Stocktake and baseline data analysis for future landscape management and monitoring in West Kalimantan

This publication is part of the Sentinel Landscape network initiative established in eight sites around the world representative of widely different biophysical and socioeconomic contexts. Here we present and summarize the results of the research and baseline studies carried out in West Kalimantan, Borneo. Within the Kapuas Hulu regency, two districts were selected as ‘sentinel sites’: (1) a traditionally managed landscape with direct influence on watersheds to the north of Danau Sentarum National Park wetlands (Batang Lupar); and (2) a contrasting area in the south with improved infrastructure along the mainsouthern road from Sintang to Putussibau, to bring a different context (Mentebah).


Introduction
Long-term socio-ecological research is lacking when it comes to understanding the impacts of forest and tree cover changes on ecosystem services, diversity and social well-being. To address this gap, an evidencebased research initiative drew on several tropical long-term research sites, called 'sentinel landscapes', under the CGIAR Research Program on Forests, Trees and Agroforestry (FTA). The ambition was to monitor these sentinel landscapes and sites so as to guide future landscape management choices and development at district, regency and provincial levels, as well as enable comparisons between countries.
The main objective of this report is to introduce the large amount of data compiled and initiatives undertaken by CIFOR and partners in the Borneo sentinel landscape region between 2010 and 2017, to identify eventual gaps in socio-ecological system monitoring, and to make recommendations for future research priorities.

Context
A core region of the Borneo-Sumatra Sentinel Landscape, the Kapuas Hulu regency (31,162 km 2 ), represents the last forest frontier in West Kalimantan province. The regency is located in the northeastern part of the To curb extensive illegal logging occurring in the 1990s up to 2003s, and also to secure its hydrological tower function for the rest of the province, the regency was declared a 'conservation district' in 2003. In 2018 it was as registered as an official UNESCO-MAB (Man and the Biosphere Programme) biosphere zone. Various institutions and projects have since attempted to contribute to solving landscape management, conservation and development in the regency. Kapuas Hulu's total population is 263,207, with a population density of approximately eight inhabitants per km 2 . Ethnically and culturally the area is very diverse; Malay inhabit the wetlands and floodplains, with Dayak generally inhabiting the interior hills. The use of river transportation coupled with poor road infrastructure results in high economic costs, affecting both the price of goods and services. Approximately 70% of the population still works in agriculture.
In Kapuas Hulu, forest designated for conservation and watershed protection makes up approximately 57% of the regency. The two national parks occupy about 30% of the area. Land allocated for alternative use (APL) increased in 2013, and is now available for agricultural development, in particular for the establishment of oil palm plantations. Local farmers manage land using traditional methods, practicing swidden agriculture to grow crops such as paddy, maize, cassava, tubers and vegetables. The agricultural land, acquired by clearing forest or secondary forest, is systematically left to fallow after the paddy has been harvested. Perennial crops, mainly rubber, provide a ready source of cash and are often grown to supplement other crops such as pepper, fruits and tengkawang (illipe) nuts. The forest is vitally important for Kapuas Hulu's population, especially for food (meat, fruit and vegetables), medicines and traditional crafts (baskets and mats). In general, although economic activities in the district and rural areas are emerging, subsistence activities still play a central role, for instance, food provisioning from fallows, fishing and the gathering of forest products.
Oil palm plantations have expanded across West Kalimantan in recent years and are now advancing within Kapuas Hulu, with the establishment of oil palm plantations spreading from the west of the regency since 1998, reaching the 'biodiversity corridor' between the two national parks in the north, and spreading rapidly in the south. The potential impact that expanding oil palm plantations have on biodiversity and ecosystem functions presents a major challenge to the regency, faced with the double challenge of mitigating the impacts Within the regency, two districts were selected as 'sentinel sites': (1) a traditionally managed landscape with direct influence on watersheds to the north of Danau Sentarum National Park wetlands (Batang Lupar); and (2) a contrasting area in the south with improved infrastructure along the main southern road from Sintang to Putussibau, to bring a different context (Mentebah). Any transformation within these landscapes could impact both the integrity of the wetland ecosystem and the communities living there.

Methods
The baseline sampling methodology consisted of several components: Institutional mapping, multi-stakeholder platforms, and stakeholder engagement using the participatory prospective analysis (PPA) approach -a semi-quantitative, expert-based approach, designed to ensure balanced integration of a diversity of perspectives, relying on the assumption that stakeholders from different backgrounds with a shared interest in the same system can interact in a way that will reveal a common vision (Bourgeois et al. 2017a).
A vegetation survey, using an equally stratified sampling design to establish a baseline for future landscape monitoring, covering natural forest, old secondary forest, old fallow, young fallow, jungle rubber and mixed gardens. The sampling unit size was 20 x 20 m for trees with diameter above 5 cm.

Figure 3. Harvesting paddy among the weeds in Batang Lupar
Credit: Yves Laumonier/CIFOR ix Soil surveys, using methods prescribed by the Land Degradation Surveillance Framework (LDSF, Vågen et al. 2010), consisting of modules for soil, landform and land cover classification.
Socio-economic baseline surveys, using research and survey tools from the International Forestry Resources and Institutions program (IFRI 2013). Data were collected in ten randomly selected villages in each site, with random sampling undertaken in each village.

Results
The results of the participatory prospective analysis highlighted several plausible scenarios for development. The finally selected scenario and associated plan of action, both collaboratively made, were characterized by an explicit narrative emphasizing: policies made jointly with the community; the public participating through monitoring and supervising the planning process; improved access to education; and changes in people's behavior, toward more environmentally sound development. A road map was developed, containing guidelines for the implementation of desired scenarios, outlining preventive and anticipatory actions to mitigate undesired scenarios.
From 80% in 1973, Kapuas Hulu's forest cover remained significant in 2019 (73%, e.g. 66% intact and 7% logged-over). Between 2000 and 2010, 15% of mixed peat swamp forest, 7% of lowland mixed dipterocarp forest and 7% of freshwater swamp forest were logged. The opening of forest for swidden agriculture (ladang, food crop fields) remains minimal. The land conversion into oil palm plantation that began around 1998 near Badau, was not at the expense of forest. Representing just 2% between 2000 and 2010, forest conversion to oil palm increased after, notably to the detriment of the mixed peat (6%) and peat (6%) swamp forests in the regency's south-west.
Comparison between sentinel sites on the north and south of the Sentarum wetlands showed significant differences. Although not much difference was seen in the number of

. A bird view of a typical swidden landscape in Batang Lupar
Jungle rubber and tengkawang (illipe nut) forest near the river, surrounded by burn fields for food crops, shrubs and young fallows. Degraded soils on slopes after swidden cultivation as indicated by the fern cover. A rare occasion of additional rain fed paddy field in the low lying areas.
Credit: Yves Laumonier/CIFOR food crop fields or the proportion of plot size under fallow, the land tenure situation was very different in the two sites. The size of land owned without a legal document in Batang Lupar was 2.5 times higher than in Mentebah. Land was first opened in 1935 in Batang Lupar, 25 years before Mentebah (1960). The pattern of Batang Lupar field opening intensity increased and decreased every 5 to 10 years, while in Mentebah it showed to continually increase, probably in relation to rapid development along the main road. On average, plots were opened up to 3 km away from the village, and were under 3 ha in size. Dietary diversity across both sites was relatively similar. Batang Lupar households (which are more traditional) were relatively secure according to their food security score (mostly in the 3-4 range) while Mentebah (with more modern influences) showed high levels of inequality between households. Some villages experienced extreme food insecurity.
The differences in the development trajectories are linked to different cultural, socio-ecological and historical contexts. Batang Lupar sentinel site is located in the vicinity of two national parks, where land is more restricted as communities face unclear boundaries with Betung Kerihun National Park to the north and, to a lesser extent, Danau Sentarum National Park to the south. All villages are mostly inhabited by ethnic Dayak (Kantu, Iban and Tamanbaloh) who demonstrate stronger customary behaviors than Mentebah communities in the south. The communities are more dependent on forest resources for both timber and nontimber forest products. Rubber production and other agroforestry commodities are also important, like fishing; while local employment opportunities are relatively limited, mainly to government roles like education, with few employment opportunities for traders and merchants. Some communities support xi the development of oil palm plantations expanding from the west, with the prospect of development and employment opportunities which have been lacking since the demise of local timber concessions. Others are less receptive, fearing the negative impacts of logging operations in the past may rematerialize under oil palm plantations. This is occasionally a source of conflict between villages, over land-use and access.
In the Mentebah sentinel site in the south, villages are mostly Malayu, especially in the flatter areas near the Kapuas river; Dayak villages are relatively remote, being on the foothills. Gold mining is an economically important activity here; communities are more dependent on it than in Batang Lupar, and much less dependent on forest resources. Dryland farming, predominately for subsistence foods, coupled with rubber production, are the main land uses. Paddy and vegetable gardens are limited, as communities can buy agricultural produce from neighboring villages using cash earned from rubber, mining and increasingly employment with oil palm plantations, as there are generally good road networks. However, some villages have difficult access, particularly during the rainy season. The main issue facing communities and the government is the level of mining activity in the area, which has many serious environmental and social impacts, as well as impacts on agriculture practices. Mining causes sedimentation and water pollution, which are detrimental to the ecosystem and human health. However, it does provide a significant source of income, for which there are few alternatives. Timber production is now minimal, having been more prolific in the past. The Mentebah site is, in some respects, more economically developed, with income from employment and rubber outweighing subsistence farming and non-timber forest product processing.

Conclusion
The regency of Kapuas Hulu, where the 'Borneo sentinel sites' are located, is of utmost importance as a water reservoir to the western part of the province, including the capital city of Pontianak. With forest cover of almost 73% in 2019, it represents the last forest frontier of West Kalimantan province. The area forms part of the Heart of Borneo initiative and is a reservoir of unique and intact flora and fauna similar to Sarawak landscapes where forest has been significantly damaged and habitat degradation has been more intense.
In future, to minimize deforestation and degradation of the area without jeopardizing development and public well-being, continuing to enhance the institutional capacity of partner research organizations is recommended, working closely with local government and communities through multi-stakeholder platforms, with the aim of more productive but equitable natural resource management.
For efficient monitoring, it is crucial to revisit former biophysical and social sampling locations to assess the ecosystems' resilience potential and smallholder communities' capacity to adapt and cope with possible future shocks. Critical research pathways, built on existing data, should encompass more research on: commodity value chains, markets and economics; ecosystem functions, particularly related to the relationship between trees, soil and water; and participatory modelling to help decision-making processes, including enabling conditions for the development of 'payments for environmental services' (PES).

Introduction
There is a mounting appreciation of the complex inter-linkages between biological diversity and ecosystem services, and of the necessity to better consider these within land-use and spatial planning decision-making for sustainable development. The need to balance oftencompeting economic priorities while mitigating environmental degradation represents a leading policy challenge. Growing concerns over issues such as food security, malnutrition and increasing international demand for agricultural commodities, exacerbate this. Sustainable development is, likewise, partially impeded by a lack of sound understanding about the relationships between these drivers, as well as the limitations of existing institutional arrangements managing such complex systems.
This requires long-term socio-ecological research into these relationships, as well as research that specifically explores how these complexities can be fairly and legibly represented within resource governance systems. Long-term socio-ecological research is lacking when it comes to understanding impacts of forest and tree cover changes on ecosystem services, diversity and the well-being of indigenous communities. To address this gap, an evidence-based research initiative drew on several tropical long-term research sites, called 'sentinel landscapes', under the Forests, Trees and Agroforestry (FTA) research program of the CGIAR. This approach responded to a key recommendation from the 2009 Stripe Review of Social Sciences, commissioned by the CGIAR Science Council.
As defined by the FTA (2011b) research program, "a sentinel landscape is a geographic area or set of areas bound by a common issue, in which a broad range of biophysical, social, economic and political data are monitored, collected with consistent methods and interpreted over the long term". These long-term data are essential to understand socio-ecological system dynamics, and therefore to address development, resource sustainability and scientific challenges, such as linking biophysical processes to human reactions, and understanding the impacts of those reactions on ecosystems. However, the major justification for sentinel landscapes is the need for a common observation ground where reliable data from biophysical and social sciences can be tracked over time, so that long-term trends can be detected, and society can make mitigation, adaptation and best-bet choices.
At the global scale, the data generated will fit into a global analysis of networks, including other sentinel landscapes, to help understand issues and processes that could be relevant to managing tropical landscapes worldwide.
The following analysis and report are based on the data collected so far (2010-2017) by CIFOR and partners, in one of the selected landscapes of the FTA-Sentinel Landscape initiative: Borneo (focusing on Kapuas Hulu, Sintang, Melawi and Ketapang regencies, West Kalimantan). The decision to conduct this stock take was taken at workshop held under the auspices of the FTA Independent Steering Committee, in June 2018. The main objective is to introduce here the large amount of data and initiatives already collated and undertaken in the Borneo sentinel landscape region, to identify eventual gaps in what should be the most suitable data for socio-ecological system monitoring, and to make recommendations for future research priorities. | 2 Yves Laumonier, Trifosa Simamora, Abraham Manurung, Sari Narulita, Uji Pribadi, Alfa Simorangkir, Selly Kharisma and Bayuni Shantiko

History of the 'Borneo -Sumatra Sentinel Landscape'
The premise behind the selection of sites was 'forest transition' theory 1 (Mather 1992;Meyfroidt and Lambin 2011) when several studies carried out in many countries highlighted signs of an increase in forest cover over recent decades (Perz 2007;Meyfroidt and Lambin 2009). This phenomenon is explained by the intensification of agricultural systems, the increase in urbanization and new reforestation policies, allowing spontaneous regeneration of the forest (Rigg et al. 2001). In tropical regions, with international pressure to take into account REDD programs (for reducing emissions due to deforestation and forest degradation) and, more recently, advocating the needs for 'ecosystem restoration', it becomes essential to deepen our knowledge on this still-debated forest transition theory (Perz 2007).
To exemplify the forest transition in Indonesian landscapes, and after negotiation between FTA centers and partners in Nairobi (2011), final agreement was made on a 'Borneo -Sumatra Sentinel Landscape' (FTA 2011a). Borneo would be assigned to CIFOR (who have a long research presence in Borneo) and Sumatra to ICRAF (who have a long research presence in Sumatra).
1 Forest Transition theory: The concept is based on observation of the historical trajectory of changes in land-use patterns in forested landscapes, the drivers of these human-induced changes. Its representation is a classic U-shaped evolution curve, from high forest cover to decreasing forest cover through degradation, then total transformation to agriculture, and sometimes to expanding tree cover again through tree plantations or agroforestry.
Within these regencies, potential sentinel sites of 10 x 10 km were randomly selected. The sites were excluded when situated in an area almost exclusively dominated by either 1) swamp/peat, 2) timber or oil palm plantations, 3) mining activities, 4) forest without settlements, or 5) urban or periurban situations. The final four selected sentinel sites represent a good illustration of the forest transition theory, from the more forested landscapes of Kapuas Hulu in West Kalimantan and Merangin in Sumatra, to the more fragmented landscapes of Sintang in West Kalimantan, and Sarolangun in Jambi, Sumatra.
Within each sentinel site (10 x10 km), a minimum of 10 villages were to be identified, first through existing base maps; their basic characteristics were later checked in the field and through discussions with local partners. In Borneo, it was sometimes difficult to find 10 villages within a 10 x 10 km square, while in Jambi, Sumatra, the density of villages was higher. In both cases, we did not limit our selection to villages, but considered hamlets (dusun) as well. In Borneo, we tried to ensure the balance between villages according to distance to the road and market access. In Sumatra we tried to ensure balance between traditional local villages and new settlements inhabited mainly by migrants from Java.
The ambition was that the monitoring of these sentinel landscapes and sites would guide future landscape management choices and development at the district, regency and provincial levels.

Background of the West Kalimantan sites
The dynamics between people and landscape are long established in West Kalimantan. The province has been largely deforested since colonial times, the oldest modern vegetation map (Hannibal 1950) showing forest already confined mostly to swamps on the west coast, and the interior hills and mountains. However, development has introduced changes and brought new pressures that have influenced these interactions, and that are having far-reaching impacts for both people and place. These now often convolute conditions, and have current and future implications for the viability of livelihoods (individuals, communities and the private sector) and the sustainability of natural resources use. The regencies and provincial governments have made commitments to conserve their natural resources, but the challenge is to ensure that this can be done while also creating economic development that ensures poverty alleviation and supports the needs of the local population.
Regencies associated with the Kapuas River basin exemplify the forest transition trajectories, and related issues in landscape management trade-off between conservation and development. The five regencies of Figure 8 represent a gradient of forest degradation balanced by traditional swidden and agroforestry systems (smallholder rubber), largely impacted now by the development of monoculture plantations, mainly oil palm, and some timber pulp plantations. These landscapes encompass representative lowland forest types (well-drained mixed dipterocarp forest, Kerangas/kerapah, and peat swamp forest of varying peat depths), the whole series of degradation, secondary regrowth/fallow and traditional agroforestry/ swidden agriculture systems, but also the environmentally-devastating traditional gold mining that is occurring in some areas, and more recent land uses such as booming smallholder oil palm plantations. They also correspond to the Kapuas Hulu basin, and watersheds impacting the Kapuas River's flow.  (Yuliani et al. 2010). The potential impact of oil palm plantation expansion, on both biodiversity and ecosystem services, presents a major challenge to the regencies where forest is still prevailing; the local government is faced with the double challenge of mitigating the impacts of economic development and of maintaining ecosystem services, preserving the environmental and social functions that they provide. The diversity of unique habitats and their importance for the conservation of Borneo's biodiversity, as well as the crucial significance of the upper Kapuas River basin as a hydrological system, led to the establishment of two national parks within the regency: Betung Kerihun National Park, with a variety of lowland, hill and mountain forest ecosystems, and Danau Sentarum National Park, the only remaining extensive wetland area in Borneo, with a unique ecosystem of interconnecting seasonal lakes, peat swamps, and periodicallyinundated freshwater swamp forests (Giessen 2000;Aglionby 2010). Both parks and the Kapuas River play essential roles in biodiversity conservation, environmental services and local community livelihoods, not only for inhabitants of Kapuas Hulu, but also for inhabitants who live downstream, including in Pontianak, the provincial capital. To curb massive illegal logging occurring in the area since the 1980s, reaching a peak from the 1990s until 2003 (Eilenberg 2012), and also to secure its hydrological tower function for the rest of the province, the local government declared the regency a 'conservation district' in 2003. In 2018, it was registered as an official UNESCO-MAB (Man and the Biosphere Programme) biosphere zone.
Local government, national park authorities, local communities that practice traditional swidden agriculture, the private sector, national and international NGOs, research institutions and academics are the main actors influencing the landscape configuration and composition. Competing perspectives over land are apparent between regencylevel agencies focused on planning, forestry and agriculture, national park authorities, local communities and the private sector. These stakeholders have different, sometimes conflicting, interests over the limited useable agricultural land within the regency. All parties believed they had rights to the land, either due to legal permits (i.e. concessionaires) or because they had lived there for generations (before formal land rules were established) and stand by their customary rights. While most villagers had a perception of high tenure security, their land rights appeared to be threatened by the government's incomplete recognition of customary institutions, unclear regulations, and most information and key documents being held by the local elite (Clerc 2012;Shantiko et al. 2013). Current land use has been influenced by unclear or ambiguous regulations and institutions related to natural resource management, tenure conflicts, the exclusion of ecosystem services in land-use planning processes, and the agenda of largescale businesses. Such issues have impacts on forest clearance and biodiversity loss, and threaten local communities' livelihoods and cultural identity.
To meet the need for development, land allocated for alternative use to forestry

Kapuas Hulu: A demographic overview
The total population of Kapuas Hulu is 263,207 (BPS 2020), and the regency has a population density of approximately eight inhabitants per km 2 . The highest population density is in the district of Hulu Gurung (32 inhabitants/km 2 ) while the lowest is in Embaloh Hulu (1 inhabitants/km 2 ). The majority of the population (67%) is of productive or working age (15-64 years old).
Kapuas Hulu is ethnically and culturally very diverse. The population can be generally classified into Malay (those speaking Malayic Dayak language) who inhabit the wetlands and floodplains, and Dayak, inhabiting the interior hills (mainly Iban, Tamanbaloh, Kantu, Bukat, Kayan, Tamankapuas, Kalis and Hovongan).
While the majority have access to basic education, 29% hold no form of educational certificate ( Table 1) and only 2% of young people graduate from higher education. A lack of teachers and poor infrastructure are contributing factors to these low levels of formal education.
Poor infrastructure can be seen in both road quality and distribution. Asphalt and concrete roads with good access throughout the year make up just 33% of all road surfaces ( Table 2).
Other road classes, i.e. those made of gravel and dirt, are passable but sometimes difficult to access, particularly during the rainy season. River transportation remains an important means of mobility and is also used for economic activities. The use of river transportation, coupled with the poor road infrastructure, | 7 causes high economic costs and affects the price of goods and services.
Employment is diverse in the district, although 69% of the population is still involved in the agricultural sector. This is followed by social services at 9.79% and the trade sector at 9.15% (Table 3).

Kapuas Hulu: An environmental, land-use and conservation overview
Kapuas Hulu, covering an area of 31,162 km 2 , has unique environmental characteristics. The annual mean temperature is 27.2 o C (mean temperature of the coldest month 22.1 o C; mean temperature of the warmest month 31.9 o C), and mean annual precipitation is 4,231 mm (Worldclim 2018; Figure 10). The whole regency has a very humid climate but long droughts occasionally occur, often during El Niño years. Lusiana (2008) and Hidayat et al. (2017) used a combination of satellite remote sensing and field observations to study the hydrological dynamics of the Kapuas basin.
Forest designated for conservation and watershed protection makes up approximately 57% of the regency, in contrast to the 26% designated across the whole of West Kalimantan. The two national parks -Danau Sentarum National Park and Betung Kerihun National Park -occupy about 30% of the regency. The area designated as production forest -including limited production forest (hutan produksi terbatas, HPT, forests on steeper slopes, with only trees over 50 cm in diameter permitted for cutting) and convertible production forest (hutan produksi konversi, HPK) -makes up about 25% of the regency. The revised spatial plan (rencana tata ruang wilayah kabupaten, RTRWK) for 2011 (officially endorsed in 2013) proposes that land allocated for alternative use (areal penggunaan lain, APL) be increased up to 19% of the regency ( Table 4). This implies that this land is now available for agricultural development, in particular for the establishment of oil palm plantations.
As far as biodiversity assessments are concerned, the Danau Sentarum National Park area has received the most attention.
Since the 1990s, it has been an active research site for ecologists researching flora and vegetation, fish, crocodiles, birds, proboscis monkeys and orangutans, as well as human resource use, including honey, fisheries, turtles, rattan, timber, culture and social and environmental economics (see review by Aglionby 2010). In contrast, the rest   (BPS 2020) of the regency, including the other park Betung Kerihun, has received very little attention outside of initial transboundary expeditions and baseline fauna flora data (Soedjito 1999;ITTO 2013). None of the southern hills and mountains had ever been surveyed until recently ( In terms of agriculture, people generally use traditional methods to manage land, practicing swidden agriculture to grow crops such as paddy, maize, cassava, tubers and vegetables. Agricultural land acquired through clearing forest and secondary forest is systematically left to a fallow period after paddy is harvested. Perennial crops, mainly rubber, provide a ready source of cash income and are often grown to supplement other crops like pepper, fruits and tengkawang (illipe) nuts. The forest is very important for the people of Kapuas Hulu, particularly for food (meat, fruits and vegetables), medicines and traditional crafts (baskets and mats), but also culturally (sacred forest areas and forest products used in rituals). Although economic activities are emerging across the district and in rural areas, subsistence activities, like sourcing food from fallow land, fishing and gathering forest products, still play a central role. Oil palm plantations have expanded across West Kalimantan in recent years. Kapuas Hulu is no exception; oil palm plantations have spread from the regency's west, reaching the north of Danau Sentarum National Park in 2012, the 'biodiversity corridor' between the two national parks, and now spreading rapidly in the south, with peatland near Putussibau being opened up in 2012. The potential impact of this expansion on biodiversity and ecosystem functions presents a major challenge to the regency, who must both mitigate economic development impacts while also maintain ecosystem services, preserving the environmental and social functions they provide.
Development has introduced changes and brought new pressures, which have influenced long-established interactions between the people and their landscape. These are having far-reaching impacts on both people and places. This unique, but now often fragile, situation has implications for the present and future viability of livelihoods (for individuals, communities and the private sector) and the sustainability of natural resources. With the regency designated a 'conservation district', Kapuas Hulu's government has committed to conserving its natural resources, but the challenge is to ensure that this can be done while creating economic development that ensures poverty alleviation and supports the local population's needs.

CGIAR and FTA partner projects
A number of CGIAR interventions have taken place in West Kalimantan, particularly in Kapuas Hulu regency (Figure 12). The projects below are CGIAR and FTA partner projects in Kapuas Hulu and other West Kalimantan regencies over the last decade.

CIFOR -Great Ape Conservation Funds: Saving the remaining orangutan population and their habitat within and surrounding the Danau Sentarum National Park, Indonesia (USFWS), 2009-2013
This project aimed to update information on the conservation status of the orangutan, conduct community/government awareness programs on reduction of forest conversion, and develop a conservation plan, collaborative monitoring system and a participatory land-use plan for the national park area.

CIRAD -Collaborative Land Use Planning and Sustainable Institutional Arrangements (CoLUPSIA), 2010-2014
This EU-funded project was implemented by the French Agricultural Research Centre for International Development (CIRAD) in partnership with CIFOR, Telapak and several local NGOs (HuMA, Riak Bumi, TOMA) and universities (Pattimura, Tanjungpura, Gadjah Mada). It aimed to contribute to reducing environmental degradation and strengthening land tenure and community rights by collaboratively integrating all stakeholders' views in land-use management and development versus conservation. Outputs revolved around stakeholder engagement for land management (including a multistakeholder platform) and assessment of possible payment of ecosystem services for biodiversity and livelihood benefits. The project focused on two Indonesian regencies: Kapuas Hulu and Central Moluccas.

CIFOR -ASEAN-Swiss Partnership on Social Forestry and Climate Change (ASFCC), 2012-2020
The focus of this SDC-funded project was to better understand swidden systems as a social forestry practice and their relevance for REDD+ and livelihoods. The aim was to understand how local knowledge, practices and social networks can be incorporated into the design of REDD+ projects, to ensure that swidden communities can participate meaningfully in and benefit from REDD+.

CIFOR -Participatory Monitoring Reporting and Verifying (PMRV), 2013-2015
The objective of the USAID/NORAD-funded PMRV project was to identify MRV systems producing credible data that are effective, verifiable, participatory and locally relevant, so that the data can be embedded into a national database. This was done by: (i) exploring new possibilities for community-based carbon monitoring; (ii) exploring community-based monitoring of the drivers of deforestation and forest degradation; and (iii) studying participation in reporting, by comparing the health and forestry sectors in Indonesia. In Kapuas Hulu the project worked in the southern part of the regency (Pengkadan village in Pengkadan Hulu district; Nanga Jemah and Sriwangi, Boyan Tanjung district).

CIFOR -Agrarian Changes, 2013-2015
The DFID-funded, CIFOR-coordinated Agrarian Changes project explored the conservation, livelihood and food security implications of land-use and agrarian change processes at the landscape scale. One of the project outputs, an associated book, provides detailed background information on seven multi-functional landscapes in Ethiopia, Cameroon, Indonesia, Nicaragua, Bangladesh, Zambia and Burkina Faso. The focal landscapes were selected as they exhibit various scenarios of changing forest cover, agricultural modification, and integration with local and global commodity markets. A standardized research protocol allowed for future comparative analyses between these sites.

CIFOR -Sustainable Wetlands Adaptation and Mitigation Program (SWAMP), 2012-2021
SWAMP is a collaborative effort by CIFOR and the US Forest Service, with support | 11 from USAID. Tropical wetlands provide a wide range of ecosystem services, such as supporting services (nutrient cycling, soil formation, primary production), provisioning services (food, fiber, and fuel), regulating services (pollution, flood and erosion control, carbon/climate), and cultural services (education and recreational). The aim of SWAMP is to generate knowledge regarding the sustainable management of wetlands, especially peatlands and mangroves in the face of changing global climate and livelihoods of local community. Such knowledge is used to inform governments for public policy-making processes.

CIFOR -Partnerships for Enhanced Engagement in Research (PEER): Integrated watershed management for enhancing local livelihoods and biodiversity conservation in Indonesia, 2015-2018
This USAID-funded project sought to promote effective implementation of integrated watershed management (IWMA), enhancing local livelihoods, biodiversity conservation, and the research capacity of the partners involved by: (i) assessing institutional arrangements for more effective IWMA; (ii) developing approaches for implementing landscape-level biodiversity conservation; and (iii) promoting IWMA for enhancing local livelihoods supported by policy and regulation frameworks at national and local levels, based on good governance principles. Using participatory action research approach, the project focused on the capacity building of NGOs as local project partners. The project complemented current research initiatives, leveraging existing social capital and building upon key recommendations. The activities were implemented at watershed level in Kapuas Hulu (West Kalimantan), South Sulawesi and Sumbawa.

CIFOR -Governing Oil Palm Landscapes for Sustainability (GOLS), 2015-2018
The USAID Indonesia-funded GOLS program supported effective and equitable implementation of private sector commitments to monitoring land-use change and halting deforestation, helping to align public and private policies and actions, and delivering targeted, research-based evidence to key stakeholders and practitioners. Extensive surveys on smallholder oil palm farmers were performed in Kapuas Hulu, Sintang and Sanggau regencies.

CIFOR -Knowfor 2: Food Security Strategy, 2016-2017
The second phase of this DFID-funded project on food security strategy built upon Phase 1's successful outcomes and knowledge-sharing related to forests, food security and nutrition, which contributed to expanded linkages to health, and a consolidation of research into both policy and practice. In this project, CIFOR played a strong role in terms of recent commitments to zero deforestation, and to monitoring and evaluating such systems.

CIFOR -Understanding the Drivers of Food Choice in the Context of Rapid Agrarian Change in Indonesia, 2017-2019
This project, funded by the Drivers of Food Choice competitive grants program, aimed to provide the research community, local communities, policymakers and international donors with evidence to help create a food environment more conducive to healthier food choices in rural Indonesia. The project was carried out across two Indonesian sites undergoing this process of transformation -in West Kalimantan (Kapuas Hulu) and in Papua. Both qualitative and quantitative data were collected to investigate the impacts of agrarian transitions on the diets of mothers and children. Research findings will inform more nutrition-sensitive decisions at national, local, industry and household levels.

ICRAF -Harnessing the Potential of Trees on Farms, 2018-2021
The 'Harnessing the Potential of Trees on Farms' project, funded by the German Federal Ministry for the Environment, Nature Conservation and Nuclear Safety (BMU), under the International Climate Initiative (IKI), aims at better management of resources in the pressing issues of deforestation for agriculture. Agricultural land continues to grow as the world gears up to feed a population expected to rise to at least 9 billion people by 2050. Farms management can be maximized to conserve some level of biodiversity and complement the efforts to save and restore forests with efforts to manage biodiversity and ecosystems services on farmland. The overall goal of the project is to improve the ability of countries to meet Aichi Target 7 of the United Nations Convention on Biodiversity (sustainably managed agricultural areas) by advancing the knowledge of the importance of trees on farms for biodiversity and human wellbeing. The project is carried out in Honduras, Indonesia, Peru, Rwanda and Uganda. In Indonesia, the project is carried out by CIFOR in West Kalimantan.

CIFOR -Collaborating to Operationalise Landscape Approaches for Nature, Development and Sustainability (COLANDS), 2018-2023
This five-year project is funded by the German Federal Ministry for the Environment, Nature Conservation and Nuclear Safety (BMU), under the International Climate Initiative. The COLANDS project focuses on operationalizing landscape approaches in Indonesia, Ghana and Zambia. It aims to use a landscape approach to address community challenges, as well as observe implementation processes and local uptake of such an approach. The project addresses the science-implementation gap and formally operationalizes the landscape approach through wide-ranging partners and stakeholders, who will assess the conceptual framework and develop methods and tools to put the approach into practice. Actors across multiple scales, with a focus on national policy and process in the target countries, will benefit. The project will simultaneously raise awareness of the value of biological diversity, promote options and build capacity for better integrating biodiversity in national sector policies and landuse planning processes. GIZ -FORCLIME, 2010-2020 FORCLIME focused on the relationships between national, provincial and district governments in making land and resource use decisions. A team of international and Indonesian advisors strengthened institutional capacity to manage and monitor forests where decentralization had left uncertainties in mandates between agencies and administrative levels. FORCLIME facilitated establishment of forest management units (FMUs), to bring the management of different forest categories from 'protection forest' to 'production forest' under one roof, is crucial to this reform. The establishment of FMUs involves the division of monitoring and inspection tasks from operational tasks and thus will place management responsibility closer to the field. During its final years, FORCLIME involved participatory mapping and agroforestry in one of the sentinel sites.

GIZ -Sustainable Agricultural Supply Chains in Indonesia (SASCI), 2020present
In collaboration with Ministry of Agriculture (Directorate General of Estate Crops), SASCI's aim is to reduce GHG emissions from deforestation and forest degradation by promoting sustainable supply chains for agricultural commodities (primarily palm oil) to smallholder farmers in Kapuas Hulu. The project promotes sustainable agriculture based on a jurisdictional approach, focusing on participatory land-use planning and mapping, and improved agricultural production practices, including deforestation-free production and dissemination of agro-biodiversity promoting approaches. In the landscape planning process, areas with protection value (

Land degradation surveillance framework (LDSF)
Biophysical surveys were carried out using methods prescribed by the extended land degradation surveillance framework (LDSF -Vågen et al. 2010), consisting of modules for soil and vegetation sampling, species assessment, and landform and land cover classification. As with other sentinel landscape sites, 160 plots were surveyed in 16 clusters distributed across the site. Vegetation and soils were sampled in each plot. Collected soil samples were processed at the site and then shipped to Nairobi to be analyzed in the soil spectral laboratory. For the species assessment, a botanist with local tree and shrub species knowledge was hired, but we also systematically collected botanical samples for proper identification at the Herbarium in Bogor, Indonesia. The main information extracted at household level was: demographic; household composition; education; migration; livelihoods; housing, water and sanitation; assets; income; main livelihood activities; remittances; credit; food security; food consumption and composition; food scarcity; social visibility; social networks; informal safety nets; use of natural resources; welfare; coping mechanisms; and mobility of the household along the poverty ladder.

Gender-disaggregated data
Gender-specific focus group discussions were organized to collect village-level data.
With household-level data, both the male and female heads of households were surveyed separately in 10 out of the 30 houses in each village. For the rest of the households, either the male household head was interviewed, or both the male and female household heads were interviewed together.

On-farm tree inventory
During the second phase of the baseline data collection, it was intended for a complete tree on-farm inventory to be conducted at 20% of surveyed households, through visits to all the household's fields with a member of this household. This tree inventory was to include species, uses, management, growth parameters and location. The enumerator was to check the land-use activities at plot level and make an assessment of field management together with the household member. However, the survey tool was never designed and perhaps due to the lack of budget, the idea was abandoned. In the case of Borneo and Sumatra's sentinel landscape locations, the definition of 'farm' also raises a conceptual challenge; when it comes to areas which use traditional swidden agriculture, what exactly is, or is included in, the definition of a farm -food crop fields, fallow, or rubber trees grown in the household garden? | 17

Landscape-level vegetation, sampling and methods of analysis
The originally prescribed sentinel landscape methodology for vegetation and tree surveys was inadequate. The information below is therefore taken from other research projects in the same area (i.e. COLUPSIA, GOLS).

Sampling
Based on the CIRAD-CIFOR large-scale ecological vegetation map units (Laumonier et al. 2020) we used an equally stratified sampling design (Hirzel and Guisan 2002) for each vegetation class in the sentinel site window (10 x 10 km). Vegetation classes covered were forest, old secondary forest, old fallow, young fallow, jungle rubber and various mixed gardens, and sampling unit size was 20 x 20 m for trees with diameter above 5 cm. The following subsections detail aspects of data that were analyzed.

Species richness, diversity indices and rarefaction curves
Species richness -the number of total species present in the sample area -is a classic biodiversity measurement, but very sample dependent. Besides species richness, various diversity indices were also used: Shannon-Weiner (H) for species richness and evenness or equitability, Fisher's Alpha (α) (independent of sample size) and Berger-Parker (if dominant species/traits were expected to be more important). Diversity indices like Fisher's α and Berger-Parker are particularly useful to compare tree communities. Fisher's α logarithmic series model (Fisher et al. 1943) describes the number of species and the number of individuals within those species. Berger-Parker gives the fraction of total sampled individuals contributed by the most abundant species. A high Berger-Parker index means that the community has been dominated by common species in the area.
Individual rarefaction curves were analyzed. The cumulative number of species (y-axis) was plotted as a function of the cumulative number of individuals (x-axis), pooled in random (Gotelli and Colwell 2011), using 'vegan: Community Ecology Package' (Oksanen et al. 2017) in R version 3.4.4.
The aboveground biomass differences of each successional stage were tested using ANOVA

Figure 13. Secondary forests in Mentebah
Credit: Yves Laumonier/CIFOR pairwise comparison tests. ANOVA tests whether the means of two or several groups are all equal or not, and therefore generalizes t test to three or more groups (Fisher 1950). The test was done using the vegan package (Oksanen et al. 2017) in R version 3.4.4.

Non-metric multidimensional scaling
Among ordination technics, non-metric multidimensional scaling (NMDS) is often the method of choice for graphical representation of community classification in ecology, because of its flexibility and generality. This is due to: (i) its dependence only on a biologically meaningful view of the data; (ii) its distance-preserving properties (Clarke 1993

Aboveground biomass
Aboveground biomass was calculated using Chave's pantropical model (Chave et al. 2014), using diameter at breast height and the total height of trees: Where D (trunk diameter at 130 cm aboveground) is in cm, H (total tree height) is in m, and WD (wood specific gravity) is in g cm -3 . The function returns the AGB in Mg ha -1 (or ton). Total aboveground biomass was calculated for all plots and sites, along with diameter (classed into 5-10 cm, 10-30 cm, 30-50 cm, and ≥ 50 cm) using the Biomass package (Rejou-Mechain et al. 2018) in R version 3.4.4.

Importance Value index and indicator species
The Importance Value index, widely used in Indonesia, combines relative density, relative frequency and relative dominance of a species. It gives an overall picture of the 'ecological importance' of the species in a community. In this case, it appeared to be of interest to assess its value against concepts such as 'indicator species'.
Indicator species are defined here as individuals specific to a particular vegetation type. To assess the indicator species in the field, the indicator value was calculated by obtaining the specificity (A) and fidelity (B) value (Dufrene and Legendre 1997). For each species i in each site group j, we computed the product of Aij, which is the mean abundance of species i in the sites of group j compared to all groups in the study, by Bij which is the relative frequency of occurrence of species i in the sites of group j, as follows: Where INDVAL is the Indicator Value of species i in site cluster j (Dufrene and Legendre 1997). Ecosystems having a broad geographic range will show low sensitivity, because each indicator species will probably only occur within a subset of the geographical range. The maximum of Aij is A=1, which is when species i is only present in cluster j, while the maximum of Bij is B=1, which is when species i is present in all plots of each vegetation class. In our analysis, site group is the vegetation class (i.e. young fallow, old fallow, secondary forest), and cluster means the number of plots in each vegetation

Institutional mapping, multi-stakeholder platforms and stakeholder engagement
Participatory Prospective Analysis (PPA) is defined as "a systematic, participatory and multi-disciplinary approach to explore midto long-term futures and drivers of change" (Bourgeois et al. 2017a). It is based on coelaborative scenario-building that aims | 19 at imagining the future; a specific form of anticipatory practice that makes it possible to build plausible futures from the explicit and implicit knowledge of diverse participants (Ahlqvist and Rhisiart 2015). It is a structured semi-quantitative, expert-based approach, designed to ensure balanced integration of a diversity of perspectives, relying on the assumption that stakeholders from different backgrounds with a shared interest in the same system can interact in a way that will reveal a common vision (Jésus and Bourgeois 2003; Pretty 1995). The 'system' refers to the issue in question, which is defined by a specific question, geographic space, set of actors and time horizon in the future. The system is made up of a set of 'forces of changes' and their interactions. The process enables stakeholders to explore alternative futures, shaped by the interaction of trends and discontinuities across the various forces considered to drive the system as a whole (Bourgeois et. al 2017b).

Field implementation
Implementation of the PPA in Kapuas Hulu was covered by three workshops of one week each, following the steps below, as per Bourgeois et al. (2017a). 1. Identification of participants. Participants in the co-elaborative scenario building process were selected through an iterative process, following interviews of key informants from different sectors and institutions related to spatial planning. The final selection of each participant was made based on their knowledge, the needed diversity of expertise within the group, and the individual capacity for sharing and learning from others. 2. Definition of the 'system'. In the first participant workshop, the 'system' was defined with regards to the core question to be addressed, as well as its geographic boundaries, the time horizon of the anticipatory work, and the identification of relevant stakeholders. 3. Identification and definition of the 'forces of change'. In the same workshop, brainstorming and discussion were used to enable participants to agree on a list of variables they considered could influence the system in the future. Once the final list of 'forces' was identified, participants worked on common definitions to be used in the next sequence of work. 4. Identification and selection of the 'driving forces'. A cross-impact analysis (or structural analysis) was conducted with the participants, in order to identify the direct influences between variables and to bring structure to the system of variables (Godet 1986(Godet , 2010Popper 2008). A binary scale was used to record the existence of a mutual influence between two variables into a matrix of influence/ dependence. Visualization graphs using the levels of influence and dependence of each variable as coordinates were used to classify the variables and identify the key forces of change. This method was complemented by the use of a formula, producing a compound measurement of the strength of each force used to rank the forces. 5. Building scenarios. Creative brainstorming and collective discussion were used to engage participants in making contrasted, mutually exclusive hypotheses regarding the future states of each key variable. Morphological analysis (Godet 2000;Álvarez and Ritchey 2015) was then applied to identify incompatibilities between future states and build plausible, contrasted scenarios, ensuring that no scenarios would entail incompatible combinations. 6. Elaboration of action plans. Three public consultations, involving a diversity of stakeholders such as community groups, government officials, private companies, NGOs and academics, were conducted through direct meetings and group discussions at both sites, at village, district and regency level. The purpose was to disseminate and discuss the scenarios and identify preferences related to each scenario. Feedback was collated during the meetings by the PPA group. The results were used to develop a road map, using system mapping (Gienapp et al. 2009).

Identification of participants
A diverse range of participants representing stakeholders linked to spatial planning was selected to participate in the scenariobuilding process (Table 5). Altogether, 17 participants from Kapuas Hulu committed to engage in a series of three PPA workshops over a three-month period.

Definition of the 'system'
The core question selected in Kapuas Hulu was, 'what could development look like in our regency in the next 20 years?' This broad theme was chosen instead of a more specific topic such as 'land-use planning' as people feared that it would be too sensitive because of the much-debated issue of oil palm development within the regency. The time horizon was set at twenty years in line with the regional medium-term development and spatial plan; a time span long enough to  consider possible and significant changes, but short enough to justify immediate action.

Identification and definition of the 'forces of change'
In Kapuas Hulu, participants identified 50 variables linked to the issue, covering social, economic, political, technological and ecological dimensions.

Identification and selection of the 'driving forces'
The groups spent a substantial amount of time on critical thinking at this stage, particularly in determining the influence of variables on each other. The results of the influence/dependence matrix allowed the selection of the key variables, according to their position in the graphs and their strengths ( Table 6).

Building scenarios
The analysis resulted in several combinations of plausible scenarios. Similar scenarios were combined, the most contrasting scenarios were assessed and the final scenarios were mutually agreed upon. A final scenario was then defined by an explicit title and narrative ( Table 7).  Policies favor, and are made jointly with, the community. The public participates by monitoring and supervising the planning process. Decisions on land use take people's aspirations into account and are acceptable in both customary and national law. Access to education improves and changes people's behavior toward environmentally sound development. An example of this is that people can master practical and environmentally-friendly technology. 2. Throw the coin but hide the hand Most policies do not address the essential development needs of the people. National law is widely accepted, and customary institutions and indigenous wisdom have been excluded; indigenous people have disappeared. Land use is not informed by traditional wisdom, leading to environmental destruction and the marginalization of people. 3. Panning for gold, getting stones Conflicts in society escalate because stakeholders are excluded from the development process. Poverty and inequality also lead to public apathy; people refuse to participate in development projects. Land-use conflicts arise as customary institutions are weakened and indigenous people are divided.

Eating poisonous fruit
Changing regency's policy priorities -coupled with policy maker's opportunistic behavior that favors certain groups' interests -has resulted in slow growth of development within the regency. Indigenous peoples and customary law are recognized but their existence serves to promote the regency's public image. People are excluded from making decisions on land use and development.

Elaboration of an action plan
The PPA group conducted three socialization meetings and workshops at the village, regency and provincial levels to validate the results and obtain feedback. These meetings allowed participants to discuss the future actions required to at least try to avoid undesired scenarios. In Kapuas Hulu, 120 people from 20 villages and 60 district-level representatives agreed that the 'Steps in harmony' scenario would be the most desirable future outcome. They were prompted to consider issues that might bring about change. A road map was developed for both sites, containing guidelines for implementation of desired scenarios, but also outlining preventive and anticipatory actions against undesired scenarios.
New landscape-level institutional arrangements that can appropriately consider the diversity of stakeholders -and their different roles, interests, power and level of influence -are still needed. Potential funding schemes to sustain such institutional arrangements should also be explored; lack of funding from local institutions often hampers implementation after donorfunded projects end.

History of land use, land cover changes
Dropping from 80% coverage in 1973, Kapuas Hulu forest cover remained significant in 2019 (73%, of which 66% was intact and 7% loggedover). The changes in land use and land cover for 2000, 2010 and 2019 are given in Figures 15 to 17, and briefly analyzed below.

Natural forest
In 1973, intact forest in Kapuas Hulu stood at 2.5 million ha (Gaveau et al. 2016). This decreased to 2.2 million ha by 2000. Between 2000-2010, 15% of mixed peat swamp forest, 7% of lowland mixed dipterocarp forest and 7% of freshwater swamp forest area were logged. Over the following decade (2010-2019) industrial logging stopped, although illegal loggingwhich is more difficult to monitor on remote sensing images -continued, especially in the districts of Putussibau Selatan and Kalis. The area of natural forest that had been opened up for swidden agriculture (ladang, food crop fields) was just 2% between 2000-2010 and remained minimal (less than 1%) in the following decade. The conversion of land to oil palm plantations that started around 1998 near Badau, was initially not at the expense of natural forest (affecting just 2% of natural forestland between 2000-2010), however the amount of forest that was converted to oil palm increased after, notably to the detriment of mixed peat (6%) and peat (6%) swamp forests in the south-west of the regency.

Secondary forest (successional stages, fallows)
Between 2000 and 2010, 5% of the secondary forest areas converted to oil palm plantations, mainly on the expense of secondary kerapah (12%) and freshwater swamp forests (4%). Over the same period, 8% of the secondary freshwater swamp forest and 5% of the secondary kerapah forest degraded into low swampy secondary regrowth. In the following decade, a further 6% of secondary forest was converted to plantations, essentially again from secondary kerapah and secondary mixed peat swamp forests, and also from the mosaic of old fallows on mineral soils. In addition, between 2000 and 2010, 5% of the old fallows became young fallow, shrubs, grassland or fernland through the swidden agriculture cycle, while in 2000-2019 just 2% of old fallow was converted into agricultural land.

Agriculture
The conversion of traditional agricultural land (food crop fields and smallholder rubber) into industrial plantation decreased slightly from 5% between 2000-2010 to 3% between 2010-2019. Around 7% of mixed gardens and 6% of food crops fields were converted between 2000-2010; between 2010-2019 another 6% of food crop fields became oil palm plantation.

Socio-economics, livelihoods, well-being and nutrition
This section provides a summary of the main findings of the sentinel landscape socioeconomic survey in the two districts concerned in Kapuas Hulu. Although belonging to the same regency, Batang Lupar and Mentebah districts have different socio-economic and cultural characteristics that may differently influence land management and economic development at each respective site.

Batang Lupar
Batang Lupar is located in the northern part of Kapuas Hulu, at the border of the Malaysian state of Sarawak (Lupar River actually flows to Malaysia Sarawak). The socio-economic survey was conducted in ten randomly selected villages. Table 8 shows the name of the villages and the number of households sampled. The area is culturally, ecologically and biologically significant, covering a biodiversity-rich forested corridor between the two national parks (Betung Kerihun National Park in the north and Danau Sentarum National Park in the south). The region is also socially and economically important; villagers have a longstanding cultural connection with the landscape as a source of livelihoods. Ethnically, local inhabitants are mostly Dayak Iban inland and Malayu in the wetlands.
In general, people in Batang Lupar live in longhouse villages, which consist of a long building containing separate family 'apartments', as well as public spaces for social life. In Dayak Iban language these partitions are called bilik. Each compartment represents one household, which may consist of one or more families. Originally organized  1940-1945. Between 1955-1960 new villages settled in the area and 40 plots were cleared for farming (Figure 20b). Cleared land size was around 3 ha on average, and mostly within 3 km from the village.
Every 10 years after, a peak in land opening is observed. This pattern may be linked to land rotation in the swidden system, where fallow land is left for years then reused. The largest surface of land use in this village was young fallow, while food crops, mixed gardens and jungle rubber garden areas were relatively similar in size. Recently, the interval between peaks in the land-opening pattern has shortened; more plots have been cleared in recent years (2000)(2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014), suggesting greater population pressure on available land area.

Food security
The food security indicators used in the survey were the Household Dietary and Diversity (HDDS), Food Consumption Score (FCS), and Food Security Score (FSS) (Ballard et al. 2013;WFP and FAO 2008;Wiesmann et al. 2009). HDDS and FCS are highly correlated, so FCS results are not presented here.
HDDS is an important food security indicator as it measures the diversity of diet (Kennedy 2010). The consumption of food was collected using a 24-hour dietary recall (Swindale and Bilinksy 2006). Questions were directed towards the household members doing the cooking and focused only on food consumed | 27 at home. HDDS scores fall between 0-10, where a higher score means higher food diversity; good dietary diversity is >6, medium dietary diversity is 4.5-6, and low dietary diversity is <4.5 (Swindale and Bilinsky 2006).

Figure 22
shows the histogram and density plot of HDDS within Batang Lupar's villages, with HDDS scores ranging between 1-9. 114 households (82%) in Batang Lupar had good dietary diversity, and 22 (16%) had medium dietary diversity; few villages -Sawah, Libung, Sungai Luar and Sei Sedik -fell within the low dietary diversity category (below 3%). Overall, categorization shows that most households across this site consume more than 6 types of foods in a day.
The FSS is used to assess household food security status (Ballard et al. 2013;Efsa 2008). The score is based on the level of food security within a household, ranging from 1 to 10. A high | 28 Yves Laumonier, Trifosa Simamora, Abraham Manurung, Sari Narulita, Uji Pribadi, Alfa Simorangkir, Selly Kharisma and Bayuni Shantiko

Figure 21. Rubber tree in jungle rubber garden in Batang Lupar
Credit: Yves Laumonier/CIFOR score means the household is food insecure. Households are considered more food insecure when, in the previous month, they experience food shortages or lacked enough money to buy food. Figure 23 shows that most of Batang Lupar households have low FSS (scores 2 and 3 having the highest frequency), meaning that, generally, food security levels are good. Looking at income levels, there was no clear relationship between income and FSS scores; some low-income households scored 0 in terms of FSS, while some richer households had FSS of 7 and 9 (demonstrating high insecurity). This points to factors other than income being more influential on food security.

Farm typology
All of Batang Lupar's households are connected to farming. Based on the survey, 16 variables were identified to characterize their farms.
Principal Component Analysis (PCA) was used to explore the potential household outliers for the contributing variable; PCA was iterated 3 times until it has no outlier and produced significant contributing variables. 104 households were selected, as well as 7 variables with the highest contribution for the farm typology classification. Variables focused on household: (1) household size; (2) number of adults; (3) household labor capacity; as well as farm characteristics: (4) number of plots; (5) farm size; and economic variables: (6) Intensification Benefit Index (IBI); and (7) Progress out of Poverty (PPI) Index (Righi et al. 2011;Madry et al. 2013;Alvarez et al. 2014;Harris 2019). Food security and nutrition variables did not contribute enough to be included.
Following the PCA, an Agglomerative Hierarchical Clustering (AHC) algorithm was used to cluster the selected households. Figure 24 shows the biplot of the first and second aspect of the PCA, displaying 3 clusters of farm types. The horizontal axis covers 55% of the variation, while the vertical axis covers 28%. Table 9 shows average values of each variable, for farm typology in Batang Lupar.
There were 37 households grouped in Cluster 1. Characteristics of this cluster were small household size and fewer adults, thus less labor capacity compared to other household clusters. This group also had high PPI, meaning that they were more likely to live below the poverty line. Cluster 2, consisting of 45 households, was the largest cluster. This cluster had the largest households, most adults per household, and greatest labor capacity. Average farm size in this cluster was also far bigger than in other clusters. The IBI variable of this cluster also had the highest average (0.72). This means that the ratio between personal daily income (USD/person/day) and economic returns from agriculture (USD/ha/year) is high. Cluster 3 consisted of 22 households. The value for the number of households, household size, and the number of adults in this cluster was close to the average value for all Batang Lupar farms. This cluster had the lowest values for farm characteristics, with an average of 3 plots and farm sizes averaging 2.78 ha. The IBI of this cluster was below average, meaning this cluster benefited less from farm income.

Mentebah
Mentebah is located in the southern part of Kapuas Hulu regency. Table 10 shows the ten randomly selected villages where households were sampled (n=300). This site represents a variety of social and economic differences, when compared to Batang Lupar. Mentebah is divided by a main road, running from Pontianak, Sintang to Putussibau, that lies between the swamps to the village's north and the hills to its south and east. The road is relatively good and accessible throughout the year, although access to villages further away from the main road is more challenging during the rainy season. The main ethnic group in this area is Malay, which is the second largest ethnic population in Kapuas Hulu. They identify as 'Senganan', originating from Kapuas Hulu as well as Sintang. This is significant because Malay from Senganan are considered different from Malay from Sambas and Pontianak; language is classified as 'Malayic Dayak'.
In terms of land use, smallholder rubber plantations (not jungle rubber gardens) dominate this area, and swidden agriculture is still practiced. Traditional gold mining activities are relatively intensive, both in the river and on land. Land-based investment and concessions have been allocated throughout this area in the past, beginning with logging concessions and continuing now with oil palm companies. Although the logging concessions are now inactive, local people continue with logging activities, benefiting from the good road access to Putussibau and Sintang.

Figure 26. Periodically waterlogged "Kerapah" forests in Mentebah
Credit: Yves Laumonier/CIFOR Figure 27 illustrates the trend in numbers of plots opened between 1960 and 2014. The first land opening was in 1960. The trend shows an exponential growth every year thereafter, especially after 1994. Only once did the number of land openings decrease, around 2010. After this, land plot openings increased in number again until 2014. The graph does not show any pattern of swidden agriculture cycles, so it is more likely that this intensification pattern is due to population increase.

Figure 27a
shows the trend of land plots opened, classified according to size. Most of the fields were opened within 3 km of the house, with very few opening further than 3 km from the village, and on average less than 3 ha in size. Looking at land use type, rubber gardens within 3 km of the village made up the largest land use; there were much fewer food crop fields than mixed garden, rubber garden, old fallow and young fallow plots.

Figure 27b
shows the trend of plots opening each year, by village. The earliest plot opening was in Mentebah Kiri in 1960, followed by Bangan Permai in 1965. In general, the number of plots opened after 1994 increased significantly for all villages. Before 1994, an average of two land plots opened in each village, with some villages having little agriculture activity, such as Akung Jaya and Sei Tekuyung. After 1994, the average number of plots opened for each village raised to 5 on average, with the highest number from Akung Jaya (n=78), Emotong (n=67), and Sei Tekuyung (n=86) villages.

Food security
There were 256 households (85%) in Mentebah classified as having a good dietary diversity, 43 (14%) households with medium dietary diversity, and a small number of households with low dietary diversity (below 3%). Overall, the categorization showed that most of the households in this site consumed more than 6 types of foods in a day. Figure 28 shows the Histogram and density plot for HDDS, by village in Mentebah.
Overall, the graph shows that HDDS among households in each village was balanced, with the average score coming in at 6-8. Most households have 6 types of foods a day. However, each village did have some households in the low HDDS category.  Figure 29 shows the FSS distribution, by village. FSS distribution across Mentabah peaks at 1, 3, 6, 8, 9 and 10 on the scale, indicating unbalanced social-economic conditions, with some villages showing a mixture of high and low FSS. Income level was no assurance of household food security; FSS for high-income households ranged between 1 and 9. It can be assumed, therefore, that in Mentebah food security appears to be influenced by factors other than income.

Farm typology
Not all households practiced farming; those households without farms were excluded from the analysis. PCA analysis showed 96 households which were not outliers, and 8 variables with the highest contribution to farm typology classification for Mentebah site. Variables were household variables: (1) household size, (2) number of adults, and Cluster 1 was the smallest in size and consisted of 37 households (Figure 30). The average of household variables in this cluster was similar to the 'total average' value for all farms in Mentebah (Table 11), while IBI and PPI Index results were also higher compared to the other two clusters. Based on the PCA results, IBI and PPI Index results had weak correlation, meaning that even households with high IBI scores were more likely to live below the poverty line.
The center (average value) of Cluster 1 is in the second quadrant, Cluster 2 in the fourth quadrant, and Cluster 3 in the third quadrant. The horizontal axis explains 42% of the variation and the vertical axis explains 27%.
Cluster 2, the biggest, consisted of 168 households. This cluster had the smallest average farm size (1.9 ha). The centroid for Cluster 2 was close to the zero-coordinate, meaning almost every household variable for this cluster was close to the total average of all households' variables. The average farm size in this cluster was smaller than in the other two clusters. The PPI value was similar to that seen for Cluster 1, indicating that having either a farm of any size resulted in a high PPI score for the household.
Cluster 3, with 84 households, was the wealthiest cluster, having the lowest PPI and highest HDA scores. Other than farm-related variables, this cluster had the highest values in each category. Farm size and number of plots were in between, compared to Cluster 1 and 2.
Overall, farm typology in Mentebah was harder to interpret due to the very diverse socio-economic conditions and the fact that many households have mixed-income sources (farm and off-farm).

Comparing based on land use
When comparing Batang Lupar and Mentabah, there were no significant differences in terms of the number of land plots opened. The total number of opened plots was similar in both sites, but the area of land opened for cultivation was 1.7 times bigger in Batang Lupar (1,897 ha) than in Mentebah (1,095 ha). Fallow land area in Batang Lupar (67 ha) was about the same in Mentebah (62 ha).
The land tenure situation is different for the two sites. The percentage of plots owned without There were no significant differences in terms of gender, when looking at plot ownership. Land ownership in the area is interchangeable between men and women, for diverse family reasons. The percentage of plots owned by males in both sites was about the same, while women owned plots 1.7 times more in Batang Lupar (32% of all plots) than in Mentebah (19% of all plots). The percentage of land owned without information on the owner's gender was 2.6 times less in Batang Lupar (5%), compared to Mentebah (13%).
Land for farming was first opened in 1935 in Batang Lupar, 25 years before the first plot was opened in Mentebah (1960). Land openings continued on a positive trend in both sites, however this trend was steeper in Mentebah. Despite the positive trend seen in both locations, there were some differences. The pattern of swidden plot opening in Batang Lupar increased and decreased every 5 to 10 years (see Section 3.1); however, in Mentebah, this trend continuously increased, except in 2010. Between 2005 and 2010, twice as many plots opened in Mentabah as in Batang Lupar.
An increase in the number of plots opened each year was seen in both sites, but while Batang Lupar continued to display more of a tradition of swidden agriculture, Mentebah was more associated with rapid development.
This rapid increase in plot openings in Mentebah could change the landscape and cause other socio-economic dynamics within this landscape.
Distribution of plots opened, by distance and plot size, showed that plots opened under 3 km and under 3 ha dominate both sites. In terms of distance, this shows that either people at both sites are getting older, or access to open plots further than 3 km remains difficult. In term of plot size, most households were cultivating small areas (less than 3 ha) of land.

Comparing based on food security and nutrition
Dietary diversity was almost identical in the two sites. Over 80% had high dietary diversity and less than 3% had low diversity in each site. No household at either site ate just 1 type of food in a day. In Batang Lupar, a minimum of 2 and a maximum of 9 types of food were eaten, while in Mentebah the range was between 3 and 10. Overall, HDDS scores were relatively similar, but varied more in Mentebah than in Batang Lupar.
FSS scores showed interesting results for both sites. Batang Lupar households were relatively secure, based on their FSS (with scores mostly in the 3-4 bracket) while Mentebah displayed high inequality; this site had the most high and low FSS scores.
The lowest FSS scores in Batang Lupar were seen in Sei Sidik village (where 5 households scored 0), while the highest score was seen In Mentebah, the distribution of FSS score was spread evenly from the lowest score of 1 to the highest score of 10. 27 households had an FSS score of 10; Sei Tengkuyung (11 households) and Menarin (13 households) had the highest proportion of high scores. The villages of Akung Jaya (7 households), Bangan Permai (5 households), and Emotong (7 households) had the lowest FSS scores. The village of Padang Jaya had 4 households experiencing critical food insecurity.
We can conclude that there was no significant relationship between FSS and income level; every income level displayed a variety of food security scores in both locations.

Site comparison based on farm typology
Farm typology showed that variables contributing as discriminant variables were the same in both sites. The only different was that HDA was included in Mentebah. Figure 10 and 14 show that the relationship between each variable was about the same in each site.
The number of clusters at each site was the same (3 clusters), but the centroid location of the cluster differed. Figure 10 and 14 show that the Cluster 2 centroid in Batang Lupar corresponded with Cluster 1 centroid in Mentebah (in the second quadrant). These clusters were characterized by farm size, number of plots, and IBI. Batang Lupar had just one cluster with low IBI scores (Cluster 3) while Mentebah had two clusters with low IBI scores (Cluster 2 and 3). This shows that even though Cluster 3 is significantly different to Cluster 2 in terms of IBI scores, PPI scores for both clusters were the same in Batang Lupar. The situation was different in Mentebah; two clusters (2 and 3) had low IBI scores, but there were significant differences in PPI scores. NMDS results for Batang Lupar and Mentebah ( Figure 31) showed a clear distinction in species composition, with low stress value = 0.182 (meaning that NMDS analysis performs well in displaying the position of communities in reduced dimensions). The ANOSIM result was R=0.24, P=<0.001. The ANOSIM R result of 0.24 means there were differences, with some overlaps. Both locations shared a few similar species (an R value closer to 1 means groups are highly different).

NMDS in Batang Lupar
It is useful to analyze the species composition of forest plots at different succession stages in Batang Lupar (Figure 32). Undertaking NMDS in Batang Lupar using ecological distance showed that the positions of secondary forest, young fallow and old fallow plots were distinct, confirming their difference in terms of species composition. The ANOSIM value R=0.4 means that there was difference, with some overlap in terms of species composition data (old fallow overlapping with young fallow). This overlapping of species composition reflects the successional process of appearance/disappearance of cohorts of species.

NMDS in Mentebah
The successional stages of forest in Mentebah were classified into young fallow, old fallow and old secondary forest. Figure 33 shows distinct clustering. Species composition among young fallow is much more similar to old fallow than to old secondary forest. ANOSIM value R=0.34 means that there were differences, with some overlap in terms of species composition data (old fallow overlapping with young fallow).
These NMDS results confirm the presence of different forest types that can be further described in terms of structure and species composition. The NMDS of different forest types and successional stages showed that old and young fallow in Mentebah and Batang Lupar were not very different, while old secondary forests in Mentebah and Batang Lupar were significantly different. The floristics of each succession stage will be described further.

Rarefaction curves for tree species
The rarefaction curve (Figure 35) leveled faster for Batang Lupar young fallow (n = 82). Thirty plots (1.2 ha in total) were surveyed for secondary forest, young and old and fallow. The richest plots in terms of species were Mentebah's old secondary forest, while Mentebah's young fallow scored lowest on this aspect.

Tree species richness and diversity
A total of 804 individual trees, belonging to 513 species (Ø ≥ 5 cm), among 101 genera and 51 families were enumerated across 30 plots in Batang Lupar and Mentebah. A comparison of the species richness of the successional stages of both sites is given in Figure 34.  The Shannon-Weiner index (H') revealed that the highest diversity of tree species was found in secondary forest (3.75), with the lowest diversity seen in old fallow (3.25).

Mentebah -tree species richness and diversity
In Mentebah, the Dipterocarpaceae family had the most species (18) (2); 20 families had just a single species in the study area.
In young fallow plots, 97 individuals belonging to 25 species among 22 genera and 20 families were recorded (Table 13). 135 individuals belonging to 39 species among 30 genera and 23 families were recorded in old fallow plots, while secondary forest plots had 192 individuals belonging to 85 species among 55 genera and 32 families.
The Shannon-Weiner index (H') showed that the highest diversity of tree species was found in secondary forest (4.16), with the lowest diversity seen in young fallow (2.58). The Berger-Parker value was highest in young fallow explaining the lower diversity of species.

Importance Value index Batang Lupar -Importance Value index
The species with the highest Importance Value index in Batang Lupar young fallow (Table 14) was Ixonanthes petiolaris. This species is found in primary forest, but more frequently in secondary regrowth on hillsides and ridges. In old fallow, Adinandra dumosa was the most striking. This species is known to adapt best to open conditions, poorly aerated soil and restricted supply of nutrients, and often occurs in secondary lowland to montane vegetation. Old secondary forest was best represented by Shorea macrophylla, one of the fastest growing Shorea species (Surianegara et al. 1994).

Mentebah -Importance Value index
The species with the highest Importance Value index (Table 15) in Mentebah's young fallow was Cratoxylum glaucum; this species usually grows in acid peat soils and in Kerangas forest on leached sandy soils. In old fallow, as it was for Batang Lupar, Ixonanthes petiolaris was most important. Old secondary forest was dominated by Shorea beccariana, commonly seen growing in deeply leached soils in lowlands and on shale or sandstone ridges (Surianegara et al. 1994).
A search for indicator species -species whose presence, absence or abundance reflect a specific environmental conditionshowed significant possible indicator tree species for Mentebah and Batang Lupar (Figure 36). In Batang Lupar, Crypteronia cumingii (and the other tree species for which A = 1) was the best indicator, because it was limited to and can only be found in Batang   Figure 37 demonstrates that a larger number of species with higher specificity -that could therefore be considered good indicators eventually -were found in old secondary forest than in young fallow and old fallow (in both Batang Lupar and Mentebah). This indicates that 'older' succession stages have more diagnostic value.

Batang Lupar
Based on the sampling intensity in this study, the five best indicator species in Batang Lupar were: Crypteronia cumingii, Flacourtia sp, Syzygium sp, Adinandra dumosa and Calophyllum tacamahaca.
In old fallow, the IVI index order of each species was 2 nd , 11 th , 3 rd , 1 st and 10 th of 32 species. Calophyllum tacamahaca was also present in secondary forest, with an IVI index order of 21 st of 56 species. It is noteworthy that the best indicator species in Batang Lupar was not Adinandra dumosa but Crypteronia cumingii, even this latter species dominated the entire site of Batang Lupar (IVI = 84.1). This is because IVI only examines the sum of species dominance, abundance and frequency; indicator species analysis, however, looks into the relationship between occurrence and abundance of species, to determine specific species that can be used to represent habitat types. An indicator species may have a high IVI value, but that high IVI value does not mean that a species is necessarily a good indicator of habitat.

Mentebah
Based on the sampling intensity, in Mentebah the 4 best indicator species were present not only in secondary forest, but also in old and young fallow. In secondary forest, these were Artocarpus nitidus, Elaeocarpus floribundus, Santiria rubiginosa and Barringtonia lanceolata; out of 85 species in the IVI index, these species ranked 66 th , 41 st , 8 th and 19 th . In old fallow, Artocarpus nitidus was the only indicator species, with an IVI order of 25 th out of 39 species. In young fallow, Artocarpus nitidus was also the only indicator species, with an IVI order of 15 th out of 25 species.
Indicator species results for Mentebah showed that IVI value alone is insufficient to decide about the most representative species of a site.

Structure
Tree diameter distribution and disturbance intensity As expected, young fallow at both sites showed the highest density for the small diameter class of trees (Figure 38)  secondary forest and young fallow, while in Mentebah, differences are striking. The number of trees recorded in each successional stage in Batang Lupar was also higher, compared to Mentebah.
The level of disturbance -mainly because of timber extraction for housing and the local market -is evidenced by the pattern of diameter class distribution (Figure 40). The shape of the diameter distribution histogram in Mentebah indicates that disturbance is higher there; people use more wood from the old fallows as the (previously logged) forest is difficult to access nowadays, with former logging roads gradually disappearing.

Aboveground biomass
Average biomass in young and old fallows in Batang Lupar was higher than those in Mentebah (Table 17). Trees with diameters above 50 cm were frequently found in young fallows in Batang Lupar, often because they were considered too hard to cut (Koompassia) or because they had food value (Artocarpus).
The total amount of biomass was similar in both locations (Figure 41). As expected, biomass was higher in old secondary forest, but young and old fallow in Batang Lupar are much denser and have higher biomass value, compared to Mentebah.
ANOVA pairwise comparison tests for young fallow, old fallow and secondary forest biomass (Figure 43), showed that in Batang Lupar, young and old fallow were not significantly different, while in Mentebah they were. Aboveground biomass was quite high in secondary forest sites, compared to old and young fallow, thus indicating that old secondary forest sites had carbon sequestration value.

Key differences in the socioecological trajectories of Batang Lupar and Mentebah sites
The differences in development trajectories are linked to different socio-ecological and historical contexts. Batang Lupar sentinel site is located in the vicinity of two national parks, where land is more restricted, as communities face unclear boundaries with Betung Kerihun National Park to the north and, to a lesser extent, Danau Sentarum National Park to the south. All villages are mostly inhabited by ethnic Dayak (Kantu, Iban and Maloh) who demonstrate stronger customary behavior than Mentebah communities in the south. The communities are more dependent on forest resources for both timber and nontimber forest products. Rubber production and other agroforestry commodities are also important, like fishing; while local employment opportunities are relatively limited, mainly to government roles like education, with few employment opportunities for traders and merchants. Some communities support the development of oil palm plantations expanding from the west, with the prospect of development and employment opportunities which have been lacking since the demise of local timber concessions. Others are less receptive, fearing the negative impacts of logging operations in the past may rematerialize under oil palm plantations. This is occasionally a source of conflict between villages, over land-use and access.    In the Mentebah sentinel site in the south, villages are mostly Malayu, especially in the flatter areas near the Kapuas river; Dayak villages are relatively remote, being on the foothills. Gold mining is an economically important activity here; communities are more dependent on it than in Batang Lupar, and much less dependent on forest resources. Dryland farming, predominately for subsistence foods, coupled with rubber production, are the main land uses. Paddy and vegetable gardens are limited, as communities can buy agricultural produce from neighboring villages using cash earned from rubber, mining and, increasingly, employment by oil palm plantations, as there are generally good road networks. Some villages do have difficult access, however, particularly during the rainy season. The main issue facing communities and the government is the level of mining activity in the area, which has serious environmental and social impacts, as well as impacts on agriculture practices. Mining causes sedimentation and water pollution, which are detrimental to the ecosystem and human health. However, it does provide a significant source of income, for which there are few alternatives. Timber production is now minimal, having been more prolific in the past. Mentabah is, in some respects, more economically developed, with income from employment and rubber outweighing subsistence farming and non-timber forest product processing.

Figure 45. Oil palm concession with remnant peat swamp forest in Mentebah
Credit: Yves Laumonier/CIFOR Although lessons learnt from other research and development projects could have been reviewed in more detail before this sentinel landscape was established, this compilation and preliminary analysis of CIFOR data for one of the sentinel landscape observatory has, at least, set the baseline for future monitoring of forests, trees, smallholder farms, industrial agriculture and settlement dynamics in this region of Borneo. The regency of Kapuas Hulu, where the Borneo sentinel sites are located, is of utmost importance as a water reservoir to the western part of the province, including the capital city of Pontianak. With forest cover of almost 80% in 2019, the regency represents the last forest frontier of West Kalimantan province. It is part of the Heart of Borneo initiative and a reservoir of unique and intact flora and fauna similar to that of Sarawak landscapes where forest has been significantly damaged and habitat degradation has been more intense. The very diverse forest types and socioecological conditions in Kapuas Hulu highlight weaknesses in the original sampling, and a necessity to increase these to a minimum of 4 to 5 windows of 100 km 2 , for this area of approximately 30,000 km 2 .
In the future, to minimize deforestation and degradation in the area without jeopardizing development and public well-being, continuing to enhance the institutional capacity of partner research organizations is recommended. Working closely with local government and communities through multi-stakeholder platforms will continue to be crucial, but for efficient monitoring, it is central to revisit former biophysical and social sampling locations to assess the ecosystems' resilience potential and smallholder communities' capacity to adapt and cope with possible future shocks. Critical research pathways, built on existing data, should encompass more research on: commodity value chains, markets and economics; ecosystem functions, mainly related to the relationship between trees, soil and water; and participatory modelling to help decisionmaking processes, including enabling conditions for the development of 'payments for environmental services' (PES).

Future potential
• Improve integration of existing data from various projects in the area; this should lead to better long-term monitoring of landscape dynamics, ecosystem services, animal and plant population studies; • Collect more data on forest and nutrition, agroforestry, trees, and traditional ecological knowledge; this will be useful in making a case for traditional agriculture as opposed to oil palm monoculture; • Train and build capacity to increase local use of data; • Increase awareness and advocacy work around the FTA's achievements within the landscape; • Initiate new sentinel landscapes in Eastern Indonesia or cover different agroecological zones in Indonesia; baseline data exists for the Moluccas; • Develop a LTSER (Long Term Socio-Ecological Research) regional network with partners in Southeast Asia, equivalent to the European LTSER.