**2. Methods**

### *2.1. Literature Review*

Starting from several recent reviews of oil palm in Indonesia, we used a 'snowball' method to identify more recent papers citing these sources and followed up on the citation network thus established to document a synthetic view on the various sub-topics.

#### *2.2. Maps*

For a more detailed analysis of the different forest designations within the forest zone and for the time frames of expansion within the forest zone, we compiled data for Jambi, a province in Sumatra around the Equator, and West-Kalimantan. The two provinces were selected to represent the two islands where oil palm is major (Sumatra and Kalimantan) while they had different historical settings on traditional land use of agroforest and its dynamics. Availability of the data also determined the choice of the two provinces. We traced the development of oil palm and agroforestry areas in the 2000–2020 period in each of the Forest-zone categories and sought the trajectories of changes and observed interlinks between oil palm, forest and agroforestry. Where the data were available, we incorporated the other major agroforestry tree-crops; otherwise, we utilized the common class of 'mixed agricultural tree-crops' mapped in most datasets. The following dataset was utilized for the analyses (Table 1).



### *2.3. Public Consultation*

In a webinar in September 2020, attended by 374 participants, the issue of overlap between Indonesian forest estate and oil palm was highlighted and a number of the basic data presented; the questions and discussions formed the start of this manuscript. Specific aspects discussed are listed in Appendix A at the end of this manuscript.

#### **3. Indonesian Context**

### *3.1. Forest Legality*

While efforts to agree on a forestry law during the Colonial period stranded [36], the postindependence (1945) constitution clarified that all the country's resources were for the benefit of the Indonesian people—with subsequent debate on the degree to which the Government of Indonesia has to respect prior claims of customary communities ('Masyarakat Hukum Adat') [37]. The Basic Agrarian Law of 1960 that recognizes colonial period documents as a basis for valid rights to land and the Forestry Law of 1967 was considered to be complementary but had gaps between them [38]. The 1999 revision of the

Forestry Law specified that state ownership claims within the designated 'forest zone' ('Kawasan Hutan') required completion of a gazettement procedure (verifying that there are no valid pre-existing claims), which has been very slow to implement [39], leaving doubts on the legality of government forestry policies that treat the Forest-zone as government-owned.

The total area of forest-zone lands in Indonesia (63% of total land area) is subdivided by primary function, as 'Production Forest' (68.8 M ha; 36.1% of the country), 'Protection Forest' (29.7 M ha; 15.6% of the country) and 'Conservation Areas' (22.1 M ha; 1.6% of the country) [40]. Logging is only allowed in the production forests, with further restrictions in part of the zone, and conversion to monocultural plantations for the pulp and paper industry in other parts. Although logging-based economic interests prevail in this zone, the production forests still have a recognized role in biodiversity conservation, carbon storage and maintenance of watershed functions. Most of the Protection Forest (the Indonesian term could also indicate '(watershed) protecting' forests) are on mineral soils and have been defined through a scoring system emphasizing slope and similar criteria for soil and water protection. For peatland areas, there are separate criteria for determining a Protection Function ('Fungsi Lindung') indication, based on peat depth and the peatland moratorium policy [41]. The total area of Protection Function in Indonesia's peatlands is 12.3 M ha [42].

Building on the specific example of the damar agroforests in Krui (West Lampung, Sumatra) where government ownership claims to forest-like vegetation could not be substantiated [43,44], it was realized that 'agroforests', of various specific histories and intensities of management, were included in the Forest-zone, either with Production or Protection Forest designations, but also claimed by local communities as existing before the Forestry law was formalized [45]. Since then, a government commitment to 'community-based forest management' and 'village forests' has allowed some conflict zones to be resolved, but much remains to be done [46,47]. Meanwhile, calls to redefine the Indonesian forest estate and get the boundaries of its Forestry Regulatory Framework and Agrarian rules right [12,48–51] were largely unheeded. Global interest in and country-level expectations of, new financing mechanisms through Reducing Emissions from Deforestation and (forest) Degradation (REDD+) created a new dynamic where issues of 'forest legality' obtained new relevance [52,53]. Indonesia's challenge to balance the new global environmental agenda on forest conservation, with the existing development deficits is part of a wider global pattern [54].

#### *3.2. Complexity of Frontier Situation*

In Southeast Asia's long engagement with global markets, cocoa, coffee, fast-growing trees, oil palm, rubber and shrimp have all had periods of rapid expansion or 'booms' when the region proved to be a low-cost producer in which land, labor, know-how, and market access could be rapidly mobilized [55]. The 'land grab' literature tries to answer key questions on who seeks to exercise control over land to grow export-oriented crops under boom conditions, how would-be producers navigate regulatory powers and market forces to gain control over land, and how booms differentially affect areas with secure and insecure land control relations.

Since the 19th century, commodities such as rubber [56], coffee [57] and cocoa [58] have boomed in specific parts of Indonesia, creating new 'forest frontier' conditions, involving local people, attracting investors to come in, but also stimulating people from other location to move in, spontaneously and/or in government-sponsored transmigration [59]. The relationships between spontaneous migrants, large-scale plantation companies, local communities and various branches of government have become complex in many locations [60–63]. Due to its location, boom crop production is closely associated with the issue of deforestation and environmental degradation [64].

Indigenous people and local communities (jointly described here as IP/LC) in forest frontier areas have had a long tradition of a 'dual economy' [65] where food security continued to depend on swiddening, while cash-crop agroforestry (rubber, coffee, resin trees and now oil palm) provided a financial basis for their livelihoods [66]. Where the terms of trade were sufficiently favorable, staple food production could be 'outsourced' [67]. Accordingly, most of the oil palm grown by IP/LC, just not to say smallholders, has not been a direct cause of deforestation as they developed in a village environment of orchards, old rubber groves and swidden fallows. Smallholders also tend to use fertilizer sparingly, as they are unable to access subsidized fertilizer supplies that are monopolized by larger plantations [68].

While global policies and markets are often held responsible for accelerated deforestation in the tropics [69], local knowledge is generally assumed to lead to overall positive outcomes in matters of conservation [70,71]. Even in matters of primary forest conversion, local communities are usually presented as the best managers to reconcile conservation and development [69]. Increased attention for forest protection in the context of REDD+ may have increased pressure on converting agroforests outside the forest zone to become oil palm monocultures [72].

The Indonesian forest frontier is home to approximately 26 thousand villages (often defined as 'forest villages') and more than 37 M peoples, with some 18% still struggling to escape from poverty based on the national standard [73]. Most of the forest villages are also not formally registered yet, as nationally only 30% of villages are already registered, mostly in Java. They have no clear and defined boundary and are dominated by state forest land, with formal restrictions to use for crops to generate food or income. In such landscapes, boom crop production, such as oil palm, has taken place. Expansion of agriculture into the forest area, often following onto a logging phase that brought people and road access (however poor) into the forest zone, is considered because of these complexities, that for decades have remained unresolved.

#### *3.3. Total Oil Palm Area*

According to the most recent official figure, the oil palm area in Indonesia was 14.7 M ha in 2019 [74]. Several organizations have also mapped oil palm areas for different objectives and with a range of methods (Table 2). Bahktiar et al. [75] claimed a substantially larger area (16.8 M ha), by including areas cleared for oil palm but without stands recognizable by remote sensing yet. The lowest recent estimate was 11,530,000 ha [33], in a study that only included fully-developed stands. Time-series data of oil palm for Borneo were analyzed for forest loss or deforestation [76] and biodiversity impacts [77]. The breakdowns of oil palm areas varied, with one study differentiating 'industrial' and 'smallholder' oil palm [33], while another study [75] focused on oil palm areas within forest zone lands. Condro et al. [28] mapped major commodities including coffee and cacao, and obtained area estimates close to the data from the Ministry of Agriculture. Across Indonesia, the largest oil palm areas are in Sumatra and the second largest in Kalimantan, covering respectively 6.4 M ha and 4.9 M ha in 2018 [74], while another study [75] mentioned 10.5 M ha and 5.7 M ha, respectively, for these two islands. Oil palm is expanding in W and SE Sulawesi, and in Papua, but current areas remain relatively small.



The distribution of oil palm in Indonesia is uneven and strongly related to climatic conditions. In North Sumatra (Figure 2A), the province with the most even rainfall distribution and virtually

no dry season oil palm can reach up to 50% of the area at district (Kabupaten) level; elsewhere in lowland Sumatra and the southern half of Kalimantan it can reach up to 30% of the area, but in most of Indonesia, oil palm forms less than 10% of the landscape or is virtually absent. For Indonesia as a whole oil palm covers about 8% of the total land area. The fraction of oil palm under smallholder ownership (see below for the distinctions within this category) is approaching 50% for Sumatra, but half of that in most of Kalimantan (Figure 2B). The recent expansion of oil palm, indicated by relatively young stands, is evident outside of the areas that already have a large oil palm fraction (Figure 2C).

**Figure 2.** Maps on 2017 oil palm distribution at the district level in Indonesia, highlighting Jambi and West Kalimantan provinces; (**A**) Area fraction under oil palm; (**B**) Share of oil palm cultivated by smallholders; (**C**) Share of oil palm that does not yield yet as it is too young (usually less than three years old); Sources: Own maps, oil palm data from [79]; district boundary and area data from [80].

Several authors have quantified oil palm development within the forest zone in Indonesia [76,78]. Estimates vary from 2.5 [76] to 3.4 M ha [75]. This implies that 15–20% of Indonesian oil palms grow within the designated forest zone. The dataset of [75] allows some further exploration of the geographical patterns of Indonesia's Forest-zone oil palm (FZ-OP). The relative share in Indonesia's FZ-OP at the province level does not match shares in total oil palm production. With 31.5% of

Indonesian oil palm, the two provinces of Riau + C Kalimantan have 66.8% of its FZ-OP (Riau 20.7% OP, 42.1% of FZ-OP; C. Kalimantan 10.9% of OP and 24.7% of FZ-OP). With 41.2% of Indonesian OP, the rest of Sumatra has 24.0% of its FZ-OP; with 23.7% of Indonesian OP, the rest of Kalimantan has 7.7% of its FZ-OP and with 3.6% of Indonesian OP, Eastern Indonesia + Java have 1.4% of its FZ-OP. Across Indonesia, 2.96% of FZ-OP is designated as conservation area, 4.71% as (watershed) protection forest, 15.15% as production forest with restrictions, 45.4% as regular production forest, and 34.8% as production forest indicated for conversion. The total area of FZ-OP is 2.88 times the amount of production forest indicated as 'conversion forest' (planned, legal deforestation).

For further exploration of that pattern as well as to showcase the development of oil palm gardens as part of agroforestry pathways, we focused on two provinces, one each in Sumatra (Jambi ranked #7 in oil palm production and #5 in FZ-OP) and Kalimantan (W. Kalimantan, ranked #3 in oil palm production and #6 in FZ-OP). Both are outside the historical oil palm core area in N. Sumatra, but they were part of the expansion since the 1990s; the district-level presence of oil palm within these provinces varies from 0–30%. The two provinces have harmonized their Kawasan Hutan areas with the provincial planning (RTRWP) with ministerial decrees that legalized the status of 'Forest-zone lands.' Therefore, issues of agroforest and oil palm practices in these two provinces are not burdened by different interpretations of forest legality between the central government (MoEF) and the provincial governments.

#### **4. Oil Palm in the Forest Zones of West Kalimantan and Jambi**

#### *4.1. Data by Province and Forest Category*

The total area of Forest-zone lands, encompassing all Conservation Areas (National Park and forest reserves), Protection Forest and all Production Forest status (HP, HPT, HPK) in West Kalimantan is 8.1 M ha or 56% of the province area, while for Jambi it is 2 M ha or 43% of the province.

For West Kalimantan, agroforest area occupied 38% of all Production Forest status and 16% of Protection Forest areas [31], while according to [32], agroforest covered <10% (Table 3). Some of the agroforests in [32] are mapped by [31] as disturbed forest, and that was the main differentiating factor of forest and agroforest areas in the two maps. Major commodities of oil palm and rubber are negligible in West Kalimantan's Forest-zone lands, the highest being oil palm in Production Forest areas (4%) mapped by [28].


**Table 3.** Areas of agroforest and major tree crops in West Kalimantan and Jambi in 2018–2019.

<sup>+</sup> processed/reclassified for this paper, # utilized for further analyses, \* percentage lower than 1% of the respective forest-zone land category.

In Jambi, agroforest occupied 19% of the land with Production Forest status, and <10% for Protection Forest and Conservation Areas [31] (Table 3). Rubber and oil palm monocultures were <10% in all Forest-zone land categories in Jambi [28,33]. A breakdown of FZ-OP by elevation and forest category (Figure A1) shows most are below 100 m above sea level.

Taking the total amount of palm oil produced in each of the provinces as point of references, the likely source areas differ between the two provinces, and according to three data sources specified in Table 4, with 6–17% derived from the forest zone. Within the forest zone, the fractions derived from production forest are highest (even after correction for the total area), followed by protection forest and conservation areas.


**Table 4.** Likely source area of palm oil produced in West Kalimantan and Jambi according to three spatial data sources, expressed as %, assuming homogenous productivity per unit land.

*4.2. Data by Time Period—Land Cover Changes Involving Agroforest and Smallholder Oil Palm*

Agroforest in West Kalimantan was mostly as rubber gardens and *tembawang* (Dayak traditional land-use system) having been practiced for generations [81,82]. Inside the Forest-zone lands in West Kalimantan, the majority (76%) of 300,000 ha agroforest areas in 2019 was already agroforest in 2000, and only 11% was forest (Figure 3)

**Figure 3.** Land cover trajectory (2000) of 2019 agroforest in each Forest-zone class in West Kalimantan (analyzed from [31,32]).

The development of independent smallholder oil palm was closely linked to the conversions of the locally managed agroforest areas due to the lucrative oil palm market in the vicinity of the villages and the decreasing popularity of rubber due to its volatile price [82]. That trend is observed in the land cover trajectory of independent smallholder areas in West Kalimantan, i.e., 57% of which was agroforest in 2000 and 54% in 2009 (Figure 4). Only 13% and 6% of independent oil palm was a forest in 2000 and in 2009, respectively.

**Figure 4.** Land cover trajectories (2000 and 2009) of 2019 independent smallholder oil palm in each Forest-zone class in West Kalimantan (based on [31,33]).

For Jambi, agroforest was originally developed by local communities with rubber as the major crop introduced in the beginning of the 1900s during the Dutch period [56,82,83]. It dominated private lands (APL), but its growth reached the areas designated as Forest-zone lands. The 2018 agroforest in Jambi's Forest-zone lands was equally agroforest (44%) and forest (44%) in 2000 (Figure 5), proving the historical presence of this locally managed land use in state's lands.

**Figure 5.** Past land cover types of agroforest in Forest-zone lands in Jambi (analyzed from [31]).

Expectations of high profitability of conversion to oil palm attracted many agroforest rubber farmers in Jambi, also signifying the changes from food sufficiency to cash-cropping [84], in addition to the volatile price mentioned earlier. Independent smallholder oil palm in Jambi's Forest-zone lands was 39% agroforest and 15% plantations in 2009, while in 2000 it was 11% agroforest and 21% plantations (Figure 6). These demonstrate the development of independent smallholder oil palm in Forest zone lands involving agroforest and plantations, mostly rubber. Growth of oil palm also reached degraded areas, shown by 15% shrubs in 2000 and 34% in 2009 (Figure 6), which included peatlands in the lowland areas [85,86].

**Figure 6.** Land cover trajectories (2000 and 2009) of 2019 independent smallholder oil palm in each Forest-zone class in Jambi (based on [31,33]).

Agroforest that has long been managed by local communities has evolved in Forest-zone lands, with the highest proportions and areas in Production Forest status (Figures 4 and 6). With the long history, this land use had faced unresolved forest zone issues since before the 1999 Forestry Law (see Section 3.1) until now, under the *Kawasan Hutan* officiated in 2014 for the two provinces. Since the past decade, oil palm has emerged to become part of the dynamics of this local farming, including in Forest-zone lands. The data show that these dynamics mostly occupied the Production Forest status (Table 3), the class allocated to support various production functions. Looking at the 'predecessor' cash-crop of rubber as part of the ever-evolving mixed-crop agroforest land use [56,81,82], independent oil palm might grow similarly, spatially and or temporally, as part of community-based production functions in land with Production Forest status.
