. 500-1000m SRTM-DEM NASA (NASA, 2012) SRTM-DEM NASA (NASA, 2012) Geo-corrected and gap filled reproduction of 1:250,000 RePPProT land systems map (Consortium to Revise the HCV Toolkit for Indonesia, 2008) (Ministry of Forestry Indonesia, year unknown)

-1

Distance to the nearest (m) Fire 2000-2008 Road 2003

MODIS Point Polyline

River City

Polyline Point

174

2009 2012

(NASA/LANCE – FIRMS, 2011) Developed by Bakosurtanal; prepared by WRI for the Interactive Atlas for Indonesia’s Forests (Bakosurtanal, 2009) Idem (Bakosurtanal, 2009) (CIESIN, 2012)

Appendix 19. Data sources Aboveground biomass Quiñones, M.J., Schut, V., Wielaard, N. & Hoekman, D. (2011) Above Ground Biomass map Kalimantan 2008 – Final report. SarVision Wageningen, 34p. Altitude and slope NASA, 2012. SRTM-DEM (http://www2.jpl.nasa.gov/srtm/) Soil type Consortium to Revise the HCV Toolkit for Indonesia (2008) Toolkit for identification of high conservation values in Indonesia. Jakarta, Indonesia. Digital Appendix 12. Ecosystem proxy shapefiles for Kalimantan ver 1.0. Land allocation Ministry of Forestry Indonesia (year unknown). Kawasan Hutan (Forest estate) land use maps, General Direktorat of Planning, Ministry of Forestry; downloaded from http://appgis.dephut.go.id/ appgis/kml.aspx. Processed and provided by Greenpeace. Prepared by the World Resources Institute (2012). Downloaded from http://www.wri.org/applications/maps/forest-cover-analyzer/ Fire NASA/LANCE – FIRMS, 2011. MODIS Hotspot/Active Fire Detections. Data set. Acquired on 17-092012 online http://earthdata.nasa.gov/data/nrtdata/firms Main cities CIESIN, 2012 (Center for International Earth Science Information Network), Columbia University; International Food Policy Research Institute (IPFRI); The World Bank; Centro Internacional de Agricultura Tropical (CIAT): http://sedac.ciesin.columbia.edu/gpw/ Main roads and rivers Bakosurtanal, 2009. Bakosurtanal, the Indonesian National Coordinating Agency for Surveys and Mapping (http://www.bakosurtanal.go.id). Data available in Minnemeyer et al. (2009). Interactive Atlas of Indonesia’s Forests CD-ROM. Washington, DC: World Resources Institute. Prepared by the World Resources Institute (2012).

175

Appendix 20. Categorisation of the variables Categorisation of altitude Lowlands 0 – 750m Midlands 750 – 1500m Highlands > 1500m Classification of soil type New class Original classes (Symbol_LS) Karst: GBJ, KPR, OKI Peat: BRH, GBT, KLR, MDW Volcanic: BTA, BTK, LPN, SMD, TBA Other: BKN, BLI, BPD, BRW, GDG, HJA, JLH, KHY, KJP, LHI, LNG, LWW, MGH, MPT, MTL, PDH, PKU, PMG, PST, PTG, RGK, SBG Categorisation of fire into burned and non-burned areas burned ≤ 500m hotspot unburned > 500m hotspot

176

Appendix 21. Estimation of AGB In 2011 and 2012, a ground data collection team measured the diameter at breast height (DBH, cm) and the total height (m, with hypsometer and clinometer) for each tree in 42 stands over a range of undisturbed and degraded forest types and in 6 plots in oil palm plantations with different age. In each stand, two adjacent (large) plots of 100 m x 10 m (2 x 0.1 ha = 0.2 ha), all trees with DBH ≥10 cm were measured and in two (small) subplots of 50 m x 3 m (2 x 0.03 ha), all trees with 1≤DBH40-70% Burnt forest (40-70% canopy cover) (8 plots, 0.2 ha) Degraded forest closed canopy (6 plots, 0.2 ha) Peat swamp forest logged Lowland dipterocarp forest logged Low disturbance sites Medium disturbance sites Disturbed forest reserves Disturbed forest reserves

No/low disturbance 14 plots 7 plots (0.2 ha) AGB stored in trees

Forest – closed CC >70 % Unlogged forest High forest closed canopy 1 & 2 Natural forest 2005/2008 Intact forest – mineral soils Dipterocarp forest Lowland forest (primary) Montane forest (primary) Peat swamp forest (primary) Heath forest (primary) Dipterocarp forest, various soil types Lowland dipterocarp forest – unlogged/ little logging Primary dipterocarp forest Primary dipterocarp forest Primary dipterocarp forest

Kutai Kartanegara, East Kalimantan Kutai Kartanegara, East Kalimantan Central Kalimantan Central Kalimantan East Kalimantan East Kalimantan Tangkulap, Sabah Deramakot, Sabah

190.2 159.9 230.8 220.0 106.0 245.0 335.8 205.5 69.7

Mean St.dev.

Sebulu, East Kalimantan Kinabalu, Sabah Lambir, Sarawak

486.0 511.5 520.0 411.0 123.4 156.3

Sabah Kutai Kartanegara, East Kalimantan East Kalimantan Borneo Borneo Borneo Borneo Borneo Borneo Borneo Central Kalimantan

Location

353.0 185.8 221.9 457.1 272.4 477.0 461.0 348.7 342.7 570.0 547.1

AGB (Mg/ha or t/ha

13-14 yrs after fire

Mean St.dev.

Notes/Assumptions

Land use/cover type

Table A 24. AGB (Mg ha-1 or t ha-1) for a variety of LULC types in Indonesia found in the literature.

Laan et al., unpublished data (Kronseder et al., 2012) (Kronseder et al., 2012) (Toma et al., 2005) (Toma et al., 2005) (Langner et al., 2012) (Langner et al., 2012)

Laan et al., unpublished data

(Yamakura et al., 1986) (Aiba S. and K., 1999) (Yamakura et al., 1986)

(Morel et al., 2011) Laan et al., unpublished data (Simbolon et al., 2010) (Slik et al., 2010a) (Katayama et al., 2013) (Budiharta et al., 2014) (Budiharta et al., 2014) (Budiharta et al., 2014) (Budiharta et al., 2014) (Paoli et al., 2008) (Kronseder et al., 2012)

Reference

Appendix 22. Overview of AGB for LULC types found in the literature

179

Rubber Rubber – average in literature Rubber agroforests – jungle rubber 15-40 yrs (lowland) Rubber – monoculture 6-16 yrs (lowland Rubber – monoculture Rubber – agroforest

Mixed cropland Cropland irrigated Cultivated lands secondary vegetation

Shrubs, grass and cleared Shrub Grass Imperata cylindrica grassland Imperata cylindrica grassland – min-max

Very open canopy CC = >10-40% Logged forest (various ages) (9 plots) Burnt forest (10-40% canopy cover) (16 plots, 0.2 ha) Degraded forest open canopy (5 plots, 0.2 ha) Regenerating forest 2-5 yrs after fire Secondary forest 2003-2008 (after 97/98 fires) Burnt forest High disturbance sites

Land use/cover type

Mean St.dev.

Total tree biomass

Total tree biomass

Mean St.dev.

Mean St.dev.

Mean St.dev.

AGB stored in trees

1970 13-14 yrs after fire

Notes/Assumptions

50.0 88.2 353.2 129.7 128.4

42.0 115.0

15.0 71.0 43.0 39.6

Sumatra Indonesia Indonesia

Sumatra Sumatra

Indonesia Indonesia

Indonesia Indonesia East Kalimantan Indonesia

Kutai Kartanegara, East Kalimantan East Kalimantan East Kalimantan Indonesia East Kalimantan

88.8 15.4 148.1 73.0 27.0 102.2 88.0 68.8 6.8 21.9 13.4 27.7 28.1

Sabah Kutai Kartanegara, East Kalimantan

Location

275.2 88.0

AGB (Mg/ha or t/ha

(Kotowska et al., 2015) (references in Khasanah et al., 2015) (references in Khasanah et al., 2015)

(Kotowska et al., 2015) (Kotowska et al., 2015)

(Prasetyo, 2000) (Prasetyo, 2000)

(references in Khasanah et al., 2015) (references in Khasanah et al., 2015) (Syahrinudin, 2005) (Germer and Sauerborn, 2008)

Laan et al., unpublished data (Hiratsuka and Toma, 2006) (Simbolon et al., 2010) (Prakoso, 2006) (Toma et al., 2005)

(Morel et al., 2011) Laan et al., unpublished data

Reference

180 Notes/Assumptions

Mean St.dev. Mean St.dev.

Settlement

Mean St.dev.

Immature, 2 plots Mature, 4 plots AGB stored in trees AGB stored in trees

Mining

Pulpwood plantation Timber plantation Timber plantation Acacia mangium 3 yr old Acacia mangium 5 yr old Acacia mangium 1 yr old (3 plots) Acacia mangium 4 yr old (3 plots) Acacia mangium 9 yr old (3 plots) Acacia mangium 6 yr old Acacia mangium 7 yr old

Oil palm Oil palm plantation (6 plots, 0.2 ha, 11 m average H) Oil palm – average in literature Oil palm – monoculture 8-15 yrs (lowland) Oil palm (3 plots) Immature Oil palm (17 plots) Mature Oil palm – 25 yrs rotation – mineral soil, company Oil palm – 25 yrs rotation – peat, company Oil palm – 25 yrs rotation – mineral, smallholder Mean St.dev.

Land use/cover type

27.7 28.1

14.0 7.6

14.5 115.5 61.0 220.9 16.8 64.8 139.2 63.2 96.6 88.1 64.7

104.0 40.0 2.4 52 84.0 80.1 75.5 54.2 34.5

10.4

AGB (Mg/ha or t/ha

Sabah Sabah East Kalimantan East Kalimantan East Kalimantan East Kalimantan East Kalimantan East Kalimantan East Kalimantan

Sumatra Sabah Sabah Indonesia Indonesia Indonesia

Kutai Kartanegara, East Kalimantan

Location

(set equal to shrubs and grass)

(set equal to shrubs and grass)

(Morel et al., 2011) (Morel et al., 2011) (Simbolon et al., 2010) (Simbolon et al., 2010) (Syahrinudin, 2005) (Syahrinudin, 2005) (Ruhiyat, 1989) (Ruhiyat, 1989) (Ruhiyat, 1989)

(Kotowska et al., 2015) (Kotowska et al., 2015) (Morel et al., 2011) (Morel et al., 2011) (Khasanah et al., 2015) (Khasanah et al., 2015) (Khasanah et al., 2015)

Laan et al., unpublished data

Reference

Appendix 23. Species richness remaining ranges of LULC types. The different LULC types have different remaining ranges of the original plant species richness, which may vary depending on management and the growth forms present (trees, shrubs or herbs). Below we describe the criteria we used to define such remaining species fractions per LULC type. Forest: For the forest types, we assumed that all species can potentially remain; this corresponds to the maximum of the remaining range (100%). However, depending on the management and disturbance level species can be lost. We assumed that this loss is proportionally correlated to the canopy cover because in the study region canopy cover is a function of disturbance and management. (1) closed canopy forest is the natural ecosystem, it has a canopy cover of 70-100% and no or extensive management. Under these conditions, we assumed that between 90-100% of the original species richness can be retained. (2) medium open canopy forests have a cover of 40-70% and a medium disturbance level. For these forests we assumed a remaining range of 60-100%. (3) open canopy forests have a cover of 10-40% and a high disturbance level. For these forests we assumed a remaining range of 30-100%. Shrubs and Grassland: Shrubs, grassland and cleared land were aggregated into one LULC type in the PLUC scenarios. For this aggregated class we assumed that no tree species will remain. Thus, the maximum species remains is set to 28%, which corresponds to the proportion of shrubs and herbs in the total number of species. After LULC change some shrubs and herbs can potentially remain; we assumed that 25% of the maximum amount of species remained. Therefore, we set the remaining range to 7-28%. Natural rubber: This LULC type corresponds to rubber produced in agroforest and monoculture plantation systems, mostly by smallholders. Smallholder rubber and rubber agroforest systems are mixed systems with medium management and disturbance levels. Danielsen and Heegaard (1995) show that conversion of forest to rubber plantations led to simple, species-poor and less diverse animal communities. We therefore assumed that the maximum remaining percentage for all growth forms (trees, shrubs and herbs) is 70%. For the monoculture rubber plantations, however, we expected that all tree species are removed and that only shrubs and herbs can remain. However, it is also possible that all species are removed. This resulted in a remaining range of 0-70% for rubber. Pulpwood plantation: these are monoculture plantations with a medium management and disturbance level. We assumed that all tree species are removed either before plantation establishment and during cultivation, thus only the shrub and herb species remain in a pulpwood plantation. However, it is also possible that all species are removed. Therefore, the remaining range is 0-28% for pulpwood. Mixed cropland: this mixed agricultural system has a medium management level and we assume that a maximum of 50% of the species of all growth forms can remain. Again all species may be removed. Therefore the remaining range is 0-50%. Oil palm: In an oil palm plantation, all tree species are removed (Danielsen et al., 2009) and only shrubs and herbs can remain or may be removed as well. Therefore, we assumed a remaining range of 0-28%. Oil palm plantations support much fewer species than forests and often also fewer than other

181

tree crops. We could not find many studies that assessed the impacts of oil palm expansion on plant species richness (Fitzherbert et al., 2008). The study of Danielsen et al. (2009) reports the absence of forest trees, lianas, epiphytic orchids and indigenous palms in oil palm plantations. Mining: Mining in the study region is mostly for coal, and for this activity all trees and shrubs are removed. We assumed that only a maximum of half of the herb species remains. Therefore, we set the remaining range to 0-9% of the original species richness. Settlements: Settlements are highly variable and so is their species richness, e.g. due to the presence of homestead gardens. Therefore, we assumed a species remaining range of 25-50%.

182

Appendix 24. Species richness of plant families and growth forms Table A 25. The plant families (Raes et al., 2013, 2009) and their presence in the landscape and percentage of lifeforms.

Moraceae Myristicaceae Sapindaceae Ericaceae Dipterocarpaceae Lauraceae Leguminosae

Fagaceae

mulberryor fig family nutmeg family soap-berry family heather family includes Meranti hardwood laurel family legume family beech family

Total

Growth form (herb, shrub, tree)

Trees %

Trees

15%

Trees 90%, Shrubs 10% Trees, herbaceous plants or lianas Herbs, dwarf shrubs, shrubs, and trees Tree

6%

Tree, vines Herbaceous perennials (majority), small annual herbs, trees Tree

Shrubs %

Herbs Remaining Min no. Max no. % value of species of species

1%

4% 9% 21%

16% 19%

10% 71.9%

9.6%

18.6%

15%

54

123

7%

20

60

4%

14

38

9%

5

75

21%

29

177

16% 19%

29 54

132 157

10%

17

84

222

846

Table A 26. Mean AGB and standard deviations per LULC type that are derived from field data and from the literature (see Table A 24 for an overview).

Forest – closed canopy Forest – medium open canopy Forest – very open canopy Shrubs and grassland Mixed cropland Rubber (mostly smallholder) Pulpwood plantation Oil palm Mining (mostly coal) Settlement

Mean AGB (t ha-1)

Standard deviation (t ha-1)

411.0 205.5 102.2 27.7 43.0 129.7 88.1 54.2 14.0 27.7

123.4 69.7 88.0 28.1 39.6 128.4 64.7 34.5 7.6 28.1

183

Appendix 25. Descriptive statistics for AGB Table A 27. Remaining mean AGB and standard deviations under all scenarios, and the decline (%) over time. Scenario

Year

Mean AGB remaining

St.dev.

LR LuR uLR uLuR LR LuR uLR uLuR

1990 2000 2009 2020 2020 2020 2020 2030 2030 2030 2030

276.20 248.63 232.01 230.34 230.93 223.76 218.79 230.85 227.53 219.22 167.66

62.61 54.46 49.17 49.56 46.82 47.08 46.03 47.72 46.66 45.24 42.29

184

Decline 1990-2009

Decline 1990-2030

Decline 2009-2030

16% 18% 21% 39%

0% 2% 6% 28%

16%

185

1990 2000 2009 2020 2020 2020 2020 2030 2030 2030 2030


-27.573 -44.199 -45.869 -45.277 -52.442 -57.415 -45.353 -48.678 -56.986 -108.545

Mean difference 0.000* 0.000* 0.000* 0.000* 0.000* 0.000* 0.000* 0.000* 0.000* 0.000*

Significance

1990

-16.626 -18.296 -17.705 -24.870 -29.842 -17.780 -21.105 -29.413 -80.973

0.000* 0.000* 0.000* 0.000* 0.000* 0.000* 0.000* 0.000* 0.000*

Significance

2000 Mean difference

* The mean difference is significant at the 0.05 level.

Year

Scenario

-1.670 -1.078 -8.243 -13.216 -1.154 -4.479 -12.786 -64.346

Mean difference

0.955 0.991 0.004* 0.000* 0.989 0.309 0.000* 0.000*

Significance

2009

0.516

Mean difference

1.000

Significance

LR 2020

-3.401

0.593

Significance

LuR 2020 Mean difference

Table A 28. Mean differences (Tukey HSD) between AGB year-pairs and significance under all scenarios.

-4.543

Mean difference

0.291

Significance

uLR 2020

-51.130

Mean difference

0.000*

Significance

uLuR 2020

Appendix 26. Pair-wise comparisons total AGB Pair-wise comparisons showed that projected AGB between 2009 and 2030 was not significantly different under the limited growth scenarios LR (Tukey HSD=1.474, p=1.000) and LuR (Tukey HSD=4.284, p=1.000). The strongest decline occurred in the lowlands in the south-east of the study region (Figure 6.4), where the stabilising expansion of agriculture and mining was projected to occur. Under the uLR scenario, total AGB declined with ~19% to ~707 Mt between 1990-2030, mostly due to the conversion of medium open canopy forest (Figure 4.5). However, the decline in AGB was not significant in both the time periods 2009-2020 (Tukey HSD=13.765, p=0.805) and 2020-2030 (Tukey HSD=11.586, p=0.927), when additional land development was hampered by the restrictions in land zoning. This restriction in land zoning caused the decrease in AGB to occur mostly in the south-east and in the centre of the region (Figure 6.4). Under the uLuR scenario, AGB is projected to decrease with 37% to ~550 Mt (±143 Mt) between 1990-2030. This significant additional decrease in AGB (Tukey HSD=182.778, p=0.000) is mostly caused by the conversion of medium open canopy and closed canopy forest in the higher altitudes in the north-west (Figure 4.5 and Figure 6.4). The standard deviations and thus uncertainties of the total AGB estimates were highest for 1990 and decreased over time, particularly under the unrestricted scenarios (Figure 6.5). This is because the uncertainty range of AGB values for agricultural and mining land was much lower than for forests.

Appendix 27. Descriptive statistics for mean species richness

Table A 29. Remaining mean species richness and standard deviations under all scenarios, and the decline (%) over time. Scenario

Year

Mean no. species remaining

St.dev.


1990 2000 2009 2020 2020 2020 2020 2030 2030 2030 2030

406.27 388.62 364.85 360.27 359.77 324.91 312.93 359.67 357.76 318.34 196.22

44.52 46.80 49.88 49.48 49.22 48.52 46.73 48.78 49.44 47.64 45.30

186

Decline 1990-2009

Decline 1990-2030

Decline 2009-2030

11% 12% 22% 52%

1% 2% 13% 46%

10%

187

1990 2000 2009 2020 2020 2020 2020 2030 2030 2030 2030


2000

2009

LR 2020

LuR 2020

uLR 2020

uLuR 2020

-17.652 -41.426 -46.006 -46.507 -81.362 -93.339 -46.599 -48.517 -87.937 -210.048

0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 -23.774 -28.354 -28.856 -63.711 -75.687 -28.947 -30.865 -70.285 -192.397

0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 -4.580 -5.082 -39.937 -51.913 -5.174 -7.092 -46.512 -168.623

0.205 0.125 0.000 0.000 0.112 0.009 0.000 0.000 -0.593

0.999 -2.010

0.883 -6.575

0.017 -116.709

0.000

Mean SignifiMean SignifiMean SignifiMean SignifiMean SignifiMean SignifiMean Signifidifference cance difference cance difference cance difference cance difference cance difference cance difference cance

1990

* The mean difference is significant at the 0.05 level.

Year

Scenario

Table A 30. Mean differences (Tukey HSD) between species richness year-pairs and significance under all scenarios.

Appendix 28. Cultivation/plantation area and production volume of the production systems Table A 31. The land uses and estate types, current cultivation area (for oil palm, only mature areas were accounted for), proportion of mature oil palm plantations in and current production volume (in t) in NorthEast Kalimantan in 2008. Agricultural production system

Yielded product

Oil palm private estates Oil palm smallholdings Wetland rice Dryland rice Rubber smallholdings Rubber private estates Total

FFB FFB Rice Rice Rubber Rubber

34 50 n/a n/a n/a n/a

311,393 (106,000*) 93,203 (46,000*) 98,000 60,000 60,000 12,000 382,000

1,611,000 481,000 441,000 145,000 43,000 6,000 2,727,000

Pulpwood

n/a

1.2 Mha

Unknown**

HTI, industrial pulpwood plantation

% of mature plantations in Cultivation area (ha) Production volume 2008 (projected using BPS in 2008 (t) in 2008 data; BPS, 2012)

* In this figure only mature oil palm plantations are accounted for, as was calculated by the correction factor in section 2.4.1. ** The production volume and the exact proportion of pulpwood from natural forest and from the different forest management systems, including HTI, could not be collected from BPS data or the literature and is thus unknown. Source: BPS (BPS, 2012, 2009)

Table A 32. Plantation area and annual production volume of pulpwood in Indonesia in 2010, planned for 2030, and projected for 2020, and projected for North-East Kalimantan in 2020 and 2030. 2008

2010

2020

2030

Plantation area Indonesia (Mha)

Unknown

5 (Obidzinski and Dermawan, 2012)

10 (projected)

15 (planned, (Obidzinski and Dermawan, 2012))

Plantation area North-East Kalimantan projected (Mha)

1 (BPS, 2012)

1 (BPS, 2012)

3 (projected)

4 (projected)

Ratio Indonesia vs. North-East Kalimantan

3.6

Annual production pulpwood Indonesia (Million m3 yr-1)

Unknown

Unknown

242 (projected)

363 (planned (Obidzinski and Dermawan, 2012))

Annual production pulpwood NorthEast Kalimantan (Million m3 yr-1)

Unknown

Unknown

68 (projected)

102 (projected)

188

Appendix 29. Disaggregation of FFB production volume from Indonesia-Malaysia to North-East Kalimantan

6.0

Share North-East Kalimantan vs. IndoMalay

5.0

Trendline 2010-2011

4.0 3.0 2.0 1.0

10 20 11 20 12 20 13 20 14 20 15 20 16 20 17 20 18 20 19 20 20

20

20

20

09

0.0

08

ShareNorth-East Kalimantan of Indomalay

To disaggregate the FFB production volumes for North-East Kalimantan, we first compared the 2010-2011 FFB production volumes for North-East Kalimantan (based on BPS, 2009, 2012) to the production volume of Indonesia and Malaysia combined (based on FAOSTAT, 2014). This resulted in a proportion of approximately 1.7-2.1% of the production volume in North-East Kalimantan to the production volume in Indonesia-Malaysia in 2010-2011 (Figure A 13). Second, we linearly extrapolated this proportion to 2020, assuming that the proportion would continue to increase until 2020 at the same rate as between 2010-2011 (Figure A 13). Third, the extrapolated proportion of 4.9% in 2020 was multiplied by the FFB production volume under the MIRAGE BAU84 scenario for Indonesia and Malaysia Table A 33, to obtain the FFB production volume of 14.6 Mt for North-East Kalimantan for 2020 under the MIRAGE projection.

Figure A 13. Proportion (%) of the production volume North-East Kalimantan to production volume Indonesia and Malaysia.

Table A 33. Production volume of oil palm FFB and CPO for 2008 and projected extra production volume demand for 2020 in the Indonesia-Malaysia region. Production volume 2008

Indonesia-Malaysia Oil palm FFB (Mt) CPO (Mt)

Extra demand in 2020

BPS

MIRAGE

Linear

157.5 31.5

143.7 28.5

n/a n/a

189

Appendix 30. Integration of by- and co-products in the palm oil production chain OPTs are available at the moment a plantation is felled and are an abundant source of biomass in countries where oil palm is planted extensively, such as in Indonesia and Malaysia. At the time of felling, a mature oil palm plantation of about 25-30 years old generates approx. 235 m3 stems ha-1 (Hromatka and Savage, 2010). Generally, after felling, the OPTs are burned or left to decompose (Hromatka and Savage, 2010), and the plantation site can be left for forest regeneration or can be replanted. Although oil palm is a non-woody plant and differs from hardwood/softwood species in its cellulose, hemicellulose and lignin content, it can be utilised as an alternative to wood or tree-based biomass. OPTs can be used for the production of compressed wood, plywood, particleboard, laminated board (laminated veneer lumbers, LVL) (Wahab et al., 2008), fibreboard (medium density fibre, MDF), furniture, and pulp and paper (See (Hromatka and Savage, 2010; Sulaiman et al., 2012). However, the commercial utilisation is still being tested. Additionally, OPTs can be used as a nutrient source, erosion control measure, animal feed (Sulaiman et al., 2012), and biofuel and plastics (Yamada et al., 2010) (and PPT of Mori, year). One important note is that OPTs have a very high moisture, sugar and starch content, and this accelerates decomposition after felling which generates high transportation costs (Hromatka and Savage, 2010). Nonetheless, positive economic analyses of the use of OPTs are shown, e.g. by (Husin M, 2000). OPT is selected for evaluation because it has land-sparing potential, given that the plywood can be used as a replacement for soft pulpwood for non-construction materials. It could then reduce the pressure on natural forests and on pulpwood plantations that are used for soft wood production. Additionally, the rotting of trunks at the plantation site can be prevented, minimising the spread of fungus and disease. OPFs become available at pruning, harvesting or replanting time, and are thus available throughout the year (Zahari et al., 2009). OPFs are a good fibre source for feeding of ruminants (Zahari et al., 2009). Traditionally, OPFs are left at the plantation for soil conservation and erosion control, and long-term nutrient cycling (Zahari et al., 2009). At replanting, the crown can consist of about 41 fronds and can yield approximately 115 kg dry fronds/palm in total (Sulaiman et al., 2012). At an average plantation size of 113 trees per ha, this can generate approx. 13,000 kg of fronds per ha. Although this is a substantial amount of biomass, we do not expect OPFs to have land-sparing potential as a result of the current use for nutrient recycling. Therefore, OPFs are not considered in the analysis. EFBs remain after oil extraction and can be used as mulch and can reduce the need for fertilisers by over 50% in immature stands and by 5% inn mature stands (Gaskell et al., 2009). The use of EFB in Malaysia is generally very limited and can be utilised only after irradiation and culture-substrate treatments (Zahari et al., 2009). Biomethane can be produced by the fermentation of EFB because of its highly cellulosic components (Tahar, 2013). Energy produced from EFBs can be used as input for the palm oil mill (Elbersen et al., 2013). The production of EFBs increases with the production of CPO. However, because EFBs are often used as an energy source in the mill, no land-sparing potential is expected, and therefore these were not included in the evaluation. POME is the discharge from CPO extraction in the mill (Zahari et al., 2009) and, if discharged untreated, is considered harmful for the environment because of its high organic content (Rupani et al., 2010). However, this high organic content is also what makes more optimised use and treatment of POME beneficial (Yusoff and Hansen, 2007). POME can be combined with palm kernel cake (PKC) and OPF to provide a cost-effective and complete ration for feeding ruminant livestock (Zahari et al., 2009). Additionally, POME may be sustainably reused as a fermentation substrate in the production of various metabolites, fertilisers, and animal feeds (Gaskell et al., 2009; Wu et al., 2009). The treatment of

190

POME in open ponds results in high amounts of CO2 and methane being emitted to the atmosphere. The methane in POME, however, can be collected when POME is treated in closed anaerobic digesters and used as a source of biogas for the production of electricity for the palm oil mill or, if surplus electricity is generated, for the grid (Elbersen et al., 2013; Wicke et al., 2008a). Although POME has a list of potential utilisations, it is assumed to have no land-sparing potential and is therefore not accounted for in this study. PKO is co-produced in the palm oil mill and the production volume is about 10% of the CPO volume produced. PKO is currently fully in use and processed in the commercial cooking industry and the oleochemical industry; it is also suitable for the production of PKO biodiesel. Additional PKO production would not reduce land requirements, however, the projected increasing palm oil production that results from the analyses in this paper would generate a 10% additional production of PKO. As it can replace other oils, additional PKO production can reduce land requirements and thus has land-sparing potential. We recommend to estimate this potential in further analyses.

Appendix 31. Potential surplus land generated by the utilisation of oil palm trunks The land-sparing potential generated by utilising OPT for the period 2008-2020 were calculated by taking the following steps: 1. Estimate the cultivation area to be available for clearing and available for replanting between 2008 and 2020 (Oil palm areareplanted). To do so, we extrapolated the cultivation area of oil palm from 2004-2011 (BPS, 2012, 2009) and 1998 (Casson, 1999) to 1995, as the actual data could not be obtained from BPS or other sources (See Figure A 14). The extrapolation of the data shows that by 1995 approx. 25,000 ha of oil palm have been planted in North-East Kalimantan (Figure A 14). We assume that by 2020, 25 years later, this area is ready for felling and replanting, independent of the growth of the palm oil demand. 2. Define the proportion of the cultivation area at which felling will take place for the utilisation of OPT for plywood production in the three scenarios low, medium and high. In the baseline scenario, we assumed that no OPT will be utilised for the production of plywood by 2020, because no evidence was found that smallholders and companies currently utilise OPT in the study area. Under the low scenario, we assumed a minimum, but just above-baseline (10%), increase in the utilisation of by/co-products, in this case OPT. Under the medium scenario, we assumed a 40% higher utilisation of by/co-products, and under the high scenario, we assumed a 70% higher utilisation of by/co-products. 3. Estimate the total volume of OPTs that can be generated from a cultivation area to be cleared and replanted (stem yieldOPT). According to Hromatka and Savage (2010)., a mature oil palm plantation can generate approx. 235 m3 OPT ha-1 25 yr-1. 4. Estimate percentage of amount of stems that would be suitable for plywood production for nonconstruction materials (fractionsuitable). 5. Estimate the average annual yield of a HTI pulpwood plantation (stem yieldHTI). Based on the literature, this is assumed to be 25 m3 pulpwood ha-1 yr (Barry, 2002) and 175 m3 pulpwood ha-1 for a 7-years rotation cycle. 6. Calculated the land-sparing potential that can be generated if the HTI pulpwood plantation would not have to be planted to supply this amount of wood. This is done by applying Equation A 1.

191

Cultivation area (Mha)

0,9 0,8

Oil palm cultivation area

0,7

trendline

y = 1847.7e0.2074x R2 = 0.98013

0,6 0,5 0,4 0,3 0,2 0,1

19 83 19 84 19 85 19 86 19 87 19 88 19 89 19 90 19 91 19 92 19 93 19 94 19 95 19 96 19 9 19 7 98 19 99 20 00 20 01 20 02 20 0 20 3 04 20 05 20 06 20 07 20 08 20 09 20 10 20 11

0,0

Figure A 14. Oil palm cultivation area in North-East Kalimantan extrapolated (year* and trendline) based on data from 1998 (Casson, 1999) and 2004-2011 (BPS, 2012, 2009).

Equation A 1 Land-sparing potential

Land – sparing potential =

Oil palm areafelled * Stem yieldOPT * fractionsuitable * RROPT – HTI wood Stem yieldHTI

Appendix 32. Improvement of the rice and palm oil production chain efficiency Table A 34. CPO (Oil, palm) and FFB (Oil, palm fruit) production volumes and the calculated oil extraction rates, for Indonesia and Malaysia. FAO product definition

2008

2009

2010

2011

2012

Indonesia

Oil, palm Oil, palm fruit OER

17,539,788 85,000,000 21

19,324,293 90,000,000 21

21,958,120 97,800,000 22

23,096,541 1.05E+08 22

26,900,000 1.13E+08 24

Malaysia

Oil, palm Oil, palm fruit OER

17,734,441 88,672,000 20

17,564,937 87,825,000 20

16,993,717 84,965,000 20

18,911,520 94,557,600 20

18,785,030 97,700,000 19

Source: FAOSTAT (2014).

192

Table A 35. Rice production volumes and losses in Indonesia and nearby countries and regions between 2008 and 2011. 2008

2009

2010

2011

Indonesia

Production volume (t) Post-harvest losses % of losses

60,251,000 4,674,000 7.8

64,399,000 4,963,000 7.7

66,469,000 5,144,000 7.7

65,741,000 5,172,000 7.9

Malaysia

Production volume(t) Post-harvest losses % of losses

2,353,000 178,000 7.6

2,511,000 195,000 7.8

2,465,000 186,000 7.5

2,576,000 198,000 7.7

South-east Asia

Production volume(t) Post-harvest losses % of losses

192,600,000 13,2021,000 6.9

197,777,000 13,611,000 6.9

204,305,000 14,248,000 7.0

202,942,000 14,400,000 7.1

Lao People’s Democratic Republic

Production volume(t)

2,970,000

3,145,000

3,071,000

3,066,000

148,000 5.0

189,000 6.0

184,000 6.0

184,000 6.0

Post-harvest losses % of losses Source: FAOSTAT (2014).

193

Appendix 33. Input information mitigation measures Table A 36. Selected input settings of the WRI Suitability Mapper, indicating suitability and oil palm crop criteria under the low, medium and high scenario (for data descriptions, data layer resolutions and data sources see WRI). Scenarios and settings

Low

Medium

High

Optimal growth conditions

Drainage and soil preparation measures needed to obtain similar or higher yields

Drainage and soil preparation measures needed to obtain similar or higher yields

Land cover (Forest, HCV areas, plantations and agricultural lands excluded)

Grassland/shrub

Peat depth (cm)

0 (to minimise C emissions)

Conservation area buffer (m)

> 1000

Water resources buffer (m)

> 100

Altitude (m) (increasing climatic restrictions on cultivation above 200-300m1,2

0-300

0-500

0-750

Slope (%) (Apply terracing if slope >10%)2

0-10

0-20

0-30

Rainfall (mm yr-1)

1500-3000 (No growth limitations1)

1500-4000 (None to moderate growth limitations1, drainage needed)

1500-5000 (None to moderate growth limitations1, drainage needed)

Soil drainage

Well, moderately well, excessive

Well, moderately well, excessive

Poor, imperfect to excessive (drainage needed)

Soil depth (cm)

> 50

> 50

> 50

Soil acidity (pH)

pH 4-6 (No measures to acidify soil are needed)

4-7.3 (Excessively acid- neutral, measures to acidify the soil may be needed)

4-7.3 (Excessively acid- neutral, measures to acidify the soil may be needed)

Soil type (rock and Histosols excluded)

Inceptisol, Oxisol, Alfisol, Ultisol, Spodosol, Entisol

1. Source: Harris et al. (2013) 2. Source: Sheil et al. (2009)

194

Table A 37. Extra CPO production and rice loss prevented (t) and the land-sparing potentials (ha) in 2020 by improving the OER and minimising rice losses under the selected scenarios compared to the baseline. Palm oil production Extra CPO in 2020 (t)* Land sparing in 2020 (ha)*

Low

Medium

High

MIRAGE Linear MIRAGE Linear

135,000 228,000 50,000 85,000

202,000 342,000 73,000 124,000

269,000 456,000 95,000 162,000

MIRAGE

4,000

7,000

10,000

Linear

4,000

6,000

9,000

MIRAGE

16,000

26,000

36,000

14,000

23,000

32,000

Rice production Rice loss prevented 2020 (t) Land sparing potential in 2020 (ha)**

Linear -1

* based on the OERs (21-22%) and FFB yields (~15-16 t ha ) under the low-high scenarios. ** based on the baseline yields for wetland and dryland rice for 2020 (~ 4-4.5 t ha-1).

Table A 38. Amount of suitable and available under-utilised land for expansion of the commodities in NorthEast Kalimantan. Setting

Low

Medium

High

Estimated suitable land in North-East Kalimantan according to the Suitability Mappera (Mha) Estimated suitable and available land in North-East Kalimantan according to the Suitability Mapper and accounting for ~40% of the areab (Mha)

1.8

2.2

2.4

0.7

0.9

1.0

a) Local land claims/rights/interests are unknown and need to be determined through field assessments. b) The portion of suitable and available land area that is not already in use by or of interest to local communities was estimated to be ~40% for West Kalimantan (Gingold et al., 2012). The same percentage was applied here for North-East Kalimantan.

195

Table A 39. Estimated land-sparing potential (ha) of the measures under the two expansion projections. Projections and measures

Commodities

Low

Medium

High

Palm oil Rice Rubber HTI pulpwood Palm oil Palm oil Rice

225,500 15,400 13,000 287,200 3,400 84,700 14,100

397,800 27,800 24,200 417,400 13,400 124,100 23,000

423,700 29,000 24,500 539,600 23,500 161,700 31,800

Low

Medium

High

118,100 17,700 7,300 213,800 3,400 50,000 16,100

208,400 31,900 13,500 310,900 13,400 73,200 26,200

222,000 33,300 13,700 401,900 23,500 95,400 36,300

Low 640,000 423,100

Medium 1,014,300 664,200

High 1,210,300 802,600

Linear projection Yield improvement

Chain integration Chain efficiency – increased OER Chain efficiency – reduced rice loss MIRAGE-based projection Yield improvement

Chain integration Chain efficiency – increased OER Chain efficiency – reduced rice loss

Palm oil Rice Rubber HTI pulpwood Palm oil Palm oil Rice

Total land-sparing potential 2020 Linear projection MIRAGE-based projection

Table A 40. Land-use demand in 2020 with and without implementation of the measures. Land-use demand 2020 Linear projection MIRAGE-based projection

Without measures Baseline 2,712,0 1,289,200

With implementation of measures Low 2,072,100 866,100

Medium 1,697,800 625,000

High 1,501,700 486,600

Table A 41. Difference between i) available and suitable land and ii) the land-use demand under different growth projections and scenarios for implementing the measures. Negative figures indicate a land shortage and positive figures show that the land-use demand is lower than the suitable and available land area. Baseline

Low

Medium

High

-1,986,400 -1,814,400 -1,737,500

-1,346,400 -1,174,500 -1,097,600

-972,100 -800,100 -723,200

-776,000 -604,100 -527,200

-563,500 -391,600 -314,600

-140,400 31,500 108,500

100,700 272,600 349,500

239,100 411,000 488,000

Linear projection Low Medium High MIRAGE-based projection Low Medium High

196

10 References Abood, S.A., Lee, J.S.H., Burivalova, Z., Garcia-Ulloa, J., Koh, L.P., 2014. Relative contributions of the logging, fiber, oil palm, and mining industries to forest loss in Indonesia. Conserv. Lett. 0, 1 – 10. doi:10.1111/conl.12103. Agus, F., Runtunuwu, E., June, T., Susanti, I., Komara, H., Syahbuddin, H., Las, I., van Noordwijk, M., 2009. Carbon dioxide emission in land use transitions to plantation. J. Penelit. dan Pengemb. Pertan. 28, 119-126. Ahrends, A., Hollingsworth, P.M., Ziegler, A.D., Fox, J.M., Chen, H., Su, Y., Xu, J., 2015. Current trends of rubber plantation expansion may threaten biodiversity and livelihoods. Glob. Environ. Chang. 34, 48 – 58. doi:10.1016/j. gloenvcha.2015.06.002. Aiba S., Kitayama, K. 1999. Structure, composition and species diversity in an altitude-substrate matrix of rain forest tree communities on Mount Kinabalu, Borneo. Plant Ecol. 140, 139 – 157. Allison, P., 1977. Testing for interaction in multiple regression. Am. J. Sociol. 83, 144-153. Alves, L.F., Vieira, S.A., Scaranello, M.A., Camargo, P.B., Santos, F.A.M., Joly, C.A., Martinelli, L.A., 2010. Forest structure and live aboveground biomass variation along an elevational gradient of tropical Atlantic moist forest (Brazil). For. Ecol. Manage. 260, 679 – 691. doi:10.1016/j.foreco.2010.05.023. Anderson, Z.R., Kusters, K., McCarthy, J., Obidzinski, K., 2016. Green growth rhetoric versus reality: Insights from Indonesia. Glob. Environ. Chang. 38, 30 – 40. doi:http://dx.doi.org/10.1016/j.gloenvcha.2016.02.008. Angelsen, A., 2009. Realising REDD+: National strategy and policy options, Realising REDD+: National strategy and policy options. doi:361. Angelsen, A., 2007. Forest Cover Change in Space and Time: Combining the von Thünen and Forest Transition. World Bank Policy Res. Work. Pap. 43. World Bank, Washington DC, US. Angelsen, A., Kaimowitz, D., 2001. Agricultural Technologies and Tropical Deforestation, Agricultural Technologies and Tropical Deforestation. CABI Publishing, CIFOR, Wallingford, UK. doi:10.1079/9780851994512.0000. Angelsen, A., Rudel, T.K., 2013. Designing and implementing effective REDD + policies: A forest transition approach. Rev. Environ. Econ. Policy 7, 91 – 113. doi:10.1093/reep/res022. Ankersen, J., Grogan, K., Mertz, O., Fensholt, R., Castella, J.C., Lestrelin, G., Nguyen, D.T., Danielsen, F., Brofeldt, S., Rasmussen, K., 2015. Vietnam’s Forest Transition in Retrospect: Demonstrating Weaknesses in Business-asUsual Scenarios for REDD+. Environ. Manage. 55, 1080 – 1092. doi:10.1007/s00267-015-0443-y. Arets, E.J.M.M., Meer, P.J. Van Der, Verwer, C.C., Hengeveld, G.M., Tolkamp, G.W., Nabuurs, G.J., Oorschot, M. Van, 2011. Global Wood Production – Assessment of industrial round wood supply from forest management systems in different global regions. Alterra, Wageningen, the Netherlands. Asner, G.P., Flint Hughes, R., Varga, T.A., Knapp, D.E., Kennedy-Bowdoin, T., 2008. Environmental and Biotic Controls over Aboveground Biomass Throughout a Tropical Rain Forest. Ecosystems 12, 261 – 278. doi:10.1007/ s10021-008-9221-5. Asner, G.P., Powell, G.V.N., Mascaro, J., Knapp, D.E., Clark, J.K., Jacobson, J., Kennedy-Bowdoin, T., Balaji, A., Paez-Acosta, G., Victoria, E., Secada, L., Valqui, M., Hughes, R.F., 2010. High-resolution forest carbon stocks and emissions in the Amazon. Proc. Natl. Acad. Sci. U. S. A. 107, 16738 – 42. doi:10.1073/pnas.1004875107. Baker, T., Phillips, O., Malhi, Y., 2004. Variation in wood density determines spatial patterns in Amazonian forest biomass. Glob. Chang. Biol. 10, 545 – 562. doi:10.1111/j.1529-8817.2003.00751.x. Bakker, L., Fristikawati, Y., 2014. Permasalahan Kehutanan di Indonesia dan Kaitannya dengan Perubahan Iklim Serta REDD+. Penerbit Pohon Cahaya, Yogyakarta, Indonesia. Bakker, L., Reerink, G., 2015. Indonesia’s land acquisition law: Towards effective prevention of land grabbing?, in: Carter, C., Harding, A. (Eds.), Land Grabs in Asia: What Role for the Law? London/New York: Routledge, pp. 83-99.

197

Bakosurtanal, 2009. Bakosurtanal, the Indonesian National Coordinating Agency for Surveys and Mapping (http:// www.bakosurtanal.go.id). Data available in Minnemeyer et al. (2009). Interactive Atlas of Indonesia’s Forests CD-ROM. World Resources Institute, Washington DC, US. Barbier, E.B., Burgess, J.C., Grainger, A., 2010. The forest transition: Towards a more comprehensive theoretical framework. Land use policy 27, 98 – 107. doi:10.1016/j.landusepol.2009.02.001. Barry, K., 2002. Heartrots in plantation hardwoods in Indonesia and Australia. Australian Centre for International Agricultural Research, Canberra, Australia. Basuki, I., Sheil, D., 2005. Local Perspectives of Forest Landscapes. A preliminary evaluation of land and soil, and their importance in Malinau, East Kalimantan, Indonesia. Center for International Forestry Research (CIFOR), Bogor, Indonesia. Batjes, N.H., 2012. ISRIC-WISE derived soil properties on a 5 by 5 arc-minutes global grid (ver. 1.2). Report 2012/01.ISRIC – World Soil Information, Wageningen (with data set, available at www.isric.org). Benhin, J., 2006. Agriculture and deforestation in the tropics: a critical theoretical and empirical review. AMBIO A J. Hum. Environ. 35, 9 – 16. Berry, N.J., Phillips, O.L., Lewis, S.L., Hill, J.K., Edwards, D.P., Tawatao, N.B., Ahmad, N., Magintan, D., Khen, C. V., Maryati, M., Ong, R.C., Hamer, K.C., 2010. The high value of logged tropical forests: lessons from northern Borneo. Biodivers. Conserv. 19, 985 – 997. doi:10.1007/s10531-010-9779-z. Bini, L.M., Diniz-Filho, J.A.F., Rangel, T.F.L.V.B., Akre, T.S.B., Albaladejo, R.G., Albuquerque, F.S., Aparicio, A., Araújo, M.B., Baselga, A., Beck, J., Isabel Bellocq, M., Böhning-Gaese, K., Borges, P.A.V., Castro-Parga, I., Khen Chey, V., Chown, S.L., de Marco, Jr, P., Dobkin, D.S., Ferrer-Castán, D., Field, R., Filloy, J., Fleishman, E., Gómez, J.F., Hortal, J., Iverson, J.B., Kerr, J.T., Daniel Kissling, W., Kitching, I.J., León-Cortés, J.L., Lobo, J.M., Montoya, D., Morales-Castilla, I., Moreno, J.C., Oberdorff, T., Olalla-Tárraga, M.Á., Pausas, J.G., Qian, H., Rahbek, C., Rodríguez, M.Á., Rueda, M., Ruggiero, A., Sackmann, P., Sanders, N.J., Carina Terribile, L., Vetaas, O.R., Hawkins, B.A., 2009. Coefficient shifts in geographical ecology: an empirical evaluation of spatial and non-spatial regression. Ecography (Cop.). 32, 193 – 204. doi:10.1111/j.1600-0587.2009.05717.x. BPS, 2015. Badan Pusat Statistik. Dalam Angka 2015. http://www.bps.go.id/Date accessed: 20 May 2016. BPS, 2014. Badan Pusat Statistik. Kaltim Dalam Angka 2014. http://kaltim.bps.go.id/index.php#accordion-daftarsubjek3. Date accessed: 15 July 2015. BPS, 2012. Badan Pusat Statistik. Kaltim Dalam Angka 2012. http://kaltim.bps.go.id/web/KDA12/. Date accessed: 14 March 2013. BPS, 2009. Badan Pusat Statistik. Kaltim Dalam Angka 2009. http://kaltim.bps.go.id/?p=60. Date accessed: 14 March 2013. BPS, 2008. Badan Pusat Statistik. Kaltim Dalam Angka 2008. http://kaltim.bps.go.id/index.php/publikasi/61. Date accessed: 14 March 2013. Brinkman, M.L.J., Wicke, B., Gerssen-Gondelach, S., van der Laan, C., Faaij, A., 2015. Methodology for assessing and quantifying ILUC prevention options: ILUC Prevention Project – Methodology report. Utrecht University, Utrecht, the Netherlands. 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+? For. Policy Econ. 18, 30 – 37. doi:10.1016/j.forpol.2011.09.004. Broich, M., Hansen, M., Stolle, F., Potapov, P., Margono, B.A., Adusei, B., 2011a. Remotely sensed forest cover loss shows high spatial and temporal variation across Sumatera and Kalimantan, Indonesia 2000 – 2008. Environ. Res. Lett. 6, 1 – 9. doi:10.1088/1748-9326/6/1/014010. Broich, M., Hansen, M.C., Potapov, P., Adusei, B., Lindquist, E., Stehman, S. V., 2011b. Time-series analysis of multi-resolution optical imagery for quantifying forest cover loss in Sumatra and Kalimantan, Indonesia. Int. J. Appl. Earth Obs. Geoinf. 13, 277 – 291. doi:10.1016/j.jag.2010.11.004.

198

Brown, S., 1997. Estimating Biomass and Biomass Change of Tropical Forests: A Primer. FAO For. Pap. 134, 55. Brown, S., Iverson, L.R., Prasad, A., Liu, D., 1993. Geographical distributions of carbon in biomass and soils of tropical Asian forests. Geocarto Int. 4, 45 – 59. 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. Glob. Environ. Chang. 31, 285 – 295. doi:10.1016/j.gloenvcha.2015.02.004. Budiharta, S., Slik, F., Raes, N., Meijaard, E., Erskine, P.D., Wilson, K.A., 2014. Estimating the Aboveground Biomass of Bornean Forest. Biotropica 0, 1 – 5. doi:10.1111/btp.12132. Budiman, A., Setiabudi, Hultera, P., Ginanjar, A., 2014. Heart of Borneo Land Cover Dynamic 1990, 2000, 2009, 2013 Kutai Barat Landscape Section Methodology and Technical Report. Jakarta, Indonesia. Budiman, A., Smit, H.H., 2011. A path towards a Green Economy in the Heart of Borneo. WWF-Indonesia, Jakarta, Indonesia. Burney, J.A., Davis, S.J., Lobell, D.B., 2010. Greenhouse gas mitigation by agricultural intensification. Proc. Natl. Acad. Sci. 107, 12052 – 12057. doi:10.1073/pnas.0914216107/-/DCSupplemental.www.pnas.org/cgi/ doi/10.1073/pnas.0914216107. 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. Proc. Natl. Acad. Sci. 112, 1328 – 1333. doi:10.1073/pnas.1412514112. Busch, J., Grantham, H.S., 2013. Parks versus payments: reconciling divergent policy responses to biodiversity loss and climate change from tropical deforestation. Environ. Res. Lett. 8, 34028. doi:10.1088/1748-9326/8/3/034028. Butler, R.A., Koh, L.P., Ghazoul, J., 2009. REDD in the red: palm oil could undermine carbon payment schemes. Conserv. Lett. 2, 67 – 73. doi:10.1111/j.1755-263X.2009.00047.x. Carlson, K.M., Curran, L.M., Asner, G.P., Pittman, A.M., Trigg, S.N., Marion Adeney, J., 2012a. Carbon emissions from forest conversion by Kalimantan oil palm plantations. Nat. Clim. Chang. 3, 283 – 287. doi:10.1038/ nclimate1702. Carlson, K.M., Curran, L.M., Ratnasari, D., Pittman, A.M., Soares-Filho, B.S., Asner, G.P., Trigg, S.N., Gaveau, D.A., Lawrence, D., Rodrigues, H.O., 2012b. Committed carbon emissions, deforestation, and community land conversion from oil palm plantation expansion in West Kalimantan, Indonesia. Proc. Natl. Acad. Sci. U. S. A. 109, 7559 – 64. doi:10.1073/pnas.1200452109. Carrasco, L.R., Larrosa, C., Milner-Gulland, E.J., Edwards, D.P. 2014. A double-edged sword for tropical forests. Science (38). 346, 38 – 41. Carriquiry, M., Dong, F., Du, X., Elobeid, A., Fabiosa, J., Hart, C., Hayes, D.J., Kovarik, K., Mulik, K., Wailes, E., Chavez, E., Moreira, M., Binfield, J., Brown, D.S., Gerlt, S., Madison, D., Meyer, S., Meyers, W.H., Thompson, W., Westhoff, P., Wilcox, L., Womack, A.W., 2010. FAPRI 2010 U.S. World Agricultural Outlook. Ames, Iowa. Casson, A., 2001. Decentralisation, Forests and Estate Crops in Kutai Barat District, East Kalimantan. CIFOR, Bogor, Indonesia. Casson, A., 1999. The hesitant boom: Indonesia’s oil palm sub-sector in an era of economic crisis and political change. CIFOR, Bogor, Indonesia. Casson, A., Deddy Multastra, Y.I.K., Obidzinski, K., 2014. Large-scale plantations, bioenergy developments and land use change in Indonesia. CIFOR Working Paper No. 170. doi:10.17528/cifor/005434. Casson, A., Obidzinski, K., 2007. From New Order to Regional Autonomy: Shifting Dynamics of Illegal Logging in Kalimantan, Indonesia, in: Tacconi, L. (Ed.), Illegal Logging: Law Enforcement, Livelihoods and the Timber Trade. Earthscan, London, pp. 43 – 68.

199

Chapin, F.S., Zavaleta, E.S., Eviner, V.T., Naylor, R.L., Vitousek, P.M., Reynolds, H.L., Hooper, D.U., Lavorel, S., Sala, O.E., Hobbie, S.E., Mack, M.C., Díaz, S., 2000. Consequences of changing biodiversity. Nature 405, 234 – 42. doi:10.1038/35012241. Chave, J., Andalo, C., Brown, S., Cairns, M. A, Chambers, J.Q., Eamus, D., Fölster, H., Fromard, F., Higuchi, N., Kira, T., Lescure, J.-P., Nelson, B.W., Ogawa, H., Puig, H., Riéra, B., Yamakura, T., 2005. Tree allometry and improved estimation of carbon stocks and balance in tropical forests. Oecologia 145, 87 – 99. doi:10.1007/s00442-0050100-x. Chave, J., Condit, R., Aguilar, S., Hernandez, A., Lao, S., Perez, R., 2004. Error propagation and scaling for tropical forest biomass estimates. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 359, 409 – 20. doi:10.1098/rstb.2003.1425. Chave, J., Condit, R., Lao, S., Caspersen, J.P., Foster, R.B., Hubbell, S.P., 2003. Spatial and temporal variation of biomass in a tropical forest: results from a large census plot in Panama. J. Ecol. 91, 240 – 252. doi:10.1046/j.13652745.2003.00757.x. Chave, J., Réjou-Méchain, M., Búrquez, A., Chidumayo, E., Colgan, M.S., Delitti, W.B.C., Duque, A., Eid, T., Fearnside, P.M., Goodman, R.C., Henry, M., Martínez-Yrízar, A., Mugasha, W.A., Muller-Landau, H.C., Mencuccini, M., Nelson, B.W., Ngomanda, A., Nogueira, E.M., Ortiz-Malavassi, E., Pélissier, R., Ploton, P., Ryan, C.M., Saldarriaga, J.G., Vieilledent, G., 2014. Improved allometric models to estimate the aboveground biomass of tropical trees. Glob. Chang. Biol. 20, 3177 – 3190. doi:10.1111/gcb.12629. Chave, J., Riera, B., Dubois, M.A., 2001. Estimation of biomass in a neotropical forest of French Guiana: spatial and temporal variability. J. Trop. Ecol. 17, 79 – 96. doi:10.1017/S0266467401001055. Chazdon, R.L., Broadbent, E.N., Rozendaal, D.M.A., Bongers, F., Zambrano, A.M.A., Aide, T.M., Balvanera, P., Becknell, J.M., Boukili, V., Brancalion, P.H.S., Craven, D., Almeida-Cortez, J.S., Cabral, G.A.L., de Jong, B., Denslow, J.S., Dent, D.H., DeWalt, S.J., Dupuy, J.M., Durán, S.M., Espírito-Santo, M.M., Fandino, M.C., César, R.G., Hall, J.S., Hernández-Stefanoni, J.L., Jakovac, C.C., Junqueira, A.B., Kennard, D., Letcher, S.G., Lohbeck, M., Martínez-Ramos, M., Massoca, P., Meave, J.A., Mesquita, R., Mora, F., Muñoz, R., Muscarella, R., Nunes, Y.R.F., Ochoa-Gaona, S., Orihuela-Belmonte, E., Peña-Claros, M., Pérez-García, E.A., Piotto, D., Powers, J.S., Rodríguez-Velazquez, J., Romero-Pérez, I.E., Ruíz, J., Saldarriaga, J.G., Sanchez-Azofeifa, A., Schwartz, N.B., Steininger, M.K., Swenson, N.G., Uriarte, M., van Breugel, M., van der Wal, H., Veloso, M.D.M., Vester, H., Vieira, I.C.G., Bentos, T.V., Williamson, G.B., Poorter, L., 2016. Carbon sequestration potential of secondgrowth forest regeneration in the Latin American tropics. Sci. Adv. 2. doi:10.1126/sciadv.1501639. CIESIN, 2012. Center for International Earth Science Information Network, Columbia University; International Food Policy Research Institute (IPFRI); The World Bank; Centro Internacional de Agricultura Tropical (CIAT), New York, US. http://sedac.ciesin.columbia.edu/gpw/. Cochrane, M.A., 2003. Fire science for rainforests. Nature 421, 913 – 919. doi:10.1038/nature01437. Consortium to Revise the HCV Toolkit for Indonesia, 2008. Toolkit for identification of high conservation values in Indonesia. Jakarta, Indonesia. Digital Appendix 12. Ecosystem proxy shapefiles for Kalimantan ver 1.0. Coomes, O.T., Takasaki, Y., Rhemtulla, J.M., 2011. Land-use poverty traps identified in shifting cultivation systems shape long-term tropical forest cover. Proc. Natl. Acad. Sci. U. S. A. 108, 13925 – 30. doi:10.1073/ pnas.1012973108. Danielsen, F., Beukema, H., Burgess, N.D., Parish, F., Brühl, C.A., Donald, P.F., Murdiyarso, D., Phalan, B., Reijnders, L., Struebig, M., Fitzherbert, E.B., 2009. Biofuel Plantations on Forested Lands: Double Jeopardy for Biodiversity and Climate. Conserv. Biol. 23, 348 – 358. doi:10.1111/j.1523-1739.2008.01096.x. Danielsen, F., Heegaard, M., 1995. Impact of logging and plantation development on species diversity: a case study from Sumatra, in: Management of Tropical Forests: Towards an Integrated Perspective. Oslo, pp. 74 – 92. de Castilho, C. V., Magnusson, W.E., de Araújo, R.N.O., Luizão, R.C.C., Luizão, F.J., Lima, A.P., Higuchi, N., 2006. Variation in aboveground tree live biomass in a central Amazonian Forest: Effects of soil and topography. For. Ecol. Manage. 234, 85 – 96. doi:10.1016/j.foreco.2006.06.024.

200

DeFries, R., Rosenzweig, C., 2010. Toward a whole-landscape approach for sustainable land use in the tropics. Proc. Natl. Acad. Sci. U. S. A. 107, 19627 – 32. doi:10.1073/pnas.1011163107. 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. Proc. Natl. Acad. Sci. U. S. A. 99, 14256 – 61. doi:10.1073/pnas.182560099. Dekker, S.C., de Boer, H.J., Dermody, B.J., Wagner-Cremer, F., Wassen, M.J., Eppinga, M.B., 2015. Journal of Geophysical Research: Biogeosciences. J. Geophys. Res. Biogeosciences 120, 254 – 269. doi:10.1002/2014JG002802. Dennis, R.A., Colfer, C.P., 2006. Impacts of land use and fire on the loss and degradation of lowland forest in 19832000 in East Kutai District, East Kalimantan, Indonesia. Singap. J. Trop. Geogr. 27, 30 – 48. doi:10.1111/j.14679493.2006.00238.x. Donough, C., Witt, C., Fairhurst, T., 2010. Yield intensification in oil palm using BMP as a management tool. International Plant Nutrition Institute, Southeast Asia Program, Penang, Malaysia. Edwards, D.P., Hodgson, J.A., Hamer, K.C., Mitchell, S.L., Ahmad, A.H., Cornell, S.J., Wilcove, D.S., 2010. Wildlife-friendly oil palm plantations fail to protect biodiversity effectively. Conserv. Lett. 3, 236 – 242. doi:10.1111/j.1755-263X.2010.00107.x. Edwards, F.A., Edwards, D.P., Hamer, K.C., Davies, R.G., 2013. Impacts of logging and conversion of rainforest to oil palm on the functional diversity of birds in Sundaland. Ibis (Lond. 1859). 155, 313 – 326. doi:10.1111/ibi.12027. Eisner, R., Seabrook, L.M., McAlpine, C.A., 2016. Are changes in global oil production influencing the rate of deforestation and biodiversity loss? Biol. Conserv. 196, 147 – 155. doi:10.1016/j.biocon.2016.02.017. Ekadinata, A., Dewi, S., Nugroho, D.K., 2008. Land cover changes in different forest transition stages in Indonesia: East Kalimantan, Jambi and Lampung. World Agroforestry Centre – ICRAF, SEA Regional Office, Bogor, Indonesia. Ekadinata, A., Vincent, G., 2011. Rubber Agroforests in a Changing Landscape: Analysis of Land Use/Cover Trajectories in Bungo District, Indonesia. For. Trees Livelihoods 20, 3 – 14. doi:10.1080/14728028.2011.975669 4. Elbersen, W., Meesters, K., Bakker, R., 2013. Valorization of palm oil (mill) residues. Identifying and solving challenges. Wageningen UR, Food & Biobased Research, Wageningen, the Netherlands. Elith, J., Leathwick, J.R., 2009. Species Distribution Models: Ecological Explanation and Prediction Across Space and Time. Annu. Rev. Ecol. Evol. Syst. 40, 677 – 697. doi:10.1146/annurev.ecolsys.110308.120159. Elith, J., Phillips, S.J., Hastie, T., Dudík, M., Chee, Y.E., Yates, C.J., 2011. A statistical explanation of MaxEnt for ecologists. Divers. Distrib. 17, 43 – 57. doi:10.1111/j.1472-4642.2010.00725.x. Englhart, S., Keuck, V., Siegert, F., 2011. Aboveground biomass retrieval in tropical forests – The potential of combined X- and L-band SAR data use. Remote Sens. Environ. 115, 1260 – 1271. doi:10.1016/j.rse.2011.01.008 Ernst, C., Mayaux, P., Verhegghen, A., Bodart, C., Christophe, M., Defourny, P., 2013. National forest cover change in Congo Basin: Deforestation, reforestation, degradation and regeneration for the years 1990, 2000 and 2005. Glob. Chang. Biol. 19, 1173 – 1187. doi:10.1111/gcb.12092. Eshel, G., Shepon, A., Makov, T., Milo, R., 2014. Land, irrigation water, greenhouse gas, and reactive nitrogen burdens of meat, eggs, and dairy production in the United States. Proc. Natl. Acad. Sci. 1402183111 –. doi:10.1073/ pnas.1402183111. ESRI, 2015. ESRI. http://www.esri.com. Fairhurst, T., McLaughlin, D., 2009. Sustainable oil palm development on degraded land in Kalimantan. World Wildlife Fund, Washington DC, US. FAOSTAT, 2016. http://faostat3.fao.org/home/E. Date accessed: 13 April 2016. FAOSTAT, 2014. FAOSTAT, 2014. http://faostat3.fao.org/home/E.

201

Fargione, J., Hill, J., Tilman, D., Polasky, S., Hawthorne, P., 2008. Land clearing and the biofuel carbon debt. Science (80-.). 319, 1235 – 8. doi:10.1126/science.1152747. Feintrenie, L., Chong, W.K., Levang, P., 2010. Why do farmers prefer oil palm? lessons Learnt from Bungo District, Indonesia. Small-scale For. 9, 379 – 396. doi:10.1007/s11842-010-9122-2. Feldpausch, T.R., Lloyd, J., Lewis, S.L., Brienen, R.J.W., Gloor, M., Monteagudo Mendoza, A., Lopez-Gonzalez, G., Banin, L., Abu Salim, K., Affum-Baffoe, K., Alexiades, M., Almeida, S., Amaral, I., Andrade, A., Aragão, L.E.O.C., Araujo Murakami, A., Arets, E.J.M.M., Arroyo, L., Aymard C., G.A., Baker, T.R., Bánki, O.S., Berry, N.J., Cardozo, N., Chave, J., Comiskey, J.A., Alvarez, E., de Oliveira, A., Di Fiore, A., Djagbletey, G., Domingues, T.F., Erwin, T.L., Fearnside, P.M., França, M.B., Freitas, M.A., Higuchi, N., Iida, Y., Jiménez, E., Kassim, A.R., Killeen, T.J., Laurance, W.F., Lovett, J.C., Malhi, Y., Marimon, B.S., Marimon-Junior, B.H., Lenza, E., Marshall, A.R., Mendoza, C., Metcalfe, D.J., Mitchard, E.T.A., Neill, D.A., Nelson, B.W., Nilus, R., Nogueira, E.M., Parada, A., Peh, K.S.H., Pena Cruz, A., Peñuela, M.C., Pitman, N.C.A., Prieto, A., Quesada, C.A., Ramírez, F., Ramírez-Angulo, H., Reitsma, J.M., Rudas, A., Saiz, G., Salomão, R.P., Schwarz, M., Silva, N., Silva-Espejo, J.E., Silveira, M., Sonké, B., Stropp, J., Taedoumg, H.E., Tan, S., ter Steege, H., Terborgh, J., Torello-Raventos, M., van der Heijden, G.M.F., Vásquez, R., Vilanova, E., Vos, V.A., White, L., Willcock, S., Woell, H., Phillips, O.L., 2012. Tree height integrated into pantropical forest biomass estimates. Biogeosciences 9, 3381 – 3403. doi:10.5194/bg-9-3381-2012. Ferry, B., Morneau, F., Bontemps, J.-D., Blanc, L., Freycon, V., 2010. Higher treefall rates on slopes and waterlogged soils result in lower stand biomass and productivity in a tropical rain forest. J. Ecol. 98, 106 – 116. doi:10.1111/ j.1365-2745.2009.01604.x. 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 Ecol. Evol. 23, 538 – 45. doi:10.1016/j.tree.2008.06.012. Food and Agriculture Organization of the United Nations, 2014. http://www.fao.org. Foody, G.M., 2003. Geographical weighting as a further refinement to regression modelling: An example focused on the NDVI – rainfall relationship. Remote Sens. Environ. 88, 283 – 293. doi:10.1016/j.rse.2003.08.004. Frangi, J., Lugo, A., 1985. Ecosystem dynamics of a subtropical floodplain forest. Ecol. Monogr. 55, 351 – 369. Fuller, D., Jessup, T., Salim, A., 2004. Loss of forest cover in Kalimantan, Indonesia, since the 1997 – 1998 El Nino. Conserv. Biol. 18, 249 – 254. Fuller, D.O., Meijaard, E.M., Christy, L., Jessup, T.C., 2010. Spatial assessment of threats to biodiversity within East Kalimantan, Indonesia. Appl. Geogr. 30, 416 – 425. doi:10.1016/j.apgeog.2009.11.004. Gaskell, J., Sunderland-Groves, J., Sheil, D., Casson, A., Meijaard, E., Noordwijk van, M., Gaskell, J., SunderlandGroves, J., Wertz, K., Kanninen, M., 2009. The impacts and opportunities of oil palm in Southeast Asia: What do we know and what do we need to know? Occasional paper no. 51, CIFOR, Bogor, Indonesia. Gatto, M., Wollni, M., Qaim, M., 2015. Oil palm boom and land-use dynamics in Indonesia: The role of policies and socioeconomic factors. Land use policy 46, 292 – 303. doi:10.1016/j.landusepol.2015.03.001. 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. doi:10.1371/journal.pone.0069887. 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., 2014. Four Decades of Forest Persistence, Clearance and Logging on Borneo. PLoS One 9, e101654. doi:10.1371/journal.pone.0101654. GCF-TF Indonesia, 2011. Governors’ Climate and Forests Task Force Indonesia. East Kalimantan’s Green Growth Journey. GCF Taskforce Meeting, Palangka Raya, Central Kalimantan, Indonesia 2011. http://www.gcftaskforce. org/meeting/documents/East. Date accessed: 15 January 2014. GDV, 2015. Rice – Transport Information Service. http://www.tis-gdv.de/tis_e/ware/getreide/reis/reis.htm. Date accessed: 5 January 2015.

202

Gerasimchuk, I., Yam Koh, P., 2013. Cutting subsidies or cutting rainforest? The International Institute for Sustainable Development, Winnipeg, Canada. Gerber, J.F., 2011. Conflicts over industrial tree plantations in the South: Who, how and why? Glob. Environ. Chang. 21, 165 – 176. doi:10.1016/j.gloenvcha.2010.09.005. Germer, J., Sauerborn, J., 2008. Estimation of the impact of oil palm plantation establishment on greenhouse gas balance. Environ. Dev. Sustain. 10, 697 – 716. doi:10.1007/s10668-006-9080-1. Giam, X., Mani, L., Koh, L.P., Tan, H.T.W., 2015. Saving Tropical Forests by Knowing What We Consume. Conserv. Lett. 0, 1 – 8. doi:10.1111/conl.12209. Gibbs, H.K., Brown, S., Niles, J.O., Foley, J.A., 2007. Monitoring and estimating tropical forest carbon stocks: making REDD a reality. Environ. Res. Lett. 2, 13. doi:10.1088/1748-9326/2/4/045023. 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. Proc. Natl. Acad. Sci. U. S. A. 107, 16732 – 16737. doi:10.1073/pnas.0910275107. Gingold, B., Rosenbarger, A., Muliastra, Y., Stolle, F., Sudana, I., Manessa, M., Murdimanto, A., Tiangga, S., Madusari, C., Douard, P., 2012. How to identify degraded land for sustainable palm oil in Indonesia. Washington DC, US. Graham, M.H., 2003. Confronting multicollinearity in ecological multiple regression. Ecology 84, 2809 – 2815. doi:10.1890/02-3114. Grau, R., Kuemmerle, T., Macchi, L., 2013. Beyond “land sparing versus land sharing”: environmental heterogeneity, globalization and the balance between agricultural production and nature conservation. Curr. Opin. Environ. Sustain. 5, 477 – 483. doi:10.1016/j.cosust.2013.06.001. Gutiérrez-Vélez, V.H., DeFries, R., Pinedo-Vásquez, M., Uriarte, M., Padoch, C., Baethgen, W., Fernandes, K., Lim, Y., 2011. High-yield oil palm expansion spares land at the expense of forests in the Peruvian Amazon. Environ. Res. Lett. 6, 044029. doi:10.1088/1748-9326/6/4/044029. Hamanakaa, A., Inouea, N., Shimadaa, H., Sasaokaa, T., Matsui, K., Miyajima, I., 2015. Design of self-sustainable land surface against soil erosion at rehabilitation areas in open-cut mines in tropical regions. Int. J. Mining, Reclam. Environ. 29, 305 – 315. doi:10.1080/17480930.2014.1000644. 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 (80-.). 342, 850 – 3. doi:10.1126/ science.1244693. Hansen, M.C., Stehman, S. V, Potapov, P. V, Arunarwati, B., Stolle, F., Pittman, K., 2009. Quantifying changes in the rates of forest clearing in Indonesia from 1990 to 2005 using remotely sensed data sets. Environ. Res. Lett. 4, 034001. doi:10.1088/1748-9326/4/3/034001. Hardjanto, Y.S., 2015. East Kalimantan reaffirms moratorium on new mining, logging concessions. MongabayIndonesia. Harris, N.L., Brown, K., Netzer, M., Gunarso, P., Killeen, T.J., 2013. Projections of oil palm expansion in indonesia, malaysia and papua new guinea from 2010 to 2050 – Appendix. Winrock International, Arlington, US. Harris, N.L., Petrova, S., Stolle, F., Brown, S., 2008. Identifying optimal areas for REDD intervention: East Kalimantan, Indonesia as a case study. Environ. Res. Lett. 3, 035006. doi:10.1088/1748-9326/3/3/035006. Hashimotio, T., Kojima, K., Tange, T., Sasaki, S., 2000. Changes in carbon storage in fallow forests in the tropical lowlands of Borneo. For. Ecol. Manage. 126, 331 – 337. doi:10.1016/S0378-1127(99)00104-8. Hiratsuka, M., Toma, T., 2006. Biomass recovery of naturally regenerated vegetation after the 1998 forest fire in East Kalimantan, Indonesia. Japan Agric. pp 277 – 282. Hodges, R.J., Buzby, J.C., Bennett, B., 2011. Postharvest losses and waste in developed and less developed countries: opportunities to improve resource use 37 – 45. doi:10.1017/S0021859610000936.

203

Hoekman, D.H., Quiñones, M.J., 2000. Land cover type and biomass classification using AirSAR data for evaluation of monitoring scenarios in the Colombian Amazon. IEEE Trans. Geosci. Remote Sens. 38, 685 – 696. Hoffmann, A.A., Hinrichs, A., Siegert, F., 1999. Fire damage in East Kalimantan in 1997/98 related to land use and vegetation classes: Satellite radar inventory results and proposals for further actions. IFFM/SFMP GTZ 26. 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. doi:10.5194/bg-7-15052010. Houghton, R., 2005. Aboveground Forest Biomass and the Global Carbon Balance. Glob. Chang. Biol. 11, 945 – 958. doi:10.1111/j.1365-2486.2005.00955.x. Houghton, R., Skole, D.L., Nobre, C.A, Hackler, J.L., Lawrence, K.T., Chomentowski, W.H., 2000. Annual fluxes of carbon from deforestation and regrowth in the Brazilian Amazon. Nature 403, 301 – 4. doi:10.1038/35002062. Houghton, R.A., Hall, F., Goetz, S.J., 2009. Importance of biomass in the global carbon cycle. J. Geophys. Res. Biogeosciences 114, 1 – 13. doi:10.1029/2009JG000935. Hromatka, T., Savage, V.R., 2010. Timber shortage and the sustainability of the Malaysian furniture industry, in: LinHeng, L., Ofori, G., Lai Choo, M.-L., Savage, V.R., Yen-Peng, T. (Eds.), Sustainability Matters: Environmental Management in Asia. World Scientific Publishing Co. Pte. Ltd, Singapore, pp. 413 – 442. Husin M, 2000. Utilization of oil palm biomass for various wood-based and other products. Adv. oil palm Res. 2, 1346 – 1412. ICRAF, 2014. Chapter 3. Tree cover transitions in tropical landscapes: hypotheses and cross-continental synthesis, in: Minang, P.A., van Noordwijk M., Kahurani, E. (Eds.), Partnership in the Tropical Forest Margins: A 20-Year Journey in Search of Alternatives to Slash-and-Burn. ICRAF, Nairobi, Kenya, pp. 17 – 22. Immerzeel, D.J., Verweij, P.A., van der Hilst, F., Faaij, A.P.C., 2014. Biodiversity impacts of bioenergy crop production: A state-of-the-art review. GCB Bioenergy 6, 183 – 209. doi:10.1111/gcbb.12067. Inoue, M., Kawai, M., Imang, N., 2013. Implications of local peoples’ preferences in terms of income source and land use for Indonesia’s national REDD-plus policy: evidence in East Kalimantan, Indonesia. Int. J. Environ. Sustain. Dev. 12, 244 – 263. IRRI, 2011. IRRI Indonesia. http://irri.org/our-work/locations/indonesia. John, A., 2014. Rodent outbreaks and rice pre-harvest losses in Southeast Asia. Food Secur. 6, 249 – 260. doi:10.1007/ s12571-014-0338-4. Karssenberg, D., Schmitz, O., Salamon, P., de Jong, K., Bierkens, M.F.P., 2010. A software framework for construction of process-based stochastic spatio-temporal models and data assimilation. Environ. Model. Softw. 25, 489 – 502. doi:10.1016/j.envsoft.2009.10.004. Kassam, A.H., van Velthuizen, H.T., Mitchell, A.B., Fischer, G.W., Shah, M.M., 1992. Agro-ecological land resources assessment for agricultural development planning. A case study of Kenya, UN-FAO, Land and Water Development Division. Rome, Italy. Katayama, A., Kume, T., Komatsu, H., Saitoh, T.M., Ohashi, M., Nakagawa, M., Suzuki, M., Otsuki, K., Kumagai, T., 2013. Carbon allocation in a Bornean tropical rainforest without dry seasons. J. Plant Res. 126, 505 – 515. doi:10.1007/s10265-012-0544-0. Khasanah, N., van Noordwijk, M., Ningsih, H., Wich, S., 2015. Aboveground carbon stocks in oil palm plantations and the threshold for carbon-neutral vegetation conversion on mineral soils. Cogent Environ. Sci. 1, 1119964. doi :10.1080/23311843.2015.1119964. Koenig, W., 1999. Spatial autocorrelation of ecological phenomena. Trends Ecol. Evol. 14, 22 – 26. Koh, L.P., Ghazoul, J., 2010. Spatially explicit scenario analysis for reconciling agricultural expansion, forest protection, and carbon conservation in Indonesia. Proc. Natl. Acad. Sci. U. S. A. 107, 11140 – 11144. doi:10.1073/pnas.1012681107.

204

Koh, L.P., Miettinen, J., Liew, S.C., Ghazoul, J., 2011. Remotely sensed evidence of tropical peatland conversion to oil palm. Proc. Natl. Acad. Sci. U. S. A. 108, 5127 – 32. doi:10.1073/pnas.1018776108. Koh, L.P., Wilcove, D.S., 2008. Is oil palm agriculture really destroying tropical biodiversity? Conserv. Lett. 1, 60 – 64. doi:10.1111/j.1755-263X.2008.00011.x. Kotowska, M.M., Leuschner, C., Triadiati, T., Meriem, S., Hertel, D., 2015. Quantifying above- and belowground biomass carbon loss with forest conversion in tropical lowlands of Sumatra (Indonesia). Glob. Chang. Biol. 21, 3620 – 3634. doi:10.1111/gcb.12979. Kronseder, K., Ballhorn, U., Böhmc, V., Siegert, F., Böhm, V., Siegert, F., 2012. Above ground biomass estimation across forest types at different degradation levels in Central Kalimantan using LiDAR data. Int. J. Appl. Earth Obs. Geoinf. 18, 37 – 48. doi:10.1016/j.jag.2012.01.010. Laborde, D., 2011. Assessing the land use change consequences of European biofuel policies, International Food Policy Research Institute. Washington DC, US. Laborte, A.G., de Bie, K.C.A.J.M., Smaling, E.M.A., Moya, P.F., Boling, A.A., Van Ittersum, M.K., 2012. Rice yields and yield gaps in Southeast Asia: Past trends and future outlook. Eur. J. Agron. 36, 9 – 20. doi:10.1016/j. eja.2011.08.005. Lambin, E.F., 1997. Modelling and monitoring land-cover change processes in tropical regions. Prog. Phys. Geogr. 21, 375 – 393. doi:10.1177/030913339702100303. Lambin, E.F., Geist, H.J., Lepers, E., 2003. Dynamics of land-use and land-cover change in tropical regions. Annu. Rev. Environ. Resour. 28, 205 – 241. doi:10.1146/annurev.energy.28.050302.105459. Lambin, E.F., Meyfroidt, P., 2011. Global land use change, economic globalization, and the looming land scarcity. Proc. Natl. Acad. Sci. U. S. A. 108, 3465 – 72. doi:10.1073/pnas.1100480108. Lambin, E.F., Turner, B.L., Geist, H.J., Agbola, S.B., Angelsen, A., Bruce, J.W., Coomes, O.T., Dirzo, R., Fischer, G., Folke, C., George, P.S., Homewood, K., Imbernon, J., Leemans, R., Li, X., Moran, E.F., Mortimore, M., Ramakrishnan, P.S., Richards, J.F., Skånes, H., Steffen, W., Stone, G.D., Svedin, U., Veldkamp, T.A., Vogel, C., Xu, J., 2001. The causes of land-use and land-cover change: moving beyond the myths. Glob. Environ. Chang. 11, 261 – 269. doi:10.1016/S0959-3780(01)00007-3. Langner, A., Miettinen, J., Siegert, F., 2007. Land cover change 2002 – 2005 in Borneo and the role of fire derived from MODIS imagery. Glob. Chang. Biol. 13, 2329 – 2340. doi:10.1111/j.1365-2486.2007.01442.x. Langner, A., Samejima, H., Ong, R.C., Titin, J., Kitayama, K., 2012. Integration of carbon conservation into sustainable forest management using high resolution satellite imagery: A case study in Sabah, Malaysian Borneo. Int. J. Appl. Earth Obs. Geoinf. 18, 305 – 312. doi:10.1016/j.jag.2012.02.006. Laurance, W.F., Balmford, A., 2013. Land use: A global map for road building. Nature 495, 308 – 9. doi:10.1038/495308a. Laurance, W.F., Fearnside, P.M., Laurance, S.G., Delamonica, P., Lovejoy, T.E., Rankin-de Merona, J.M., Chambers, J.Q., Gascon, C., 1999. Relationship between soils and Amazon forest biomass: a landscape-scale study. For. Ecol. Manage. 118, 127 – 138. doi:10.1016/S0378-1127(98)00494-0. Laurance, W.F., Koh, L.P., Butler, R., Sodhi, N.S., Bradshaw, C.J.A., Neidel, J.D., Consunji, H., Vega, J.M., 2010. Improving the performance of the roundtable on sustainable palm oil for nature conservation. Conserv. Biol. 24, 377 – 381. doi:10.1111/j.1523-1739.2010.01448.x. Laurance, W.F., Sayer, J., Cassman, K.G., 2014. Agricultural expansion and its impacts on tropical nature. Trends Ecol. Evol. 29, 107 – 16. doi:10.1016/j.tree.2013.12.001. 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. Biol. Conserv. 186, 276 – 286. doi:10.1016/j.biocon.2015.03.004.

205

Lee, J.S.H., Abood, S., Ghazoul, J., Barus, B., Obidzinski, K., Koh, L.P., 2014a. Environmental Impacts of LargeScale Oil Palm Enterprises Exceed that of Smallholdings in Indonesia. Conserv. Lett. 7, 25 – 33. doi:10.1111/ conl.12039. Lee, J.S.H., Garcia-Ulloa, J., Ghazoul, J., Obidzinski, K., Koh, L.P., 2014b. Modelling environmental and socioeconomic trade-offs associated with land-sparing and land-sharing approaches to oil palm expansion. J. Appl. Ecol. 51, 1366 – 1377. doi:10.1111/1365-2664.12286. Lee, J.S.H., Ghazoul, J., Obidzinski, K., Koh, L.P., 2014c. Oil palm smallholder yields and incomes constrained by harvesting practices and type of smallholder management in Indonesia. Agron. Sustain. Dev. 34, 501 – 513. doi:10.1007/s13593-013-0159-4. Lucey, J., Hill, J., Reynolds, G., 2015. Co-benefits for biodiversity and carbon in land planning decisions within oil palm landscapes. Luttrell, C., Resosudarmo, I.A.P., Muharrom, E., Brockhaus, M., Seymour, F., 2012. The political context of REDD+ in Indonesia: Constituencies for change. Environ. Sci. Policy 1 – 9. doi:10.1016/j.envsci.2012.10.001 Mace, G.M., Norris, K., Fitter, A.H., 2012. Biodiversity and ecosystem services: A multilayered relationship. Trends Ecol. Evol. 27, 19 – 25. doi:10.1016/j.tree.2011.08.006. Macedo, M.N., DeFries, R.S., Morton, D.C., Stickler, C.M., Galford, G.L., Shimabukuro, Y.E., 2012. Decoupling of deforestation and soy production in the southern Amazon during the late 2000s. Proc. Natl. Acad. Sci. U. S. A. 109, 1341 – 6. doi:10.1073/pnas.1111374109. Makarim, A., 2000. Bridging the rice yield gap in indonesia, in: Papademetriou, M., Dent, F., Herath, E. (Eds.), Bridging the Rice Yield Gap in the Asia-Pacific Region. Food and Agriculture Organization of the United Nations, Bangkok, pp. 112 – 121. Margono, B.A., Potapov, P. V, Turubanova, S., Stolle, F., Hansen, M.C., 2014. Primary forest cover loss in Indonesia over 2000 – 2012. Nat. Clim. Chang. 4, 1 – 6. doi:10.1038/NCLIMATE2277. Marlier, M.E., DeFries, R.S., Kim, P.S., Koplitz, S.N., Jacob, D.J., Mickley, L.J., Myers, S.S., 2015. Fire emissions and regional air quality impacts from fires in oil palm, timber, and logging concessions in Indonesia. Environ. Res. Lett. 10, 085005. doi:10.1088/1748-9326/10/8/085005. Massawe, F., Mayes, S., Cheng, A., 2016. Crop Diversity: An Unexploited Treasure Trove for Food Security. Trends Plant Sci. 21, 365 – 368. doi:10.1016/j.tplants.2016.02.006. Mather, A.S., 1992. The Forest Transition. Area 24, 367 – 379. Mather, A.S., Needle, C.L., 1998. The forest transition: a theoretical basis. Area 30, 117 – 124. Matricardi, E.A.T., Skole, D.L., Pedlowski, M.A., Chomentowski, W., Fernandes, L.C., 2010. Assessment of tropical forest degradation by selective logging and fire using Landsat imagery. Remote Sens. Environ. 114, 1117 – 1129. doi:10.1016/j.rse.2010.01.001. Mayaux, P., Pekel, J.-F., Desclée, B., Donnay, F., Lupi, A., Achard, F., Clerici, M., Bodart, C., Brink, A., Nasi, R., Belward, A., 2013. State and evolution of the African rainforests between 1990 and 2010. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 368, 20120300. doi:10.1098/rstb.2012.0300. MEA, 2005. Ecosystems and human well-being: Synthesis., Millennium Ecosystem Assessment. doi:10.1196/ annals.1439.003. Melo, F.P.L., Arroyo-Rodríguez, V., Fahrig, L., Martínez-Ramos, M., Tabarelli, M., 2013. On the hope for biodiversity-friendly tropical landscapes. Trends Ecol. Evol. 28, 461 – 468. doi:10.1016/j.tree.2013.01.001 Mertens, B., Lambin, E., 2000. Land-cover-change trajectories in southern Cameroon. Ann. Assoc. Am. Geogr. 90, 467 – 494. Mertens, B., Lambin, E.F., 1997. Spatial modelling of deforestation in southern Cameroon. Appl. Geogr. 17, 143 – 162. doi:10.1016/S0143-6228(97)00032-5. Mertz, O., Padoch, C., Fox, J., Cramb, R.A., Leisz, S.J., Lam, N.T., Vien, T.D., 2009. Swidden change in southeast Asia: Understanding causes and consequences. Hum. Ecol. 37, 259 – 264. doi:10.1007/s10745-009-9245-2.

206

Meyfroidt, P., Lambin, E.F., 2011. Global Forest Transition: Prospects for an End to Deforestation, Annual Review of Environment and Resources. doi:doi:10.1146/annurev-environ-090710-143732. Meyfroidt, P., Rudel, T.K., Lambin, E.F., 2010. Forest transitions, trade, and the global displacement of land use. Proc. Natl. Acad. Sci. 107, 21300 – 21305. doi:10.1073/pnas. Miettinen, J., Shi, C., Liew, S.C., 2011. Deforestation rates in insular Southeast Asia between 2000 and 2010. Glob. Chang. Biol. 17, 2261 – 2270. doi:10.1111/j.1365-2486.2011.02398.x. Ministry of Forestry Indonesia, 2008. Digital Forest Land Use Map. Ministry of Forestry Indonesia (year unknown), Kawasan Hutan (Forest estate) land use maps, General Direktorat of Planning, Ministry of Forestry; downloaded from http://appgis.dephut.go.id/appgis/kml.aspx. Processed and provided by Greenpeace. Prepared by the World Resources Institute (2012). Mitchard, E.T.A., Saatchi, S.S., Lewis, S.L., Feldpausch, T.R., Woodhouse, I.H., Sonké, B., Rowland, C., Meir, P., 2011a. Measuring biomass changes due to woody encroachment and deforestation/degradation in a forest – savanna boundary region of central Africa using multi-temporal L-band radar backscatter. Remote Sens. Environ. 115, 2861 – 2873. doi:10.1016/j.rse.2010.02.022. Mitchard, E.T.A., Saatchi, S.S., White, L.J.T., Abernethy, K.A., Jeffery, K.J., Lewis, S.L., Collins, M., Lefsky, M.A., Leal, M.E., Woodhouse, I.H., Meir, P., 2011b. Mapping tropical forest biomass with radar and spaceborne LiDAR: overcoming problems of high biomass and persistent cloud. Biogeosciences Discuss. 8, 8781 – 8815. doi:10.5194/bgd-8-8781-2011. Mitchard, E.T.A., Saatchi, S.S., Woodhouse, I.H., Nangendo, G., Ribeiro, N.S., Williams, M., Ryan, C.M., Lewis, S.L., Feldpausch, T.R., Meir, P., 2009. Using satellite radar backscatter to predict above-ground woody biomass: A consistent relationship across four different African landscapes. Geophys. Res. Lett. 36, L23401. doi:10.1029/2009GL040692. Molenaar, J.W., Orth, M., Lord, S., Meekers, P., Taylor, C., Hanu, M.D.A., Elson, D., Ginting, L., 2010. Analysis of the agronomic and institutional constraints to smallholder yield improvement in Indonesia. Amsterdam, the Netherlands. Morel, A.C., Fisher, J.B., Malhi, Y., 2012. Evaluating the potential to monitor aboveground biomass in forest and oil palm in Sabah, Malaysia, for 2000 – 2008 with Landsat ETM+ and ALOS-PALSAR. Int. J. Remote Sens. 33, 3614 – 3639. doi:10.1080/01431161.2011.631949. 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. For. Ecol. Manage. 262, 1786 – 1798. doi:10.1016/j.foreco.2011.07.008. Mosnier, A., Havlík, P., Obersteiner, M., Aoki, K., Schmid, E., Fritz, S., McCallum, I., Leduc, S., 2014. Modeling Impact of Development Trajectories and a Global Agreement on Reducing Emissions from Deforestation on Congo Basin Forests by 2030. Environ. Resour. Econ. 57, 505 – 525. doi:10.1007/s10640-012-9618-7. MPOB, 2014. http://www.mpob.gov.my/Date accessed: 18 February 2015. Müller, D., Sun, Z., Vongvisouk, T., Pflugmacher, D., Xu, J., Mertz, O., 2014. Regime shifts limit the predictability of land-system change. Glob. Environ. Chang. 28, 75 – 83. doi:10.1016/j.gloenvcha.2014.06.003. Murdiyarso, D., Adiningsih, E.S., 2006. Climate anomalies, Indonesian vegetation fires and terrestrial carbon emissions. Mitig. Adapt. Strateg. Glob. Chang. 12, 101 – 112. doi:10.1007/s11027-006-9047-4. Murdiyarso, D., Skutsch, M., Guariguata, M., Kanninen, M., Luttrell, C., Verweij, P., Stella, O., 2008. Measuring and monitoring forest degradation for REDD. Implic. Ctry. circumstances. CIFOR, Bogotm Indonesia. Murray, J.P., Grenyer, R., Wunder, S., Raes, N., Jones, J.P.G.G., 2015. Spatial patterns of carbon, biodiversity, deforestation threat, and REDD+ projects in Indonesia. Conserv. Biol. 00, n/a – n/a. doi:10.1111/cobi.12500. NASA, 2016. VIRS 375m Active Fire Detections. NASA LANCE FIRMS, 2016 VIIRS 375m Act. Fire Prod. Data set. Available on-line [https//earthdata.nasa.gov/firms]. Accessed: 18 Febr. 2016. NASA, 2012. http://www2.jpl.nasa.gov/srtm/

207

NASA/LANCE – FIRMS, 2011. MODIS Hotspot/Active Fire Detections. Data set. Acquired on 17-09-2012 .online http://earthdata.nasa.gov/data/nrtdata/firms. Obidzinski, K., Andriani, R., Komarudin, H., Andrianto, A., 2012. Environmental and social impacts of oil palm plantations and their implications for biofuel production in Indonesia. Ecol. Soc. 17. doi:10.5751/ES-04775170125. Obidzinski, K., Chaudhury, M., 2009. Transition to timber plantation based forestry in Indonesia: towards a feasible new policy. Int. For. Rev. 11, 79 – 87. Obidzinski, K., Dermawan, A., 2012. Pulp industry and environment in Indonesia: is there sustainable future? Reg. Environ. Chang. 12, 961 – 966. doi:10.1007/s10113-012-0353-y. OECD/FAO, 2015. OECD/FAO Agricultural Outlook 2015. OECD Publishing, Paris. Ostrom, E., Janssen, M.A., Anderies, J.M., 2007. Going beyond panaceas. Proc. Natl. Acad. Sci. U. S. A. 104, 15176 – 8. doi:10.1073/pnas.0701886104. Page, S., Siegert, F., Rieley, J., Boehm, H., 2002. The amount of carbon released from peat and forest fires in Indonesia during 1997. Nature 1999, 61 – 65. doi:10.1038/nature01141.1. Page, S.E., Rieley, J.O., Banks, C.J., 2011. Global and regional importance of the tropical peatland carbon pool. Glob. Chang. Biol. 17, 798 – 818. doi:10.1111/j.1365-2486.2010.02279.x. Paoli, G.D., Curran, L.M., Slik, J.W.F., 2008. Soil nutrients affect spatial patterns of aboveground biomass and emergent tree density in southwestern Borneo. Oecologia 155, 287 – 99. doi:10.1007/s00442-007-0906-9. Paoli, G.D., Wells, P.L., Meijaard, E., Struebig, M.J., Marshall, A.J., Obidzinski, K., Tan, A., Rafiastanto, A., Yaap, B., Ferry Slik, J., Morel, A., Perumal, B., Wielaard, N., Husson, S., D’Arcy, L., 2010. Biodiversity Conservation in the REDD. Carbon Balance Manag. 5, 7. doi:10.1186/1750-0680-5-7. Pearce, D., Moran, D., 1994. The Economic Value of Biodiversity. Earthscan Publications Limited, London, UK. Perera, A.H., Drew, C.A., Johnson, C.J. (Eds.), 2011. Expert Knowledge and Its Application in Landscape Ecology. Springer Science & Business Media. Perz, S.G., 2007. Grand Theory and Context-Specificity in the Study of Forest Dynamics: Forest Transition Theory and Other Directions. Prof. Geogr. 59, 105 – 114. doi:10.1111/j.1467-9272.2007.00594.x. Petit, C., Scudder, T., Lambin, E., 2001. Quantifying processes of land-cover change by remote sensing: Resettlement and rapid land-cover changes in south-eastern Zambia. Int. J. Remote Sens. 22, 3435 – 3456. doi:10.1080/01431160010006881. Phalan, B., Bertzky, M., Butchart, S.H.M., Donald, P.F., Scharlemann, J.P.W., Stattersfield, A.J., Balmford, A., 2013. Crop expansion and conservation priorities in tropical countries. PLoS One 8, e51759. doi:10.1371/journal. pone.0051759. Phalan, B., Green, R., Balmford, A., 2014. Closing yield gaps: perils and possibilities for biodiversity conservation. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 369, 20120285. doi:10.1098/rstb.2012.0285. Phalan, B., Onial, M., Balmford, A., Green, R.E., 2011. Reconciling food production and biodiversity conservation: Land sharing and land sparing compared. Science (80-.). 333, 1289 – 1291. Phillips, S.J., Anderson, R.P., Schapire, R.E., 2006. Maximum entropy modeling of species geographic distributions. Ecol. Modell. 190, 231 – 259. doi:10.1016/j.ecolmodel.2005.03.026. Phillips, S.J., Dudík, M., Schapire, R.E., 2004. A maximum entropy approach to species distribution modeling, in: Proceedings of the Twenty-First International Conference on Machine Learning. pp. 655 – 662. doi:10.1145/1015330.1015412. Pirard, R., Cossalter, C., 2006. The revival of industrial forest plantations in Indonesia’s Kalimantan provinces. Will they help eliminate fiber shortfalls at Sumatran pulp mills or feed the China market? Working Paper No. 37, CIFOR, Bogor, Indonesia. Platts, 2015. Indonesian government committed to biodiesel mandate: official. http://www.platts.com/latest-news/ agriculture/singapore/indonesian-government-committed-to-biodiesel-27042311. Accessed on 11-06-2016.

208

Prakoso, K.U., 2006. Tropical Forest Mapping Using Polarimetric and Interferometric SAR Data. Prasetyo, L.B., 2000. Development for Data Base for Eco-System Changes and Emission Changes of GHG. Priadjati, A., 2002. Dipterocarpaceae: Forest Fires and Forest Recovery. Wageningen University. TropenbosKalimantan series 8. Tropenbos International. Wageningen, The Netherlands. Propastin, P., 2012. Modifying geographically weighted regression for estimating aboveground biomass in tropical rainforests by multispectral remote sensing data. Int. J. Appl. Earth Obs. Geoinf. 18, 82 – 90. doi:10.1016/j. jag.2011.12.013. Pye-Smith, C., 2011. Rich rewards for rubber? Research in Indonesia is exploring how smallholders can increase rubber production, retain biodiversity and provide additional environmental benefits. Nairobi. Quiñones, M.J., Schut, V., Wielaard, N., Hoekman, D., 2011. Above Ground Biomass map Kalimantan 2008 – Final report. SarVision Wageningen 34. Raes, N., 2012. Partial versus full species distribution models. Nat. a Conserv. 10, 127 – 138. doi:10.4322/ natcon.2012.020. Raes, N., Roos, M.C., Slik, J.W.F., Van Loon, E.E., Steege, H. Ter, 2009. Botanical richness and endemicity patterns of Borneo derived from species distribution models. Ecography (Cop.). 32, 180 – 192. doi:10.1111/j.16000587.2009.05800.x. Raes, N., Saw, L.G., van Welzen, P.C., Yahara, T., 2013. Legume diversity as indicator for botanical diversity on Sundaland, South East Asia. South African J. Bot. 89, 265 – 272. doi:10.1016/j.sajb.2013.06.004. Ramdani, F., Hino, M., 2013. Land use changes and GHG emissions from tropical forest conversion by oil palm plantations in Riau Province, Indonesia. PLoS One 8, e70323. doi:10.1371/journal.pone.0070323. Republic of Indonesia, 2013. Instruction of the President of the Republic of Indonesia, 6/2013 on Suspension of new license issuance and improvement of primary natural forest and peatland governance. Jakarta, Indonesia. Rist, L., Feintrenie, L., Levang, P., 2010. The livelihood impacts of oil palm: Smallholders in Indonesia. Biodivers. Conserv. 19, 1009 – 1024. doi:10.1007/s10531-010-9815-z. Rudel, T., Coomes, O., Moran, E., 2005. Forest transitions: towards a global understanding of land use change. … Environ. Chang. 15, 23 – 31. doi:10.1016/j.gloenvcha.2004.11.001. Rudel, T.K., Defries, R., Asner, G.P., Laurance, W.F., 2009. Changing drivers of deforestation and new opportunities for conservation. Conserv. Biol. 23, 1396 – 1405. doi:10.1111/j.1523-1739.2009.01332.x. Rudel, T.K., Schneider, L., Uriarte, M., 2010. Forest transitions: An introduction. Land use policy 27, 95 – 97. doi:10.1016/j.landusepol.2009.09.021. Ruhiyat, D., 1989. Die Entwicklung der Standurlichen Nährstoffvorräte bei Naturnaher Walderwirtshcaftung und im Plantagenbetrieb, Ostkalimantan-Indonesien. Diss. Forst. Univ. Goettingen. Rupani, P., Singh, R., Ibrahim, M., Esa, N., 2010. Review of current palm oil mill effluent (POME) treatment methods: vermicomposting as a sustainable practice. World Appl. Sci. J. 11, 70 – 81. Rusch, A., Birkhofer, K., Bommarco, R., Smith, H.G., Ekbom, B., 2014. Management intensity at field and landscape levels affects the structure of generalist predator communities. Oecologia 175, 971 – 983. doi:10.1007/s00442014-2949-z. 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. Proc. Natl. Acad. Sci. U. S. A. 108, 9899 – 904. doi:10.1073/ pnas.1019576108. Saatchi, S.S., Houghton, R.A., Dos Santos Alvalá, R.C., Soares, J.V., Yu, Y., 2007. Distribution of aboveground live biomass in the Amazon basin. Glob. Chang. Biol. 13, 816 – 837. doi:10.1111/j.1365-2486.2007.01323.x. Santosa, S.J., 2008. Palm Oil Boom in Indonesia: From Plantation to Downstream Products and Biodiesel. CLEAN – Soil, Air, Water 36, 453 – 465. doi:10.1002/clen.200800039. SarVision, 2015. SarVision BV – Applications in Remote Sensing. http://www.sarvision.nl/

209

Sayer, J., Ghazoul, J., Nelson, P., Klintuni Boedhihartono, A., 2012. Oil palm expansion transforms tropical landscapes and livelihoods. Glob. Food Sec. 1, 114 – 119. doi:10.1016/j.gfs.2012.10.003. Senior, M.J.M., Hamer, K.C., Bottrell, S., Edwards, D.P., Fayle, T.M., Lucey, J.M., Mayhew, P.J., Newton, R., Peh, K.S.H., Sheldon, F.H., Stewart, C., Styring, A.R., Thom, M.D.F., Woodcock, P., Hill, J.K., 2013. Trait-dependent declines of species following conversion of rain forest to oil palm plantations. Biodivers. Conserv. 22, 253 – 268. doi:10.1007/s10531-012-0419-7. Siegert, F., Hoffmann, A.A., 2000. The 1998 Forest Fires in East Kalimantan (Indonesia): A Quantitative Evaluation Using High Resolution, Multitemporal ERS-2 SAR Images and NOAA-AVHRR Hotspot Data. Remote Sens. Environ. 4257. Siegert, F., Ruecker, G., Hinrichs, A., Hoffmann, A.A., 2001. Increased damage from fires in logged forests during droughts caused by El Niño. Nature 414, 437 – 40. doi:10.1038/35106547. Simbolon, H., Igarashi, T., Ruskandi, A., Dirman, 2010. Amount of C stored in aboveground biomass of trees of mixed dipterocarp forest of East Kalimantan and its dynamics., in: Proceedings of International Seminar on CDM Plantation and Biodiversity – Results of a Collaborative Research in East Kalimantan, 24 Feb 2009 in FFPRI Tsukuba. Forestry and Forest Products Research Institute, Tsukuba, pp. 7 – 13. 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., 2010a. Environmental correlates of tree biomass, basal area, wood specific gravity and stem density gradients in Borneo’s tropical forests. Glob. Ecol. Biogeogr. 19, 50 – 60. doi:10.1111/j.1466-8238.2009.00489.x. Slik, J.W.F., Bernard, C.S., Van Beek, M., Breman, F.C., Eichhorn, K.A.O., 2008. Tree diversity, composition, forest structure and aboveground biomass dynamics after single and repeated fire in a Bornean rain forest. Oecologia 158, 579 – 88. doi:10.1007/s00442-008-1163-2. Slik, J.W.F., Breman, F.C., Bernard, C., van Beek, M., Cannon, C.H., Eichhorn, K.A.O., Sidiyasa, K., 2010b. Fire as a selective force in a Bornean tropical everwet forest. Oecologia 164, 841 – 9. doi:10.1007/s00442-010-1764-4 Slik, J.W.F., Verburg, R.W., Keßler, P.J.A., 2002. Effects of fire and selective logging on the tree species composition of lowland dipterocarp forest in East Kalimantan, Indonesia. Biodivers. Conserv. 85 – 98. Smit, H.H., Meijaard, E., Van der Laan, C., Mantel, S., Budiman, A., Verweij, P.A., 2013. Breaking the Link between Environmental Degradation and Oil Palm Expansion: A Method for Enabling Sustainable Oil Palm Expansion. PLoS One 8, e68610. doi:10.1371/journal.pone.0068610. Sodhi, N.S., Koh, L.P., Brook, B.W., Ng, P.K.L., 2004. Southeast Asian biodiversity: an impending disaster. Trends Ecol. Evol. 19, 654 – 60. doi:10.1016/j.tree.2004.09.006. Sodhi, N.S., Koh, L.P., Clements, R., Wanger, T.C., Hill, J.K., Hamer, K.C., Clough, Y., Tscharntke, T., Posa, M.R.C., Lee, T.M., 2010. Conserving Southeast Asian forest biodiversity in human-modified landscapes. Biol. Conserv. 143, 2375 – 2384. doi:10.1016/j.biocon.2009.12.029. Sulaiman, O., Salim, N., Nordin, N., Hashim, R., Ibrahim, M., Sato, M., 2012. The potential of oil palm trunk biomass as an alternative source for compressed wood. BioResourcessources 7, 2688 – 2706. Susanti, A., Burgers, P., 2012. Oil palm expansion in Riau Province, Indonesia: serving people, planet, profit. Backgr. Pap. Eur. Rep. Dev. Syahrinudin, 2005. The potential of oil palm and forest plantations for carbon sequestration on degraded land in Indonesia. Tahar, A., 2013. Electricity from Biomass: Indonesia ’ s Potential in the Palm Oil Sector and How to Tap it?, in: Energy from Biomass and Biogas in the Agro-Industry. Jakarta, pp. 1 – 27. Tegegne, Y.T., Lindner, M., Fobissie, K., Kanninen, M., 2016. Evolution of drivers of deforestation and forest degradation in the Congo Basin forests: Exploring possible policy options to address forest loss. Land use policy 51, 312 – 324. doi:10.1016/j.landusepol.2015.11.024.

210

Tempo.co, 2015. Kalimantan Plans Moratorium on Mine, Forest, Palm Oil Permits. http://en.tempo.co/read/ news/2015/03/11/056649082/Kalimantan-Plans-Moratorium-on-Mine-Forest-Palm-Oil-Permits. Date accessed: 1 July 2015. Thomas, C.D., Anderson, B.J., Moilanen, A., Eigenbrod, F., Heinemeyer, A., Quaife, T., Roy, D.B., Gillings, S., Armsworth, P.R., Gaston, K.J., 2013. Reconciling biodiversity and carbon conservation. Ecol. Lett. 16, 39 – 47. doi:10.1111/ele.12054. Tilman, D., Balzer, C., Hill, J., Befort, B.L., 2011. Global food demand and the sustainable intensification of agriculture. Proc. Natl. Acad. Sci. 108, 20260 – 20264. doi:10.1073/pnas.1116437108. Toma, T., Ishida, A., Matius, P., 2005. Long-term monitoring of post-fire aboveground biomass recovery in a lowland dipterocarp forest in East Kalimantan, Indonesia. Nutr. Cycl. Agroecosystems 71, 63 – 72. doi:10.1007/s10705004-0381-1. Turner, E.C., Snaddon, J.L., Fayle, T.M., Foster, W.A., 2008. Oil palm research in context: Identifying the need for biodiversity assessment. PLoS One 3, 2 – 5. doi:10.1371/journal.pone.0001572. UNEP, 2012. Converting Waste Oil Palm Trees Into a Resource. Japan. UN-REDD, 2013. www.un-redd.org/AboutREDD/tabid/102614/Default.aspx. USDA, 2015a. Indonesia Oilseeds and Products Annual 2015. Washington DC, US. USDA, 2015b. Indonesia Biofuels Annual Report 2015, Global Agricultural Information Network (GAIN) Report. Washington DC, US. USDA, 2012. Indonesia: Stagnating rice production ensures continued need for imports. Commodity Intelligence Report. http://www.pecad.fas.usda.gov/highlights/2012/03/Indonesia_rice_Mar2012/Date accessed: 18 February 2015. Washington DC, US. van der Hilst, F., Verstegen, J.A., Karssenberg, D., Faaij, A.P.C., 2012. Spatiotemporal land use modelling to assess land availability for energy crops – illustrated for Mozambique. GCB Bioenergy 4, 859 – 874. doi:10.1111/ j.1757-1707.2011.01147.x. van der Laan, C., Verweij, P.A., Quiñones, M.J., Faaij, A.P.C., 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 Manag. 9, 8. doi:10.1186/s13021-014-0008-z. van der Laan, C., Wicke, B., Verweij, P.A., Faaij, A.P.C., 2016. Mitigation of unwanted direct and indirect land-use change – an integrated approach illustrated for palm oil, pulpwood, rubber and rice production in North and East Kalimantan, Indonesia. GCB Bioenergy 1 – 16. doi:10.1111/gcbb.12353. 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. Nat. Geosci. 2, 737 – 738. doi:10.1038/ngeo671 Van Nieuwstadt, M., 2001. The ecological consequences of logging in the burned forests of East Kalimantan, Indonesia. Conserv. Biol. 15, 1183 – 1186. Van Nieuwstadt, M. Van, Sheil, D., 2005. Drought, fire and tree survival in a Borneo rain forest, East Kalimantan, Indonesia. J. Ecol. 93, 191 – 201. doi:10.1111/j.1365-2745.2005.00954.x. Veldkamp, A., Lambin, E., 2001. Predicting land-use change. Agric. Ecosyst. Environ. 85, 1 – 6. doi:10.1016/S01678809(01)00199-2. Veldkamp, A., Verburg, P.H., 2004. Modelling land use change and environmental impact. J. Environ. Manage. 72, 1 – 3. doi:10.1016/j.jenvman.2004.04.004. Verheye, W. (eds.., 2010. Land Use, Land Cover and Soil Sciences. Encyclopedia of Life Support Systems (EOLSS), UNESCO-EOLSS Publishers, Oxford, UK. Vermeulen, S., Goad, N., 2006. Towards better practice in smallholder palm oil production. London, UK. Verstegen, J.A., Karssenberg, D., van der Hilst, F., Faaij, A., 2012. Spatio-temporal uncertainty in Spatial Decision Support Systems: A case study of changing land availability for bioenergy crops in Mozambique. Comput. Environ. Urban Syst. 36, 30 – 42. doi:10.1016/j.compenvurbsys.2011.08.003.

211

Verstegen, J.A., Karssenberg, D., van der Hilst, F., Faaij, A.P.C., 2016. Detecting systemic change in a land use system by Bayesian data assimilation. Environ. Model. Softw. 75, 424 – 438. Verweij, P., Schouten, M., Beukering, P. Van, Triana, J., Leeuw, K. Van Der, Hess, S., 2009. Keeping the Amazon forests standing: a matter of values. WWF the Netherlands, Zeist, the Netherlands. Wahab, R., Samsi, H., Mohamed, A., 2008. Utilization Potential of 30Year-old Oil Palm Trunks Laminated Veneer Lumbers for Non-structural Purposes. J. Sustain. Dev. 1, 109 – 113. Wang, Q., Ni, J., Tenhunen, J., 2005. Application of a geographically-weighted regression analysis to estimate net primary production of Chinese forest ecosystems. Glob. Ecol. Biogeogr. 14, 379 – 393. doi:10.1111/j.1466822X.2005.00153.x. Warren-Thomas, E., Dolman, P.M., Edwards, D.P., 2015. Increasing Demand for Natural Rubber Necessitates a Robust Sustainability Initiative to Mitigate Impacts on Tropical Biodiversity. Conserv. Lett. 8, 230 – 241. doi:10.1111/conl.12170. Whittaker, R., Bormann, F., 1974. The Hubbard Brook ecosystem study: forest biomass and production. Ecol. Monogr. 44, 233 – 254. Wibowo, A., Giessen, L., 2015. Absolute and relative power gains among state agencies in forest-related land use politics: The Ministry of Forestry and its competitors in the REDD+ Programme and the One Map Policy in Indonesia. Land use policy 49, 131 – 141. doi:10.1016/j.landusepol.2015.07.018. Wicke, B., 2011. Bioenergy production on degraded and marginal land: Assessing its potentials, economic performance, and environmental impacts for different settings and geographical scales. Utrecht University, Utrecht, the Netherlands. Wicke, B., Dornburg, V., Junginger, M., Faaij, A., 2008a. Different palm oil production systems for energy purposes and their greenhouse gas implications. Biomass and Bioenergy 32, 1322 – 1337. doi:10.1016/j. biombioe.2008.04.001. Wicke, B., Sikkema, R., Dornburg, V., Faaij, A., 2011. Exploring land use changes and the role of palm oil production in Indonesia and Malaysia. Land use policy 28, 193 – 206. doi:10.1016/j.landusepol.2010.06.001. Wicke, B., Sikkema, R., Dornburg, V., Junginger, M., Faaij, A.P.C., 2008b. Drivers of land use change and the role of palm oil production in indonesia and malaysia. Overview of past developments and future projections. Wilcove, D.S., Giam, X., Edwards, D.P., Fisher, B., Koh, L.P., 2013. Navjot’s nightmare revisited: Logging, agriculture, and biodiversity in Southeast Asia. Trends Ecol. Evol. 28, 531 – 540. doi:10.1016/j.tree.2013.04.005. Willemen, L., Drakou, E.G., Dunbar, M.B., Mayaux, P., Egoh, B.N., 2013. Safeguarding ecosystem services and livelihoods: Understanding the impact of conservation strategies on benefit flows to society. Ecosyst. Serv. 4, 95 – 103. doi:10.1016/j.ecoser.2013.02.004. World Resources Institute, 2015. WRI Suitability Mapper. http://www.wri.org/resources/maps/suitability-mapp. Date accessed: 12 January 2015. World Resources Institute, 2014. http://www.wri.org/Date accessed: 15 January 2014. Wu, T.Y., Mohammad, A.W., Jahim, J.M., Anuar, N., 2009. A holistic approach to managing palm oil mill effluent (POME): biotechnological advances in the sustainable reuse of POME. Biotechnol. Adv. 27, 40 – 52. doi:10.1016/j.biotechadv.2008.08.005. Yamada, H., Tanaka, R., Sulaiman, O., Hashim, R., Hamid, Z., Yahya, M., Kosugi, A., Arai, T., Murata, Y., Nirasawa, S., 2010. Old oil palm trunk: A promising source of sugars for bioethanol production. Biomass and Bioenergy 34, 1608 – 1613. doi:10.1016/j.biombioe.2010.06.011. Yamakura, T., Hagihara, A., Sukardjo, S., Ogawa, H., 1986. Above ground biomass of tropical forest stands in Indonesian Borneo. Vegetatio 68, 71 – 82. Yang, C., Huang, H., Wang, S., 2011. Estimation of tropical forest biomass using Landsat TM imagery and permanent plot data in Xishuangbanna, China. Int. J. Remote Sens. 32, 5741 – 5756. doi:10.1080/01431161.201 0.507677.

212

Yin, Z., Dekker, S.C., Rietkerk, M., van den Hurk, B.J.J.M., Dijkstra, H.A., 2016. Network based early warning indicators of vegetation changes in a land – atmosphere model. Ecol. Complex. 26, 68 – 78. doi:10.1016/j. ecocom.2016.02.004 Yusoff, S., Hansen, S.B., 2007. LCA Case Studies Feasibility Study of Performing an Life Cycle Assessment on Crude Palm Oil Production in Malaysia. Int. J. Life Cycle Assess. 12, 50 – 58. Zahari, M.W., Alimon, A.R., Wong, H.K., 2009. Utilization of oil palm co-products as feeds for livestock in Malaysia 243 – 302.

213

214

English Summary Land use change in the tropics Due to the increasing world population and overall living standard, there is a growing demand for food, feed, fibre and fuel. The production of the commodities that fulfil these demands, requires large areas of land and result in land use and land cover (LULC) change. Such LULC change occurs particularly in rural regions in the tropics, where climate, the seeming presence of large land areas, low population densities and the lack of clear land tenure systems seem to provide optimal potential for land development. On the one hand, LULC change provides income to certain groups in society, ranging from multinational companies to local companies, smallholders and government actors. On the other hand, these land developments result in land use conflicts, for example between companies and communities. Additionally, LULC change can lead to a decline of forest cover, plant and animal species, carbon stored in vegetation and soil (carbon stocks), and local food production. This is also the case in the outer islands of Indonesia, where unwanted LULC change occurs on a large scale for the production of palm oil, rubber, timber, pulpwood and mining resources. In this dissertation, we analysed LULC change that has occurred in the West Kutai and Mahakam Ulu districts, which are part of a landscape in East Kalimantan that is under high land development pressure. We focused the past analyses on the timescale 1990 to 2009 and the future projected analyses on the timescale 2010 to 2020/2030. We identified the main land use types that are involved in LULC change and forest loss and integrated these in a land use change model to project future LULC change. In addition, we show the projected LULC changes and impacts on aboveground carbon stocks and plant species in the study region. Finally, we assessed four measures to minimise unwanted LULC change. In the introduction in Chapter 1 and in the synthesis in Chapter 8, we placed our findings in the context of the Forest Transition Theory. Forest Transition Theory and aim of the dissertation According to the Forest Transition Theory, a country or region first goes through a stage of high forest cover and low deforestation rates, and then over time, loses forest cover, after which, eventually, forest cover can stabilise or even increase. In Chapter 1, we describe land development for the production of palm oil and other cash crop commodities and the related high rate of forest cover loss in Indonesia in the context of the Forest Transition Theory. Based on this, we formulated the main research question of this dissertation, namely: How to analyse LULC change processes and drivers in tropical landscapes in transition, the consequent impacts on forest cover, carbon stocks and biodiversity, and the mitigation of these impacts? In Chapter 1, we also provide an outline of the dissertation and the research questions per chapter. The introduction shows that Indonesia is a very important example of a tropical country that shows widespread LULC change and forest loss. In Chapter 2, we describe the study region of the provinces of North and East Kalimantan, and the districts of West Kutai and Mahakam Ulu therein, which are representative of processes that are occurring in the outer islands of Indonesia. In this chapter, we

215

describe biophysical and socio-economic conditions and processes that are related to LULC change in these regions. Subdivision dissertation into three parts The remaining chapters of the dissertation are subdivided into three parts. Part I focuses on LULC change and its impacts on forest cover in Chapter 3 and 4. Part II focuses on the impacts of land development and forest loss on carbon stocks and biodiversity in Chapter 5 and 6. Part III focuses on measures to mitigate unwanted land use change in Chapter 7. Part I: Chapter 3. LULC change in the study region between 1990-2009 In Chapter 3, we aimed to characterise, quantify and visualise the LULC change processes and trajectories that contributed to forest loss and degradation in West Kutai and Mahakam Ulu districts in East Kalimantan from 1990-2009. We applied quantitative and spatially-explicit analyses using Landsat-based LULC maps, supported by field information and expert knowledge. Over the time period analysed, about one third of the land use in the study area changed and forest cover declined substantially with approximately 9%, mostly by forest degradation and deforestation. Over time, however, a shift occurred from mainly forest degradation and deforestation and the development of small-scale agriculture in 1990-2000 to an increased development of large-scale monoculture agriculture in 2000-2009. Overall, LULC change became more dynamic and complex over time and in space and is characterised by a sequence of changes, defined as trajectories. We compiled a shortlist of LULC types that replaced forest cover, namely smallholder rubber, pulpwood plantations, mixed cropland and oil palm plantations. Mixed cropland and smallholder rubber were found to be intermediate land use types, with a high likelihood of conversion. We conclude that not only forests are vulnerable for conversion, but also small-scale mixed land uses, resulting in land use intensification. Accounting for LULC change trajectories is essential to improve future projections of LULC change and to support spatial planning policies. We recommend involving industries and local communities, and particularly smallholders, in the spatial planning process, as some could play an important role in agricultural development and/or may be affected by LULC change. Because most of the monocultures and concessions were developed in the designated land allocation zones and thus in accordance with the government’s spatial planning, the government can and should play an important role in guiding the expansion of agriculture in more sustainable directions that involve the maintenance of forests and local food production. Part I: Chapter 4. LULC change in the study region projected for 2010-2030 In Chapter 4, we aimed to explore the impacts of the land allocation zoning policies and the impact of different levels of land development on LULC. We hereby particularly focused on eight land uses for the production of e.g. palm oil, rubber, pulpwood and mixed cropland and the impact of LULC change on forest cover and local food production. We modelled and visualised the impacts of LULC change on forest cover and local food production in the West Kutai and Mahakam Ulu districts under four contrasting scenarios between 2009 and 2030 with the spatio-temporal PCRaster Land Use Change (PLUC) model. According to our findings, most of the LULC change between 2009 and 2030 occurs at the cost of forests and shrub- and grasslands under all scenarios, with varying effects between the scenarios. With limited development, LULC change occurs mostly in the lowlands in the south-east and forest loss between 2009 and 2030 is low (~ 0.1 Mha, ~4%). Under the unlimited-restricted scenario, however, forest cover declines stronger with 0.4 Mha (~17%), and under the unlimited-unrestricted scenario,

216

forest cover decline is strongest (1.6 Mha, ~60%). For the latter scenario, the LULC change shifts from the lowlands to the regions at higher altitudes, with a stronger decline of medium open and closed canopy forests in these areas. Meanwhile, under the two unlimited development scenarios, mixed cropland and thus food production in the lowlands is displaced and relocated to the higher altitudes where it results in forest loss. The relocation of food production to the higher altitudes is not desirable given the remoteness of particularly the midlands and mountains areas in this region. We conclude that in order to maintain primary forests, peatlands and local food production, the production of palm oil, rubber, pulpwood and coal needs to be slowed down and needs to take place under restricted land zoning and responsible spatial planning. This includes the extension and the strengthening of the current moratorium, if reconciliation of all different land functions is sought. However, the implementation of the moratorium can only be effective if reclassification of Forest area to Non-forest area is no longer allowed or strongly discouraged in forested lands or peatlands, and is additionally based on land availability and suitability for the selected crop. Conservation should also focus on medium open and very open canopy forests, because of their large regeneration potential. With limited development, sufficient grass- and shrublands are present for land development, however, the application of a robust method such as the Responsible Cultivation Method and thorough field validation need to define whether these lands are underutilised for responsible land development Part II: Chapter 5. Relationship aboveground biomass and predictor variables More insights into the spatial variation of aboveground biomass (AGB) are crucial to minimise carbon emissions and global climate change from tropical deforestation, forest degradation and agricultural expansion. In Chapter 5, we aimed to analyse how a set of biophysical and anthropogenic variables were related to and contributed to the spatial variation of aboveground biomass (AGB) North and East Kalimantan. We applied non-spatial and spatial multiple linear regression, including geographically weighted regression. We were able to explain about 59-64% of the high spatial variation in AGB. Further, we found strong positive relationships between AGB and the tested variables; altitude, slope, land allocation zoning, soil type, and distance to the nearest fire, road, river and city. Mean AGB was relatively higher at the higher altitudes, on karst and volcanic soils, with increasing distance from fire hotspots, in the Limited production forest zone, and in the Watershed protection and Conservation forest zones. Because of the strong effects of these factors on variation in AGB, efforts to minimise carbon emissions, such as REDD+, should incorporate these factors. The high correlations between the explanatory variables showed that the variables were interrelated and thus that AGB variation cannot be explained by one single variable. Instead, spatial analyses should integrate a variety of biophysical and anthropogenic variables to provide a better understanding of spatial variation in AGB. Part II: Chapter 6. Impact LULC change scenarios on carbon stocks and biodiversity In Chapter 6, we aimed to estimate the impacts of four contrasting LULC change scenarios with varying levels of land development and varying restrictions of land zoning on carbon and plant biodiversity at the landscape level. Additionally, we aimed to assess whether the responses of carbon and plant biodiversity to LULC change are similar and aim to find which areas are considered to have the largest threat of LULC change. We applied an exploratory approach, thereby using spatial and statistical analyses to quantify these impacts, using existing AGB and plant species richness data from the literature and newly collected AGB data in the case study region West Kutai and Mahakam Ulu districts in East Kalimantan.

217

Our results show that between 1990 and 2009, AGB in the region declined significantly from 877 ± 223 Mt to 732 ± 176 Mt, and mean species richness declined significantly from ~406 species to ~365 species per 10 x 10 km raster cell. Under the unlimited-unrestricted scenario, AGB is projected to decline further to 550 ± 143 Mt in 2030 and mean species richness to ~196 per raster cell in 2030. The impact on mean species richness (-22% to -52%) was stronger than the impact on total AGB (-19% to -37%) between 1990 and 2030. Under all scenarios, the lowlands showed the highest risk of LULC change. Over time, however, the threat to forest loss, and thus carbon and biodiversity losses, increases in the higher altitudes in the north-west, particularly under the unlimited development scenarios. The most responsible pathway would be to slow down land development and apply restricted zoning that prevents the development and displacement of agriculture and mining in lands with high carbon stocks and plant species richness. We recommend the promotion and active support of land swaps by the government if land in the Non-Forest area, that is designated for mining or agriculture, contains forest high in carbon stock and/or species richness. Meanwhile, the maintenance of carbon and biodiversity can be promoted by conservation planning strategies, such as REDD+. These are shown to be very effective for carbon and biodiversity, but only if carbon and biodiversity are both incorporated in the spatial planning process. Tackling corruption and improving law enforcement in order to prevent illegal logging and anthropogenic fires and illegal encroachment of forested land in the Forest area zones is then also very important. Next to carbon and biodiversity maintenance, land should be zoned for food production versus cash crop cultivation, so as to ensure sufficient production of food in this remote rural area. Further research on this subject should focus on the impacts of LULC change on access to food and on how displacement of food production can be prevented. Part III: Chapter 7. Mitigation of unwanted LULC change In Chapter 7, we aimed to analyse whether and how the production of palm oil can be reconciled with the production of pulpwood, rice and rubber, while mitigating unwanted LULC change. We investigated the technical potential of four measures to mitigate unwanted LULC change between 2008 and 2020 under low, medium, and high scenarios, referring to the intensities of the mitigation measures compared with those implemented in 2008. These measures are related to land sparing through (i) the improvements of yields, (ii) chain efficiencies, (iii) chain integration, and (iv) the steering of any expansion of these commodities to suitable and available under-utilised (potentially degraded) lands. Our analyses resulted in a land-sparing potential of 0.4-1.2 Mha (i.e., 24-62% of the total land demand of the commodities) between 2008 and 2020, depending on the land-use projection of the four commodities and the scenario for implementing the mitigation measures. Additional expansion on under-utilised land is the most important mitigation measure (45-62% of the total potential), followed by yield improvements as the second most important mitigation measure (32-46% of the total potential). We conclude that reconciling the production of palm oil, pulpwood, rice and rubber with the maintenance of existing agricultural lands, forests and peatlands is technically possible only by: i) halting the projected linear or even exponential expansion of oil palm and other commodities in the region, ii) implementing responsible land zoning and enforcement so that only available and suitable under-utilised land is used for land development, and iii) increasing the resource efficiency and productivity of agricultural production by improving the capacity of smallholdings and their access to finance and markets. However, if these conditions are not met, the four mitigation measures that were analysed in this study will not be sufficient to mitigate unwanted LULC change. Moreover, the implementation and success of these measures strongly depend on political and societal awareness and

218

willingness to mitigate unwanted LULC change. Thus, strong law enforcement, policy implementation and the tackling of corruption are fundamental to the success of the mitigation of unwanted LULC change in North-East Kalimantan and elsewhere. Chapter 8. Discussion and synthesis In Chapter 8, we provide a synthesis of the dissertation. In this chapter, we reflect on our findings in the context of the Forest Transition Theory. Firstly, we estimate that our study region is only at the start of the forest transition curve, possibly in stage 2, meaning the presence of a high forest cover, a high annual deforestation and degradation rate, and a relatively low regeneration rate. The main contributing land use types to forest loss identified in the study region are smallholder rubber, oil palm plantations, pulpwood plantations and mixed cropland. The contribution of oil palm and rubber to total LULC change is expected to increase under various land development and land zoning scenarios. Secondly, we envision for the near future that a stabilising forest cover is possible, but only with a limit-to-growth to land development for the production of the main commodities. An integrated perspective on mitigating unwanted LULC change and stabilising a high forest cover includes significant efforts in: i) Establishing a limit-to-growth on land development for the production of palm oil, rubber, pulpwood and coal in order to mitigate unwanted LULC change and forest cover loss at the regional level; ii) Land development planning and zoning, accounting for the maintenance of forests and peatlands, the regeneration potential of these lands and the land tenure; iii) Developing or strengthening incentives to maintain forests; iv) Implementing additional measures, such as the improvement of yields, chain efficiencies and chain integration of the main commodities; v) Supporting the measures by capacity building and enabling policies. Stabilising forest cover loss in the study regions, without displacement of land uses elsewhere, is already incredibly challenging We therefore expect that it will not be possible for the study region or Indonesia to reach the reforestation stage, unless reforestation would involve plantation forest, which is not desirable from a biodiversity-perspective. Instead, priority should be given to maintaining and improving existing natural forests and peatlands and developing underutilised (degraded) lands for agriculture and mining. When forest cover increases, carbon accumulates over time. However, plant species populations that were diminished may not recover, species that went extinct locally may not return, and species with a limited distribution may become extinct all together. Thirdly, it is important to assess and monitor whether the mitigation measures are working and whether forest cover in the study region and in Indonesia nation-wide is stabilising. We recommend to apply a nation-wide landscape-level approach, using wall-to-wall remote sensing data to be able to monitor displacement of land use change and under-utilised lands. Remote sensing data should be analysed on an annual basis, so that LULC change trajectories incorporating a range of land uses can be clearly identified. It is important that the remote sensing data is validated by remote sensing data from other sources, statistics, thorough (drone-based) ground checks and expert knowledge. Additionally, transparency in and sharing of concession and land allocation zoning data is essential. Most importantly, non-profit organisations, communities, governments and companies need to work together on a trust-basis in responsible land use planning, in order to mitigate unwanted LULC change and stabilise a high forest cover.

219

220

Ikhtisar Bahasa Indonesia Perubahan penggunaan lahan di kawasan tropis Akibat semakin meningkatnya populasi dunia dan standar hidup secara keseluruhan, terdapat peningkatan permintaan terhadap pangan, pakan, serat, dan bahan bakar. Produksi komoditas untuk memenuhi permintaan tersebut membutuhkan lahan yang luas dan menyebabkan perubahan penggunaan dan tutupan lahan (land use and land cover/LULC). Perubahan LULC terutama terjadi di wilayah-wilayah pedalaman di kawasan tropis, di mana iklim, lahan yang luas, kepadatan populasi yang rendah, dan ketidakjelasan sistem penguasaan atas tanah memiliki potensi optimal bagi pengembangan lahan. Di sisi lain, perubahan LULC menghasilkan pendapatan bagi kelompok-kelompok tertentu di masyarakat, dari pihak perusahaan multinasional hingga perusahaan lokal, petani dan pemerintah. Pengembangan lahan juga mengakibatkan konflik penggunaan lahan, misalnya antara perusahaan dan masyarakat. Selain itu, perubahan LULC juga bisa menyebabkan berkurangnya tutupan hutan, spesies flora dan fauna, karbon yang tersimpan di tumbuhan dan tanah (stok karbon), dan produksi pangan setempat. Ini juga menjadi masalah di pulau-pulau luar Indonesia, tempat perubahan LULC yang tidak dikehendaki terjadi dalam skala besar akibat produksi minyak kelapa sawit, karet, kayu, bubur kayu, dan pertambangan. Dalam disertasi ini, kami menganalisis perubahan LULC yang terjadi di Kabupaten Kutai Barat dan Mahakam Ulu, bagian dari Provinsi Kalimantan Timur yang saat ini berada di bawah tekanan tinggi dalam hal pengembangan lahan. Kami menitikberatkan analisis lampau pada kurun waktu 1990 hingga 2009 dan analisis proyeksi masa depan pada kurun waktu 2010 hingga 2020/2030. Kami mengidentifikasi tipe-tipe penggunaan lahan utama yang terlibat dalam perubahan LULC dan kehilangan hutan, dan mengintegrasikannya dalam model perubahan penggunaan lahan untuk memperkirakan perubahan LULC di masa depan. Sebagai tambahan, kami memperlihatkan hasil perkiraan perubahan LULC dan dampaknya pada stok karbon di atas tanah dan spesies tanaman di wilayah penelitian. Akhirnya, kami menggunakan empat tolok ukur untuk meminimalisir perubahan LULC yang tidak dikehendaki. Dalam pendahuluan di Bab 1 dan sintesis di Bab 8, kami menempatkan penemuan kami di konteks Teori Transisi Hutan. Teori Transisi Hutan dan tujuan disertasi Menurut Teori Transisi Hutan, suatu negara atau wilayah mula-mula mengalami tahap tutupan hutan tinggi dan angka deforestasi rendah, kemudian seiring waktu kehilangan tutupan hutan, dan sesudah beberapa lama akhirnya tutupan hutan bisa stabil atau bahkan bertambah. Di Bab 1, kami mendeskripsikan pengembangan lahan untuk produksi minyak kelapa sawit dan komoditas tanaman lainnya yang mendatangkan keuntungan dan memiliki keterkaitan besar dengan hilangnya tutupan hutan di Indonesia dalam konteks Teori Transisi Hutan. Berdasarkan ini, kami menyusun pertanyaan riset utama disertasi ini, yakni: Bagaimanakah cara menganalisis proses dan pemicu perubahan LULC di lanskap tropis yang tengah bertransisi, dampaknya pada tutupan hutan, stok karbon, dan keanekaragaman hayati, dan mitigasi dampaknya? Di Bab 1, kami juga menyampaikan garis besar disertasi dan pertanyaan riset per bab. Bagian pendahuluan menunjukkan bahwa Indonesia adalah contoh sangat penting dari negara tropis yang menunjukkan perluasan perubahan LULC dan kehilangan hutan. Di Bab 2, kami membahas tentang wilayah penelitian di Provinsi Kalimantan Utara dan Timur, dan Kabupaten Kutai

221

Barat dan Mahakam Ulu untuk menunjukkan proses yang terjadi di pulau-pulau luar di Indonesia. Di bab ini, kami menguraikan mengenai kondisi biofisik dan sosial ekonomi dan proses-proses yang berhubungan dengan perubahan LULC di wilayah tersebut. Pembagian disertasi ke dalam tiga bagian Sisa bab disertasi dibagi ke dalam tiga bagian. Bagian I yang berfokus pada perubahan LULC dan dampaknya pada tutupan hutan di Bab 3 dan 4. Bagian II yang berfokus pada dampak pengembangan lahan dan kehilangan hutan pada stok karbon dan keanekaragaman hayati di Bab 5 dan 6. Bagian III yang berfokus pada tindakan untuk mengurangi perubahan penggunaan lahan yang tidak diinginkan di Bab 7. Bagian I: Bab 3. Perubahan LULC di wilayah penelitian antara 1990-2009 Di Bab 3, tujuan kami adalah memaparkan karakteristik, mengukur, dan memvisualisasikan proses dan pemicu perubahan LULC yang berkontribusi pada kehilangan dan degradasi hutan di Kabupaten Kutai Barat dan Mahakam Ulu di Kalimantan Timur sejak 1900-2009. Kami menerapkan analisis kuantitatif dan eksplisit secara spasial menggunakan peta LULC berbasis Landsat, yang didukung oleh informasi dari lapangan dan pengetahuan para ahli. Sepanjang kurun waktu yang dianalisis, sekitar sepertiga area penelitian berubah dan tutupan hutan berkurang secara substansial sekitar 9%, sebagian besar karena degradasi hutan dan deforestasi. Seiring waktu, bagaimanapun, penyebab utama bergeser dari degradasi hutan dan deforestasi serta pengembangan pertanian berskala kecil pada 1990-2000 menjadi peningkatan pengembangan pertanian monokultur berskala besar pada 2000-2009. Secara keseluruhan, perubahan LULC menjadi semakin dinamis dan kompleks seiring ruang dan waktu, dan ditandai oleh serangkaian perubahan, yang diistilahkan sebagai pemicu. Kami menyusun daftar jenis LULC yang mengganti tutupan hutan, yaitu perkebunan karet rakyat, perkebunan kayu, campuran antara lahan pertanian dan perkebunan kelapa sawit. Pertanian campuran dan perkebunan karet rakyat ternyata termasuk dalam tipe penggunaan lahan menengah, dengan kemungkinan konversi tinggi. Kami menyimpulkan bahwa bukan hanya hutan yang rawan konversi, melainkan juga penggunaan lahan campuran berskala kecil, yang mengakibatkan intensifikasi penggunaan lahan. Melihat pemicu perubahan LULC penting untuk memperbaiki perkiraan perubahan LULC di masa depan dan untuk mendukung kebijakan perencanaan wilayah. Kami merekomendasikan untuk melibatkan masyarakat dan industri lokal, terutama petani, dalam proses perencanaan wilayah, karena sebagian dari mereka bisa memegang peran penting dalam pengembangan pertanian dan/atau terpengaruh oleh perubahan LULC. Karena sebagian besar pertanian monokultur dan konsesi dikembangkan di kawasan yang telak dialokasikan dan selaras dengan perencanaan spasial pemerintah, pemerintah bisa dan sebaiknya memegang peran penting dalam memandu ekspansi pertanian ke arah yang lebih berkelanjutan, yang memperhatikan kelestarian hutan dan produksi pangan setempat. Bagian I: Bab 4. Perkiraan perubahan LULC di wilayah penelitian untuk 2012-2030 Di Bab 4, tujuan kami adalah menguraikan dampak kebijakan alokasi lahan dan dampak berbagai level pengelolaan lahan bagi LULC. Kami berfokus pada delapan penggunaan lahan untuk produksi mis. kelapa sawit, karet, kayu, dan lahan pertanian campuran, serta dampak perubahan LULC pada tutupan hutan dan produksi pangan setempat. Kami membuat model dan visualisasi dampak perubahan LULC pada tutupan hutan dan produksi pangan lokal di Kabupaten Kutai Barat dan Mahakam Ulu di bawah empat skenario yang saling bertolak belakang antara 2009 hingga 2030 menggunakan model PCRaster Land Use Change (PLUC) spasio-temporal.

222

Menurut hasil temuan kami, sebagian besar perubahan LULC antara 2009 dan 2030, menggunakan semua skenario, berdampak buruk pada hutan, perdu, dan padang rumput, meskipun setiap skenario memiliki efek yang berbeda. Dengan pengelolaan terbatas, perubahan LULC terjadi sebagian besar di dataran rendah di tenggara dan kehilangan hutan antara 2009 dan 2030 tergolong rendah (~0,1 Mha, ~4%). Di bawah skenario tidak terbatas-dijaga (unlimited-restricted), bagaimanapun, angka kehilangan hutan lebih besar dengan 0.4Mha (~17%), dan di bawah skenario tidak terbatas-tidak dijaga (unlimited-unrestricted), angka kehilangan hutan paling tinggi (1,6 Mha, ~60%). Untuk skenario terakhir, perubahan LULC bergeser dari dataran rendah ke wilayah lebih tinggi, dengan kehilangan lebih besar di hutan berkanopi terbuka menengah dan tertutup di wilayah ini. Sementara itu, di bawah dua skenario pengembangan tidak terbatas, lahan pertanian campuran dan produksi pangan di dataran rendah diganti dan dipindahkan ke daerah yang lebih tinggi, sehingga menyebabkan kehilangan hutan di sana. Pemindahan produksi pangan ke wilayah yang lebih tinggi tidak diinginkan, karena jauhnya lokasi lahan berketinggian menengah dan pegunungan di wilayah ini. Kami menyimpulkan bahwa untuk menjaga hutan primer, lahan gambut dan produksi pangan lokal, maka produksi kelapa sawit, karet, kayu, dan batu bara perlu diperlambat dan dilakukan di zona terbatas dan dengan perencanaan wilayah yang bertanggung jawab. Ini termasuk perpanjangan dan memperkuat moratorium yang berlaku saat ini, jika yang dituju adalah rekonsiliasi semua fungsi lahan yang berbeda-beda. Bagaimanapun, implementasi dari moratorium hanya berfungsi efektif jika perubahan area hutan menjadi non-hutan tidak diizinkan lagi atau dilarang keras untuk dilakukan di lahan berhutan atau lahan gambut, dan juga didasarkan pada ketersediaan lahan dan kesesuaian tanaman. Konservasi sebaiknya juga berfokus ke hutan dengan kanopi medium dan sangat terbuka, yang memiliki potensi regenerasi besar. Dengan pengelolaan terbatas, akan terdapat cukup rumput dan perdu untuk pengolahan lahan, meskipun penerapan metoda yang mantap seperti Metoda Pengelolaan yang Bertanggung jawab (Responsible Cultivation Method) dan validasi lapangan menyeluruh diperlukan untuk menjabarkan apakah penggunaan lahan yang dimaksud kurang layak, sehingga diperlukan pengembangan lahan yang bertanggung jawab. Bagian II: Bab 5. Hubungan antara biomassa di atas permukaan tanah dan variabel prediktif Pemahaman lebih mendalam terhadap variasi spasial biomassa di atas permukaan tanah (aboveground biomass/AGB) penting untuk meminimalisir emisi karbon dan perubahan iklim dunia dari deforestasi, degradasi hutan, dan ekspansi pertanian di kawasan tropis. Di Bab 5, tujuan kami adalah menganalisis bagaimana set variabel biofisik dan atropogenik berhubungan dan berkontribusi pada variasi spasial biomassa di atas permukaan tanah (AGB) di Kalimantan Utara dan Timur. Kami menerapkan regresi multi-linear non-spasial dan spasial, termasuk regresi geografis. Kami bisa menjelaskan variasi spasial tinggi AGB sekitar 59-64%. Lebih jauh lagi, Kami menemukan relasi positif yang kuat antara AGB dan variabel-variabel yang diuji: ketinggian, kemiringan, zona alokasi lahan, jenis tanah, dan jarak ke sumber api, jalan, sungai, dan kota terdekat. AGB relatif lebih tinggi di wilayah yang lebih tinggi, di tanah karst dan vulkanis, dengan jarak yang semakin jauh dari sumber api, di zona hutan produksi Terbatas, dan di zona proteksi DAS dan Konservasi. Karena faktor-faktor tersebut berefek besar pada variasi AGB, upaya untuk meminimalisir emisi karbon, seperti REDD+, sebaiknya mengikutsertakan faktor tersebut. Korelasi yang tinggi antara variabel yang dijelaskan menunjukkan bahwa variabel-variabel tersebut saling berhubungan, sehingga variasi AGB tidak bisa dijelaskan hanya dengan satu variabel. Sebaliknya, analisis spasial sebaiknya mengintegrasikan berbagai variabel biofisik dan antropogenik untuk memberikan pemahaman yang lebih baik mengenai variasi spasial dalam AGB.

223

Bagian II: Bab 6. Dampak skenario perubahan LULC pada stok karbon dan keanekaragaman hayati Di Bab 6, kami bertujuan memperkirakan dampak empat skenario perubahan LULC yang saling bertolak belakang, dengan berbagai level pengelolaan lahan dan berbagai batasan zona lahan pada karbon dan keaneragaman hayati tumbuhan di level lanskap. Sebagai tambahan, kami bertujuan menelaah apakah respon karbon dan keaneragaman hayati tumbuhan pada perubahan LULC serupa, dan bermaksud mencari area mana yang dianggap memiliki ancaman terbesar perubahan LULC. Kami menerapkan pendekatan eksplorasi, dengan menggunakan analisis spasial dan statistik untuk menghitung dampak, menggunakan data AGB yang ada dan kekayaan spesies tumbuhan dari literatur dan data AGB baru di wilayah studi kasus Kabupaten Kutai Barat dan Mahakam Ulu di Kalimantan Timur. Hasil kami menunjukkan bahwa antara 1990 dan 2009, AGB di wilayah itu turun secara signifikan dari 877 ± 223 Mt menjadi 732 ± 176 Mt, dan kekayaan spesies utama menurun secara signifikan dari ~406 spesies menjadi ~365 spesies per 10 x 10 km sel raster. Di bawah skenario tidak terbatas-tidak dijaga (unlimited-unrestricted), AGB diperkirakan turun lebih jauh menjadi 550 ± 143 Mt pada 2030 dan kekayaan spesies utama menjadi ~196 per kilometer sel raster pada 2030. Dampak pada jumlah spesies utama (-22% hingga -%52%) lebih kuat daripada dampak AGB total (-19% hingga -37%) antara 1990 dan 2030. Pada semua skenario, dataran rendah menunjukkan risiko tertinggi perubahan LULC. Seiring waktu, bagaimanapun, ancaman kehilangan hutan, termasuk kehilangan karbon dan keanekaragaman hayati, meningkat di wilayah yang lebih tinggi di barat laut, terutama pada skenario pengembangan tidak terbatas (unlimited). Solusi paling bertanggung jawab yang bisa diambil barangkali adalah memperlambat pengelolaan lahan dan menerapkan pembatasan zona untuk mencegah pengelolaan dan perubahan lahan pertanian dan pertambangan di kawasan yang memiliki stok karbon dan kekayaan spesies besar. Kami merekomendasikan pengenalan dan dukungan aktif pertukaran lahan oleh pemerintah jika lahan di kawasan Non-Hutan, yang ditujukan untuk pertambangan atau pertanian, ternyata memiliki hutan dengan stok karbon dan/atau kekayaan spesies yang tinggi. Sementara itu, pemeliharaan stok karbon dan kenaekaragaman hayati bisa dilakukan dengan strategi perencanaan konservasi semacam REDD+. Hal ini telah terbukti sangat efektif untuk karbon dan kenaekaragaman hayati, namun hanya jika karbon dan kenaekaragaman hayati tercakup di dalam proses perencanaan wilayah. Menangkal korupsi dan perbaikan undang-undang untuk mencegah penebangan ilegal, kebakaran dan perambahan oleh masyarakat di kawasan Hutan juga sangat penting. Selain menjaga karbon dan kenaekaragaman hayati, lahan juga harus dibagi menjadi zona-zona produksi pangan dan pembudidayaan tanaman perdagangan, untuk memastikan kecukupan produksi pangan di kawasan pedesaan/terpencil. Riset lebih jauh mengenai subyek ini akan berfokus ke dampak perubahan LULC pada akses pangan dan cara mencegah kehilangan wilayah produksi pangan. Bagian III: Bab 7. Mitigasi perubahan LULC yang tidak dikehendaki Pada Bab 7, kami bertujuan menganalisis apakah dan bagaimana produksi minyak kelapa sawit bisa diselaraskan dengan produksi bubur kayu, beras, dan karet, sembari melakukan mitigasi terhadap perubahan LULC yang tidak dikehendaki. Kami meneliti potensi teknis dari empat tolok ukur untuk melakukan mitigasi terhadap perubahan LULC yang tidak dikehendaki antara 2008 hingga 2020 di bawah skenario rendah, sedang, dan tinggi, dengan mengacu pada intensitas tolok ukur mitigasi dibandingkan dengan implementasi pada 2008. Keempat tolok ukur ini berkaitan dengan penghematan lahan melalui (i) perbaikan produksi, (ii) efisiensi rantai, (iii) integrasi rantai, dan (iv) pengarahan ekspansi komoditas ini ke lahan kurang layak (berpotensi terdegradasi) yang sesuai dan tersedia.

224

Analisis kami membuahkan potensi penghematan lahan 0,4-1,2 Mha (24-62% dari total permintaan lahan untuk komoditas) antara 2008 hingga 2020, tergantung pada perkiraan penggunaan lahan untuk keempat komoditas dan skenario penerapan tolok ukur mitigasi. Ekspansi tambahan ke lahan kurang layak merupakan tolok ukur mitigasi terpenting (45-62% dari total potensi), diikuti oleh perbaikan produksi sebagai tolok ukur mitigasi terpenting kedua (32-46% dari total potensi). Kami menyimpulkan bahwa menyelaraskan produksi minyak kelapa sawit, bubur kayu, beras, dan karet dengan pengelolaan lahan pertanian, hutan, dan lahan gambut yang telah ada secara teknis mungkin dilakukan hanya dengan: i) menghentikan proyeksi linear atau bahkan ekpansi eksponensial kelapa sawit dan komoditas lainnya di wilayah itu, ii) mengimplementasikan penetapan zona dan penggunaan lahan yang bertanggung jawab sehingga hanya lahan kurang layak yang sesuai dan tersedia yang bisa digunakan untuk pengembangan lahan, dan iii) menambah efisiensi sumber daya dan produktifitas pertanian dengan memperbaiki kapasitas petani dan akses mereka ke pendanaan dan pasar. Bagaimanapun, jika kondisi ini tidak tercapai, keempat tolok ukur mitagsi yang dianalisis dalam kajian ini tidak akan cukup untuk memitigasi perubahan LULC yang tidak dikehendaki. Oleh karena itu, penerapan undang-undang yang tegas, implementasi kebijakan, dan pemberantasan korupsi penting bagi keberhasilan mitigasi perubahan LULC yang tidak dikehendaki di Kalimantan Utara-Timur dan tempat-tempat lainnya. Bab 8. Diskusi dan sintesis Di Bab 8 kami menyajikan sintesis disertasi. Di bab ini, kami merefleksikan temuan kami dalam konteks Teori Transisi Hutan. Pertama, kami memperkirakan bahwa wilayah penelitian kami berada di awal kurva transisi hutan, kemungkinan di tahap 2, yang berarti keberadaan tutupan hutan yang tinggi, angka deferostasi dan degradasi tahunan yang tinggi, dan angka regenerasi yang relatif rendah. Tipe penggunaan lahan yang berkontribusi utama terhadap kehilangan hutan di wilayah penelitian adalah perkebunan karet berskala kecil, perkebunan kelapa sawit, hutan tanaman industri, dan lahan pertanian campuran. Kontribusi kelapa sawit dan karet pada keseluruhan perubahan LULC diperkirakan akan bertambah di bawah berbagai skenario pengelolaan dan zona lahan. Kedua, kami memperkirakan bahwa di masa yang akan datang stabilitas tutupan hutan mungkin didapatkan, namun hanya dengan pengembangan lahan dengan batasan pertumbuhan (limit to growth) untuk produksi komoditas utama. Perspektif terintegrasi terkait mitigasi perubahan LULC yang tidak dikehendaki dan stabilisasi tutupan hutan tinggi mencakup upaya-upaya yang nyata dalam: i) Menerapkan batasan pertumbuhan (limit to growth) pada pengelolaan lahan untuk produksi minyak kelapa sawit, karet, bubur kayu, dan batu bara dengan tujuan melakukan mitigasi terhadap perubahan LULC yang tidak dikehendaki dan kehilangan tutupan hutan di level daerah; ii) Perencanaan dan penetapan zona pengelolaan lahan, memperhitungkan pelestarian kawasan hutan dan lahan gambut, potensi regenerasi lahan tersebut dan hak atas lahan; iii) Mengembangkan atau memperkuat insentif untuk pelestarian hutan; iv) Mengimplementasikan tolok ukur tambahan, misalnya peningkatan produksi, efisiensi rantai, dan integrasi rantai dari komoditas utama; v) Mendukung tolok ukur dengan peningkatan kapasitas dan penerapan kebijakan. Upaya stabilisasi kehilangan hutan di wilayah penelitian, tanpa mengubah penggunaan lahan di tempat lain sudah luar biasa menantang. Karena itu kami memperkirakan bahwa mustahil bagi lokasi penelitian atau Indonesia untuk mencapai tahap reforestasi, kecuali membangun hutan tanaman industri, yang tidak dikehendaki dari perspektif biodiversitas. Prioritas justru harus diberikan pada upaya menjaga dan memperbaiki hutan alami dan lahan gambut yang ada dan mengembangkan lahan

225

kurang layak (terdegradasi) untuk pertanian dan pertambangan. Ketika tutupan hutan bertambah, stok karbon akan berangsur-angsur bertambah. Bagaimanapun, populasi spesies tumbuhan yang telah mati tidak akan bisa dihidupkan lagi, spesies yang punah secara lokal tidak akan bisa dikembalikan, dan spesies dengan distribusi terbatas kemungkinan akan punah. Ketiga, penting untuk menilai dan memonitor apakah tolok ukur mitigasi dapat bekerja dan tutupan hutan di lokasi penelitian atau di seluruh Indonesia mulai stabil. Kami merekomendasikan penggunaan pendekatan level lanskap di tingkat nasional, dengan data penginderaan jauh dindingke-dinding (wall-to-wall) untuk memonitor pemindahan perubahan penggunaan lahan dan lahan kurang layak. Data penginderaan jauh sebaiknya dianalisis setiap tahun, agar pemicu perubahan LULC yang terjadi di cakupan penggunaan lahan tertentu bisa diidentifikasi secara jelas. Penting untuk diingat bahwa data penginderaan jauh mendapatkan validasi dari data penginderaan jauh dari sumber lain, statistik, pemeriksaan medan secara menyeluruh (berbasis drone), dan pengetahuan para ahli. Sebagai tambahan, transparansi dan pembagian data konesi dan penetapan zona alokasi lahan sangat diperlukan. Yang terpenting adalah lembaga non-pemerintah, masyarakat, pemerintah, dan perusahan perlu bekerja sama dengan saling percaya.

226

Nederlandse samenvatting Landgebruiksverandering in de tropen De toenemende wereldbevolking en levensstandaard per persoon gaan hand in hand met de groeiende productie van bijvoorbeeld palmolie, rijst, papier en rubber. Grote arealen land zijn nodig voor de ontwikkeling van deze producten en dit leidt tot landgebruiksverandering, voornamelijk in de tropen. In de tropen lijken de aanwezigheid van veel land, een lage bevolkingsdichtheid en onduidelijke landrechten de optimale omstandigheden te creëren voor landontwikkeling. Enerzijds resulteren deze ontwikkelingen in inkomens voor bedrijven, boeren en overheidsactoren. Anderzijds leiden deze ontwikkelingen tot grootschalige ongewenste landgebruiksverandering, met als gevolg ontbossing, landconflicten en een afname van de lokale voedselproductie. In dit proefschrift hebben wij allereerst in kaart gebracht welke landgebruiksprocessen tussen 1990 en 2009 hebben bijgedragen aan ontbossing in de West Kutai en Mahakam Ulu districten. Deze twee districten staan onder grote druk van landontwikkeling en zijn onderdeel van de provincie Oost Kalimantan in het Indonesische deel van Borneo. Daarnaast hebben we de belangrijkste landgebruikstypen geïdentificeerd en deze geïntegreerd in een computermodel om toekomstig landgebruik in te schatten. Verder laten we de geprojecteerde landgebruiksveranderingen en de effecten op koolstofvoorraden en plantensoorten in het studiegebied zien door middel van kaartmateriaal. Uiteindelijk hebben we vier maatregelen getest waarmee ongewenste landgebruiksveranderingen kunnen worden verminderd. In de introductie in Hoofdstuk 1 en in de synthese in Hoofdstuk 8 plaatsen we onze bevindingen in de context van de Bostransitie-theorie (de “Forest Transition Theory”). Bostransitie-theorie en het doel van dit proefschrift Volgens de Bostransitie-theorie gaat een land of regio eerst door een stadium van een hoog bosareaal en een lage jaarlijkse ontbossing en vervolgens een sterker verlies van bos, waarna het bosareaal stabiliseert en mogelijk weer toeneemt. In Hoofdstuk 1 beschrijven we landontwikkelingen voor de productie van palmolie en andere producten in Indonesië en het verlies aan bos dat hiermee gepaard gaat in de context van de Bostransitie-theorie. Op basis hiervan hebben we de onderzoeksvraag voor dit proefschrift geformuleerd: Hoe kunnen landgebruiksprocessen in tropische gebieden en de onderliggende factoren, de gerelateerde effecten op bosareaal, koolstof en plantensoorten, en mitigatiemaatregelen geanalyseerd worden? In Hoofdstuk 1 geven we ook een overzicht van het proefschrift en beschrijven we de onderzoeksvragen per hoofdstuk. In Hoofdstuk 2, beschrijven we het studiegebied in Noord- en Oost-Kalimantan en de daarin gelegen West Kutai en Mahakam Ulu districten. Deze gebieden zijn representatief voor processen die plaatsvinden in de buitenste eilanden van Indonesië. Onderverdeling proefschrift in drie delen De overige hoofdstukken zijn onderverdeeld in drie delen. Deel I richt zich op landgebruiks verandering en de effecten op bosareaal in Hoofdstuk 3 en 4. Deel II richt zich op de effecten van landontwikkeling en ontbossing op koolstofvoorraden en plantensoorten in Hoofdstuk 5 en 6. Deel III richt zich op maatregelen om ongewenste landgebruiksverandering te verminderen in Hoofdstuk 7. Het proefschrift sluit af met de synthese in Hoofdstuk 8.

227

Deel I: Hoofdstuk 3. Landgebruiksverandering in het studiegebied tussen 1990-2009 Het doel van Hoofdstuk 3 was om landgebruiksverandering in West Kutai en Mahakam Ulu te typeren, te kwantificeren en te visualiseren. We hebben hiervoor een reeks aan kwantitatieve en ruimtelijkexpliciete methoden toegepast, ondersteund door veldgegevens en advies van experts. Tussen 1990 en 2009 is het landgebruik van bijna een derde van het studiegebied veranderd en 10% van het bosareaal verdwenen. We zagen echter een verschuiving van hoofdzakelijk ontbossing en bosdegradatie en de ontwikkeling van kleinschalige landbouw in 1990-2000, naar een toenemende ontwikkeling van grootschalige monoculturen in 2000-2009. Landgebruiksverandering werd dynamischer en complexer en werd gekarakteriseerd door een specifieke reeks aan veranderingen oftewel ‘trajecten’. De gewassen die het meest aan ontbossing hebben bijgedragen zijn rubber-, pulp- en oliepalmplantages, waarvan de laatste in toenemende mate. We concluderen dat niet alleen bossen kwetsbaar zijn voor omzetting, maar ook kleinschalige landgebruiken, waaronder kleinschalige voedselproductie. Als gevolg hiervan vindt landintensivering plaats. We adviseren om industrieën en lokale gemeenschappen, waaronder kleinschalige boeren, in het ruimtelijke planningsproces te betrekken. Zij spelen een belangrijke rol in landgebruiksverandering en/ of kunnen door landgebruiksverandering beïnvloed worden. De meeste monoculturen en concessies zijn ontwikkeld in de aangewezen landallocatiezones, en dus in lijn met de ruimtelijke planning door de overheid. Om deze reden kan de overheid een belangrijke rol spelen in het verduurzamen van landbouw, waarbij bos en lokale voedselproductie behouden blijven. Deel I: Hoofdstuk 4. Landgebruiksverandering geprojecteerd in het studiegebied tussen 2010-2030 Het doel van Hoofdstuk 4 was om de effecten van landzonering en landontwikkeling op landgebruik en bosareaal te onderzoeken. We hebben ons specifiek gericht op acht landgebruikstypen voor de productie van onder andere palmolie, rubber en pulp en de effecten daarvan op bosareaal en lokale voedselproductie. We hebben deze effecten gemodelleerd voor de West Kutai en Mahakam Ulu districten onder vier scenario’s voor de periode 2009-2030 met het PCRaster landgebruiksveranderingsmodel PLUC. De vier scenario’s waren: 1) beperkte ontwikkeling – strikte zonering, 2) beperkte ontwikkeling – niet-strikte zonering, 3) ongelimiteerde ontwikkeling – strikte zonering, 4) ongelimiteerde ontwikkeling – niet-strikte zonering. De resultaten voor 2009-2030 laten een sterke afname in bosareaal (4-60%) zien met een groot verschil tussen de scenario’s. Met beperkte ontwikkeling van oliepalm-, pulp- en rubberplantages en koolmijnen vindt relatief weinig ontbossing plaats (~0.1 Mha, ~4%) en voornamelijk in de laaglanden. Met ongelimiteerde ontwikkeling en strikte landzonering vindt sterkere ontbossing plaats (~0.4 Mha, ~17%), terwijl met ongelimiteerde ontwikkeling en niet-strikte landzonering de meeste ontbossing plaatsvindt, namelijk ~1.6 Mha (~60%). Onder de twee laatst genoemde scenario’s vindt ook een verschuiving plaats van voedselproductie naar de hooglanden. Hieruit concluderen we dat de ontwikkeling van de eerder genoemde landgebruikstypen moet worden afgeremd en dat deze moet plaatsvinden onder strikte landzonering en verantwoorde ruimtelijke planning. Dit omvat de verlenging en versterking van het moratorium op nieuwe concessies. Echter, het moratorium kan alleen effectief zijn als bebost gebied of veenland dat geclassificeerd is als Bos-zone (“Forest area”) niet meer omgezet kan worden naar een niet-Bos-zone voor ander gebruik (“Non-forest area”) volgens het beleid op landzonering. Daarnaast moet de bescherming van bos zich niet alleen richten op onaangetast bos, maar ook op gedegradeerd bos met een medium tot open bladerendek. Met een beperkte ontwikkeling van de eerdergenoemde landgebruikstypen, inclusief oliepalm, pulp, rubber en mijnbouw, is voldoende grasland aanwezig voor deze ontwikkeling. Dan is het echter van belang dat wordt onderzocht of deze gebieden al in gebruik zijn, bijvoorbeeld met methoden als de “Responsible

228

Cultivation Areas method”, oftewel de methode gericht op de ontwikkeling van verantwoorde landbouwgebieden. Deel II: Hoofdstuk 5. Relatie bovengrondse biomassa en verklarende variabelen Meer inzichten in de ruimtelijke variatie van bovengrondse biomassa zijn nodig om koolstofemissies en klimaatverandering als gevolg van landbouw en ontbossing in de tropen te verminderen. Het doel van hoofdstuk 5 was om de relatie tussen variatie in bovengrondse biomassa en een set van biofysische en antropogene variabelen te onderzoeken in Noord- en Oost-Kalimantan. We hebben hiervoor multiple lineaire regressie toegepast, waarmee we ongeveer 59-64% van de ruimtelijke variatie in bovengrondse biomassa konden verklaren. Bovengrondse biomassa was relatief hoger in de hooglanden, op karst en vulkanische gronden, met een toenemende afstand tot branden en in drie specifieke zones die gericht zijn op behoud of minimaal gebruik van bos. Initiatieven zoals REDD+ (het reduceren van emissies door ontbossing en bosdegradatie) zouden deze factoren moeten meenemen. Deel II: Hoofdstuk 6. Effecten van landgebruiksverandering op koolstofvoorraden en plantensoorten Hoofdstuk 6 had als doel om de effecten te onderzoeken van de vier eerdergenoemde scenario’s op koolstofvoorraden en plantensoorten. Daarnaast hebben we onderzocht in welke gebieden het grootste risico bestond op verlies in koolstof en soortenrijkdom. We hebben hiervoor statistische en ruimtelijke methoden toegepast met data uit de literatuur en data verzameld in het studiegebied in Oost Kalimantan. Onze resultaten laten zien dat tussen 1990 en 2009, bovengrondse biomassa significant is afgenomen van ongeveer 877 naar 732 Mt en soortenrijkdom van ongeveer 406 tot 365 soorten per rastercel van 10x10km. Verder verwachten wij dat in 2030, onder het scenario van ongelimiteerde ontwikkeling en niet-strikte landzonering, biomassa verder afgenomen is tot ongeveer 550 Mt en soortenrijkdom naar ongeveer 196 soorten per rastercel. Het effect op soortenrijkdom (-22% tot -52%) is sterker dan het effect op biomassa (-19% tot -37%) tussen 1990 en 2030. De meest verantwoorde ontwikkeling zou inhouden dat landontwikkeling wordt afgeremd en dat strikte landzonering wordt toegepast, zodat de ontwikkeling van landbouw en mijnbouw niet plaatsvindt in gebieden met hoge koolstofvoorraden en soortenrijkdom. We adviseren dat concessies die uitgegeven zijn in gebieden met hoge koolstofvoorraden en hoge soortenrijkdom worden ingetrokken en opnieuw uitgegeven in gebieden waar deze laag zijn. Initiatieven zoals REDD+ blijken effectief voor het behoud van koolstof en soortenrijkdom, maar alleen als beide factoren worden meegenomen in ruimtelijke planning. Het aanpakken van corruptie en het verbeteren van wethandhaving ter voorkoming van illegale houtkap en bosbranden is hiervoor cruciaal. Naast landzonering ten behoeve van koolstof en soortenrijkdom, is landzonering voor voedselproductie ook heel belangrijk, vooral in dit afgelegen plattelandsgebied. Deel III: Hoofdstuk 7. Vermindering van ongewenste landgebruiksverandering In Hoofdstuk 7 hebben we onderzocht hoe palmolie, pulp, rijst en rubber geproduceerd kunnen worden in Noord en Oost Kalimantan zonder ongewenste effecten. We hebben hiervoor vier maatregelen onderzocht, waaronder i) het verhogen van de opbrengst van ieder gewas per hectare, ii) het verbeteren van efficiëntie in de productieketen, iii) het verbeteren van bijproducten in de productieketen en iv) het gebruik van beschikbare en geschikte braakliggende (mogelijk gedegradeerde) gronden.

229

Onze resultaten laten tussen 2008 en 2020 een landbesparingspotentieel van ongeveer 0.41.2 Mha zien. Het gebruik van braakliggende gronden en de verhoging van de opbrengst van de gewassen bleken de meest belangrijke mitigatiemaatregelen (respectievelijk 45-62% en 32-46% van het totale potentieel). Uit dit onderzoek concluderen wij dat de productie van palmolie, pulp, rijst en rubber alleen kan samengaan met het behoud van bestaande landbouw, bossen en veengebieden als: i) de geprojecteerde lineaire of zelfs exponentiële groei van oliepalm en andere gewassen in de regio afgeremd wordt, ii) verantwoordelijke landzonering wordt uitgevoerd, zodat alleen beschikbare en geschikte braakliggende gronden gebruikt worden voor landontwikkeling en iii) de efficiëntie en productiviteit van landbouw worden verhoogd door het verbeteren van de capaciteit van boeren en hun toegang tot financiering en de markt. Echter, als aan deze condities niet wordt voldaan, dan zullen de vier geanalyseerde maatregelen niet voldoende zijn om ongewenste landgebruiksverandering te verminderen. Het succes van de maatregelen is sterk afhankelijk van politiek en maatschappelijk bewustzijn en bereidwilligheid om deze ongewenste effecten te verminderen. Strikte handhaving van de wet, implementatie van beleid en het aanpakken van corruptie zijn fundamenteel voor het succes van de mitigatie van ongewenste landgebruiksverandering in Noord- en Oost-Kalimantan en daarbuiten. Hoofdstuk 8. Discussie en synthese In Hoofdstuk 8 geven we een synthese van het proefschrift. In dit hoofdstuk reflecteren we op onze bevindingen in de context van de bostransitie-theorie. Allereerst is onze inschatting dat het studiegebied, met het grote bosareaal en de grootschalige ontbossing, nog maar aan het begin van de bostransitie-curve staat, mogelijk in fase 2. De meest belangrijke gewassen in dit studiegebied zijn oliepalm- en pulpplantages, kleinschalige rubberplantages en kleinschalige gemixte gewassen. We verwachten dat in de toekomst de bijdrage van oliepalm en rubber aan totale landgebruiksverandering toeneemt. Ten tweede concluderen wij dat verdere ontbossing in de nabije toekomst voorkomen kan worden, echter alleen door significante inspanningen ten aanzien van: i) een limiet op de ontwikkeling van de belangrijkste producten palmolie, rubber, pulp en steenkool; ii) verantwoordelijke planning en zonering van land, daarbij in beschouwing nemend het behoud van bossen en veengebieden en hun potentieel voor herstel; iii) het ontwikkelen of versterken van initiatieven om bossen te beschermen; iv) het implementeren van aanvullende maatregelen, zoals het verhogen van de opbrengst van landbouw, efficiëntie in de productieketen en het gebruik van bijproducten in de keten; v) ondersteunende capaciteitsopbouw en beleid. Het stabiliseren van bosareaal, zonder de verplaatsing van landbouw naar andere gebieden, is op zichzelf al een enorme uitdaging. We verwachten daarom dat het niet mogelijk zal zijn voor het studiegebied of Indonesië om de herstelfase te bereiken, tenzij dit inhoudt dat houtplantages meegerekend worden tot het bosareaal. Dit is echter niet gewenst vanuit een biodiversiteitsperspectief. In plaats daarvan verdienen het behoud en herstel van bestaande natuurlijke bossen en veengebieden prioriteit, en ook de ontwikkeling van braakliggende mogelijk (gedegradeerde) gronden voor land- en mijnbouw. Wanneer bos toeneemt, nemen ook de koolstofvoorraden toe. Echter, plantensoorten die lokaal uitsterven, kunnen niet altijd herstellen. Ten derde is het belangrijk om te testen en monitoren of de mitigatiemaatregelen werken en of het bosareaal in het studiegebied en in Indonesië gelijk blijft. Ons advies is om satellietgegevens op regionaal niveau te onderzoeken op de aanwezigheid van braakliggende gronden en de mogelijke verplaatsing van landbouw. Satellietgegevens zouden jaarlijks geanalyseerd moeten worden, zodat

230

allerlei kleinschalige landgebruiksveranderingen bepaald kunnen worden. Validatie van deze satellietgegevens met behulp van satellietgegevens op andere schaalniveaus, uitgebreide veldgegevens en kennis van experts is fundamenteel. Daarnaast zijn transparantie in de informatie over concessies en landzones en het delen van gegevens daarover essentieel. Het meest belangrijk is dat non-profit organisaties, onderzoekers, gemeenschappen, overheden en bedrijven met elkaar samenwerken op het terrein van duurzame landgebruiksplanning, om ongewenste landgebruiksverandering tegen te gaan en een groot bosareaal in stand te houden.

231

232

Acknowledgements / Dankwoord I always tell my friends and family that I’ve been through more challenges in the past 5 years of my PhD, than in the first 30 years of my life. And it’s true. It has been one roller coaster ride. Nonetheless, an exciting one, and the challenges were always followed up by good times, laughter, joy and optimism. The bumps in the road enabled me to grow professionally and personally and I look back with gratitude and a big smile. The following persons walked with me on this journey, supported me, and sometimes even cried & laughed with me. And I’m sincerely grateful to all. (the order of appearance doesn’t necessarily reflect the importance! And if your name is not listed, please forgive me in these busy times of finalising my dissertation) First and foremost, a big thanks to my supervisors, Pita, Stefan and André. Pita, many thanks for the opportunity to do this PhD in the first place. I still remember that you contacted me in 2009, telling me that an opportunity came up to do a PhD with the topics palm oil – forests – Indonesia. This combination of topics was exactly what I wished for in a PhD. Also thank you so much for your helpful feedback and moral support, particularly in times that I wasn’t sure anymore whether a PhD was really my thing. Stefan, thank you so much for joining the supervision team later on. Your everlasting smile, optimism, and at the same time useful feedback was so incredibly motivating and useful. You were the ideal process supervisor, bringing structure and clarity in my goals and workflow. André, thank you for all your support and for your positive and motivating spirit. You could always sketch the bigger picture of my research and then suddenly things became so clear. Many times you told me that I could do this, motivating me to do it even better. I am grateful to Aisha for supporting me along the way and in the last months of finalising my dissertation. Thank you so much Margot Stoete for supporting me with the layout of my (this) dissertation, and Ton Markus for supporting me with the layout of the figures. I also thank Siham, Petra, Fiona, Ineke and Annemarieke for their valuable back-up support throughout the years. Fulco Teunissen, thank you for the English proofreading of some of my articles. Antie and Rizki, thanks for helping me with the translation of my summary to Bahasa Indonesia. You all did such a great job! Maarten Zeylmans, thank you for the valuable GIS support throughout the years. Many sweet thanks to my lovely ‘roomy’ Rosalien. Your positive spirit, warmth and sincerity brought (and still brings ;)) so much joy to the office. I will miss our chats and laughter. Of course our journey together doesn’t end by me leaving the office. See you soon! A big thank you also to all my other roomies: Asier, Barano, Nathalie, Bart and Michiel. Thanks for all the laughter and in-depth conversations we had in our office ;)

233

Many thanks to the students with whom I was able to collaborate: Inge, Rianne, Elgin, Alexandra, Leah and the late Nina. Not only did you help me to improve my teaching and supervision skills, but you also held up a mirror. You were the best teachers ever.. ;) Thanks to all the people and organisations I collaborated with in the past years. First of all, Paul, Ari, Suseno, Pak Oka, Pita and Annelies, I am thankful that I was part of our Agriculture Beyond Food ‘Sliding from greasy lands’ project team. I hope our collaboration and inspiring discussions will continue after we finish our sub-projects! I am also indebted to Cora Govers, Sikko Visscher, Huub Loffler, Jacqueline Vel, Loes van Rooijen and many other people who contributed to the NWO/KNAW Agriculture Beyond Food program. Niels Wielaard and Marcela Quiñones of SarVision and Satelligence, you’ve delivered quite some fundamental data and knowledge to my research for which I am thankful. Our brainstorming sessions were also very useful. I gratefully acknowledge the contributions of the co-authors to my articles: Birka, I learned so much from our collaboration and from your fundamental input to the ILUC-report and paper. Judith, Laurens and Maria, thanks for our pleasant collaboration and for sharing your expertise with me. Arif Budiman, Wiwin Effendy and Arif Data Kusuma, thank you for your useful input to the articles and for making it possible for me to visit the field. Hans Smit, I enjoyed our collaboration on the paper about the Responsible Cultivation Areas and I appreciate it that our collaboration on this topic continues. I am also very grateful for my collaboration with many knowledgeable and passionate employees of World Wide Fund for Nature in Indonesia, including Barano, Arif, Wiwin, Data, Paul, Yudi, Syamsuardi, Ernawati and many other staff in Jakarta, Pekanbaru, Samarinda, Melak and beyond. Many thanks are also due to Rob Stuebing, Monica, Emile Jurgens and the staff of Universitas Mulawarman for our collaboration and all the support received in Indonesia. So many thanks go to my colleagues, many of whom I consider as friends, including Ana, Anna, Atse, Barano, Barbara (Let’s do this! ;)), Bhavya, Boudewijn, Boudewijn, Dorith, Floor, Geert, Gert-Jan, Gijs, Heleen, Hu Jing, Jesús, Jonathan, Joeri, Lotte, Marnix, Mijndert, Niels, Ody, Rene, Ric, Rodrigo, Sara, Sarah, Will and many more. Thank you all – and many other colleagues whom I haven’t mentioned by name – for making the van Unnik building a fun and inspiring place to work! David and Iris, thanks for co-organising the sustainability drinks. It was fun being part of our “sustainable drinking team”! Let’s keep this tradition! I also would like to thank my former colleagues Akshay, Anne Sjoerd, Bothwel, Ingeborg, Jos, Krishna, Desirée, Hans, Loek, Marlinde, Michiel, Sanne, Steven (thanks for the English check!) and Takeshi. Also Janske and Bart, thanks for the fun diners and our journey through Turkey, including a weekend in Istanbul! I would also like to express my gratitude to Niels Raes, Ferry Slik and Erik Meijaard and the (anonymous) referees and reading committee for their valuable advice and comments on the manuscripts and dissertation.

234

I am very grateful to my dear friend Pepi and her lovely family, who were always there for me when I arrived in and travelled through Indonesia. You created a second home for me, with a comfy bed and delicious Indonesian food. I am also grateful to the KEHATI group, Rina, Tillah, Juanita, Mega, Yonky, Indah, Juneka, Candra, Ali and many more! Veel dank aan mijn lieve tantes, ooms en neven en nichten! Ook lieve dank voor mijn lieve vriendinnen en vrienden Kit & Slave, Lin & Arnout, Cis & Mike, Nath & Robin, Pascal & Loes, Rianne, Mimi, Petra, Frouk, Bapke, Erik, Annemiek, Janske, Rosalien, Renée, Desmond, Thijs en vele-vele anderen. Mét jullie waren de hobbels tijdens mijn PhD zoveel minder bumpy. Dank jullie voor de mooie festivals, sushi-moments en warme knuffels tussendoor. Mara, thank you also for helping me with the statistics! Jenny, we met while organising a SENSE workshop and became friends over the years! Thanks for all our enjoyable talks. Mijn lieve Aikido- en yogavrienden: Aikido en yoga geven de beste workout en mindfulness voor promovendi ! Dank jullie voor de fijne trainingen, seminars én voor de leuke activiteiten daarna. Carla, Ruud, Steven en de lieve Huizenga & Kasten familie, dank jullie voor alle steun en mooie herinneringen aan de 10 jaar richting mijn PhD. Lieve Lukas, thank you for all the support you gave me in the last year of my PhD. I appreciate your patience and I enjoyed and still enjoy our loving, relaxing and fun times together, including our cycling tours, cooking sessions and all the board games you ‘learned’ me ;) En dan ‘last, but not least’, Kees, Marianna, Annemieke, Jelle, Sara, Marjolijn, Marijn en Lars: Jullie hebben mij écht bij elke stap gesteund en, waar nodig, geholpen. Paps & mams, dank jullie voor al jullie steun en alles wat jullie mij gegeven hebben. Jullie hebben nooit aan mijn keuzes getwijfeld en gaven mij kracht om mijn gevoel te volgen en door te zetten. Lieve zussen Annemieke en Marjolijn, dank jullie voor jullie steun op zoveel verschillende manieren, voor onze mooie en leerzame momenten én voor het voorhouden van een kritische spiegel. Ik heb zoveel van onze zussenband geleerd en kijk ernaar uit om samen verder te groeien. Lieve Jelle & Sara, dank jullie voor wie jullie nu al zijn, voor jullie kinderlijke, maar ó zo slimme en grappige inzichten en voor de weekendjes logeren, inclusief de fietstochtjes door Utrecht! En natuurlijk Lars, ik ben benieuwd naar alle mooie avonturen die wij samen nog gaan beleven! Mét mijn lieve familie, vrienden en collega’s was en is elke keuze en elke stap zoveel makkelijker. Dank jullie allen!

235

236

Curriculum Vitae Carina van der Laan was born on June 3, 1981, in Amsterdam, the Netherlands. She studied Sustainable Development, track Land Use, Environment and Biodiversity, at Utrecht University (2005-2008). Her MSc thesis focused on the development of sustainable cassava agriculture as an alternative to illegal logging and encroachment in the Tesso Nilo Landscape in Sumatra. In this project she worked together with WWF Indonesia and TDI in Wageningen. Prior to her master’s studies, she studied Life sciences and Chemistry at the University of Applied Sciences Utrecht (Hogeschool van Utrecht, 1999-2003), after which she worked at the Vrije Universiteit Amsterdam for two years. After her master’s studies, she worked for WWF Netherlands, as an Advisor to a Nature and Poverty program in Kenya, Mozambique and Cameroon. In 2010, she started her PhD research at the Copernicus Institute of Sustainable Development of Utrecht University on a part-time basis. The project focused on the challenges and opportunities of reconciling sustainable agricultural development, particularly for oil palm, rubber and pulpwood plantations, with the maintenance of forests, carbon stocks, biodiversity and food production in Kalimantan, Indonesia. To this purpose, she conducted field work on land use and land cover (change), forest degradation and carbon stocks. Additionally, she collaborated in the ILUC prevention project together with Birka Wicke and contributed to the development of the Responsible Cultivation Area method of Smit et al. (2013). This method is based on the principles and criteria of the Roundtable for Sustainable Palm Oil (RSPO). The results of her PhD work are presented in this dissertation. She recently contributed to a special broadcast on the ongoing forest fires in Indonesia of NOS Nieuwsuur on national television and of ‘Bureau Buitenland’ on national radio. Over the last few months, she started working as an independent consultant on sustainable development of agriculture and maintenance of forests in the tropics. Her current projects are taking place in Papua, Indonesia, and Ghana.

Publications Mitigation of unwanted direct and indirect land-use change – an integrated approach illustrated for palm oil, pulpwood, rubber and rice production in North and East Kalimantan, Indonesia. Global Change Biology Bioenergy 1-16, 2016. Authors: Carina van der Laan, Birka Wicke, Pita Verweij, André Faaij 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 September 19, 2014. Authors: Carina Van Der Laan, Marcela Quiñones, Pita Verweij, André Faaij

237

Breaking the Link between Environmental Degradation and Oil Palm Expansion. PLoS ONE September 12, 2013. Authors:, Hans Harmen Smit, Erik Meijaard, Carina Van Der Laan, Arif Budiman, Stephan Mantel, Pita Verweij Exploratory Study on the Use of ALOS PALSAR to Assess Aboveground Tree Biomass of Degraded Tropical Forests: a Case Study in East Kalimantan. 19th European Biomass Conference and Exhibition June 2011. Authors: Carina Van Der Laan, Niels Wielaard, Deddy Hadriyanto, Pita Verweij Can Community Forestry contribute to livelihood improvement and biodiversity? World Wide Fund for Nature, the Netherlands May 2010. Authors: Hans J.J. Beukeboom (WNF), Carina Van Der Laan (WNF), Arnold van Kreveld (Ulucus Consultancy), George Akwah (Consultant) Sustainable cassava agriculture in support of nature conservation in the Tesso Nilo Landscape, Sumatra. MSc Thesis, Universiteit Utrecht, WWF Indonesia February 2008. Author: Carina Van Der Laan. Award: Van Heest Thesis Award for Nature Conservation 2010

238