Indonesian peatland concession system

A concession divided land without dividing its water

Tropical peat accumulates because saturated conditions slow decomposition. Plantation canals lower the water table so equipment and crop roots can operate; oxygen then enters the peat, decomposition and subsidence increase, and dry peat becomes more combustible. Water moves across title and permit lines. A canal in one parcel can change fire and flood risk elsewhere in the same dome.

Indonesia's official Peat Hydrological Unit, or PHU, recognizes that physical fact. The World Bank's institutional review describes river-to-river units that cross administrative boundaries and says protection of a dome requires limiting drainage beyond the protected plot. Yet ministries, districts, forest managers, estate-crop licenses, villages, protected areas, and corporate concessions use different operational boundaries and maps.¹

That mismatch was built into commodity coordination. A concession could assemble finance, roads, canals, seedlings, workers, mills, and buyers around planted hectares and output. It did not automatically create an authority responsible for water across a PHU. Nor did it settle customary or village tenure beneath a state permit. The same World Bank review found that detailed concession boundaries were not readily shared in usable form, obstructing community and civil-society efforts to test overlaps or pursue social-forestry rights.²

A hotspot identifies fire, not the person who caused it

Fire enters Indonesian landscapes through land preparation, accident, conflict, speculation, and other human action; drought determines how readily it escapes. Drainage makes varied ignitions more consequential because peat can smolder underground and persist after surface flames appear extinguished. Attribution therefore requires more than placing a satellite hotspot inside a concession.

A peer-reviewed satellite and atmospheric-model study of the 2006 season estimated that fires inside oil-palm, timber, and logging concessions produced 41 percent of total fire emissions in Sumatra and 27 percent in Kalimantan. The figures cover emissions located within mapped concession boundaries; they do not prove who ignited each fire, assign parent-company control, or imply that every fire outside a concession was caused by a small farmer.³

Stakeholders also diagnose the system differently. Carmenta and colleagues used Q-method interviews across local, national, corporate, civil-society, and international actors and found agreement about health harm and support for fire-free alternatives, but sharp disagreement over effective interventions. Blame affects whether resources go to arresting individual burners, regulating concession drainage, securing tenure, financing non-fire cultivation, enforcing corporate duties, or rewetting whole landscapes.⁴

This is why modular responsibility fails. A permit holder can point to an unknown burner; a parent firm to a subsidiary; a buyer to a supplier standard; a district to a national license; and an agency to another agency's map. None of those statements answers who maintained the dry conditions, who controlled land access, who benefited from conversion, or who could have prevented spread.

The 2015 crisis produced several different ledgers

Indonesia's official land data, as compiled by the World Bank, estimated that 2.611 million hectares burned from June through October 2015. About one-third of the burned area was peat. The World Bank estimated at least US$16.1 billion (IDR 221 trillion) in Indonesian damage and loss, equal to 1.9 percent of 2015 GDP. That estimate covered 94 percent of the mapped burned area and included agriculture, forestry, trade, tourism, transportation, emergency response, suppression, environmental loss, short-term health effects, and school closure. It excluded long-term health effects, many ecosystem services, and regional and global losses.⁵

The same report counted 24,773 affected schools and 4,692,537 students. Those are administrative and economic-impact measures, not diagnoses of illness. Likewise, the US$16.1 billion is a modeled disaster-loss estimate over a defined period, not company liability or the market value of all carbon and biodiversity.⁶

Atmospheric studies answer a different question. Huijnen and colleagues combined MODIS fire-radiative-power observations, MOPITT carbon-monoxide observations, atmospheric modeling, and smoke measurements near Palangka Raya. They estimated 884 million tonnes of CO2 for the 2015 regional fire year, 97 percent originating in Indonesia, and up to 1.2 billion tonnes CO2-equivalent after including other gases. Their September–October estimate had explicit uncertainty; peat burn depth and emission factors could not be observed everywhere.⁷

Other inventories produce different totals. The World Bank cited a GFED estimate of roughly 1.75 billion tonnes CO2-equivalent for Indonesian fires in 2015. Differences in domain, satellite inputs, burned area, gases, peat depth, and conversion to CO2-equivalent mean those numbers should be presented as alternative estimates, not averaged into one “true” emissions figure.⁷

Koplitz and colleagues modeled population-weighted smoke exposure and estimated 100,300 excess premature deaths across Indonesia, Malaysia, and Singapore. That is not an observed death certificate count. It applies an atmospheric transport model, fire inventories, population data, and epidemiological concentration-response relationships, while assuming broadly similar smoke transport pathways in 2006 and 2015. It is evidence of potentially immense mortality burden with model uncertainty, not a precise official toll.⁸

The asymmetry is organizational. Commodity income enters company, worker, smallholder, and state accounts. Smoke enters bodies far from the originating parcel; carbon enters the atmosphere; school and work disruption enter household time. Neither a concession balance sheet nor a hotspot map contains all of those effects.

Communities were rights-holders, workers, farmers, and responders

“Community” does not name one relationship to peat. Residents may hold customary claims, formal titles, or no recognized document; grow rice, coconut, rubber, sago, or oil palm; fish and collect forest products; work for a plantation; or combine several livelihoods. A fire ban, canal block, concession, and water-table rule therefore distribute costs differently within and among villages.

Human Rights Watch interviewed more than 90 residents of Seruat Dua, Mengkaleng Jambu, and Olak-Olak in West Kalimantan about PT Sintang Raya's oil-palm development. Residents described inadequate consultation and compensation, farmland and crop losses, food insecurity, intimidation, and an unenforced court order to return some land. The company, its identified parent companies, and officials did not respond to HRW's preliminary findings. The account preserves resident testimony and legal history, but it is one three-village rights investigation, not a national sample or a final adjudication of every disputed fact.⁹

The case also complicates a simple jobs-versus-environment story. Some women reported taking difficult, low-paid plantation work after losing farmland. World Bank estimates counted large 2015 losses to both estate companies and smallholders. Employment and market access were real benefits for some people; they did not erase altered water, insecure land, or smoke exposure.⁹⁵

Local institutions can reduce risk when they hold knowledge and usable authority. A 2023 east-Sumatran village case, based on participatory observation, 35 interviews, and documents, described community canal blocks, patrols, and firefighting with researchers, donors, government, and a timber company. The authors reported far less burned area after these arrangements, while also observing police and military domination of emergency decisions. It is evidence of one collaborative pathway, not a national causal evaluation.¹⁰

Reform began to follow hydrology

After the 2015 crisis, Government Regulation 57/2016 strengthened peat protection, restoration rules used a groundwater threshold of 40 centimeters below the surface, and PHU mapping made hydrology a formal management unit. Private concession holders were assigned responsibility for financing and reporting restoration inside their concessions, while government agencies retained responsibility outside them. A rule can assign duties without proving compliance; the World Bank found overlapping mandates, inconsistent sectoral targets, map problems, capacity constraints, and incomplete monitoring.¹¹

Presidential Instruction 5/2019 made the halt on new permits in mapped primary forest and peatland permanent. Its exceptions include previously approved projects, existing-permit extensions, national food and energy programs, strategic infrastructure, restoration, defense, and public safety. It therefore constrains new licensing within a periodically revised indicative map; it does not cancel all existing concessions or prohibit every future activity on every peat soil.¹²

President Joko Widodo created the Peatland Restoration Agency, BRG, in January 2016 to coordinate mapping, zoning, infrastructure, education, and restoration of about two million hectares. Its mandate sat alongside ministries and concession holders rather than replacing them. The agency later became BRGM, adding mangroves and extending through 2024.¹³

Restoration used “3R”: rewetting, revegetation, and livelihood revitalization. Canal blocks and wells can raise water; plants can rebuild cover; fire-free crops and enterprises can make wet peat economically usable. Four Central Kalimantan projects studied by Puspitaloka and colleagues nevertheless faced insecure funding, drainage inherited from earlier development, and social-engagement costs that project accounting often omitted. Their estimate that indirect social costs could reach half of total cost was exploratory, not a national budget rule.¹⁴

Community participation cannot be inferred from infrastructure counts. A government-linked 2025 study surveyed 32 West Kalimantan villages and found generally favorable views of restoration, especially canal blocking, but also concerns about agricultural disruption and uneven involvement in planning. Men dominated physical implementation; women participated in other roles but were underrepresented. Because BRGM personnel coauthored the study and it lacked a comprehensive before-and-after outcome evaluation, its positive findings need to be read as a bounded community-perception study.¹⁵

The dedicated restoration agency ended; the obligation did not

BRGM's mandate expired on 31 December 2024. A June 2025 analysis from the Indonesian Parliament's expertise service said duties had been distributed among the Forestry, Environment, and Marine Affairs and Fisheries ministries. It credited BRGM with 1.6 million hectares of peat restoration and extensive community work, while warning that dispersed authority risked losing technical capacity, funding continuity, and cross-sector coordination. Those achievement figures are institutional reports repeated in a parliamentary policy brief, not independently verified ecological outcomes on every hectare.¹⁶

In November 2025 the Forestry Ministry said zoning, early warning, rewetting, groundwater monitoring, and local economic programs were continuing. This is an official commitment and activity claim after BRGM, not proof that the new multi-ministry arrangement reproduced its field capacity.¹⁷

Fire also continued. The Ministry's SiPongi fire-monitoring system reported that cumulative burned area in January–April 2026 exceeded the same period in several comparison years. It calculates indicative area from Landsat imagery overlaid with hotspots, ground checks, and Manggala Agni suppression reports. The comparison is partial-year and dynamic; it establishes continuing fire and monitoring, not failure of every reform or a concession-specific cause.¹⁸

Fire changes more than carbon

A 16-year study in a Central Kalimantan peat forest measured 17 ecosystem properties and 21 biodiversity variables. Relative to unburned forest, burned sites lost 58–98 percent of habitat structure, tree density, and canopy cover; trees, damselflies, butterflies, fish, water quality, and threatened vertebrates were among the affected measures. Some vegetation indicators improved across 12 years, but recurrent high-intensity fire compromised recovery.¹⁹

That evidence is unusually broad but geographically bounded. It supports claims about habitat and nonhuman life without converting one research landscape into a census of all Indonesian peatlands. It also shows why “hectares restored” is not the same as restored function: water, forest structure, species composition, and livelihood use recover at different rates.

Authority must reach the water table and the rights-holder

Emergency suppression remains necessary, but pumps, aircraft, patrols, arrests, and temporary closures arrive after the landscape is combustible. Prevention requires reconciled public maps; enforceable water management across a PHU; liability that follows operational and corporate control; independent monitoring of concession restoration; and land rights strong enough for communities to invest in long-term care.

Guatemala's ACOFOP community forest concessions offer a bounded contrast: conditional use rights, monitoring, fire control, production, and political representation sit in federated community organizations. Indonesian peat hydrology, commodity chains, state forest law, and village diversity are different. The comparison identifies a missing organizational capacity, not a transferable template: a rights-holder whose authority and livelihood extend across the work of keeping a landscape alive.

Nepal's community forest user groups offer a second analytical comparison in federated local rights, monitoring, and resource governance. Farmer-managed natural regeneration in Niger offers a third in how usable tenure and livelihood incentives can make distributed restoration practical. The comparisons do not claim direct influence or interchangeability: peat hydrology, Indonesian concession law, community heterogeneity, and global commodity chains impose different constraints, and the reviewed sources establish no institutional borrowing among the cases.²⁰

The peatland concession system shows why a clean contractual boundary can be an ecologically false boundary. Investment becomes modular because plots, permits, and output can be assigned. Water, smoke, carbon, subsidence, biodiversity, and customary relations remain connected. Governance becomes credible only when authority follows those connections and the people who bear them can inspect the maps, challenge the permit, and shape restoration before the next dry season.

Concept fingerprint: hydrology exposed the permit system's false modules

Purpose, mission, and institutional legitimacy has supporting weight. Concessions coordinate production, investment, jobs, taxes, and export, while restoration institutions pursue fire prevention, climate, ecological, and livelihood goals. Legitimacy fails when commodity purpose is legible plot by plot but customary rights and shared smoke, water, carbon, and subsidence remain outside the authorized ledger.¹⁵⁹

Authority, legitimacy, and acceptance has supporting weight. Ministries, districts, concession holders, villages, customary claimants, regulators, courts, and restoration bodies possess different rights over land, licenses, water, enforcement, and livelihood. Map opacity and unenforced tenure claims show why a permit's legal authority does not by itself establish acceptance by people whose land and water it changes.²⁹

Delegation, decentralization, and responsibility is defining. Protection rules divide restoration finance and reporting between concession holders and public agencies; BRG and later ministries coordinate across rather than replace them. A hotspot, permit, parent-subsidiary relation, or agency boundary cannot alone assign ignition, drainage, prevention, and repair responsibility.^3111316

Coordination, communication, and common understanding is defining because effective prevention must reconcile PHUs, concessions, canals, communities, fire data, restoration duties, enforcement, and commodity chains. The village case shows one bounded collaboration, while BRGM's closure tests whether multi-ministry coordination can retain staff, funding, field knowledge, and common measures.¹⁰¹⁶¹⁷

Structure, hierarchy, and scale is defining. Plot-level permits and administrative jurisdictions sit inside river-to-river peat hydrological units, atmospheric smoke transport, carbon systems, and global markets. Harm therefore moves across levels more readily than authority. Regional emissions and mortality models make that scale visible without identifying the actor responsible for each fire.¹⁷⁸

Decision making, judgment, and bounded rationality has supporting weight. Officials, firms, villages, and researchers choose among suppression, arrest, tenure reform, water control, livelihood support, restoration, and commodity production under disputed causal accounts. Different stakeholder diagnoses are evidence about structured disagreement, not a vote that decides which intervention works.⁴¹⁵

Measurement, accounting, and control is defining. Concession maps, hotspots, burn scars, water-table thresholds, hectares, emissions, modeled deaths, economic losses, canal blocks, and species indicators each control or evaluate part of the landscape. Their boundaries must remain explicit: co-location is not ignition attribution, modeled mortality is not a death registry, and hectares treated are not restored function.³⁵⁸¹⁹

Cooperation, incentives, and organizational equilibrium has supporting weight. Commodity revenue, jobs, smallholder access, land claims, fire-free costs, enforcement risk, and restoration finance can align or divide firms, governments, and villages. The available case studies show mixed livelihood positions and collaboration, but they do not quantify one national incentive system or one motive for every burner and permit holder.⁹¹⁰¹⁴

Work design, productivity, and automation remains at score zero as an independent lens. Plantation work, smallholder cultivation, patrols, firefighting, canal blocking, monitoring, and restoration all require labor, and women and men did not participate in the surveyed programs in the same way. The record lacks representative wage, task, safety, workload, contracting, and automation evidence across concessions and communities.⁹¹⁵

Knowledge, expertise, and professional autonomy has supporting weight. Hydrologists, remote-sensing teams, atmospheric modelers, ecologists, fire crews, officials, firms, and local residents know different parts of the system. PHU science challenges parcel governance, while bounded community studies show why embedded knowledge and consent cannot be replaced by satellite or infrastructure counts.¹¹⁰¹⁵

Learning, quality, and reliability is defining. The 2015 crisis produced strengthened protection, a permit moratorium, BRG/BRGM, rewetting and livelihood programs, and expanded monitoring. Agency activity and reported hectares do not prove durable water, biodiversity, or fire outcomes; recurrent fire and uneven ecological recovery keep the learning claim open.^1113161819

Strategy, competition, and adaptation has supporting weight. Concessions made commodity investment modular, while post-2015 policy adapted around hydrological units, permanent limits on new permits, and landscape restoration. Exceptions and existing licenses mean the moratorium changed the opportunity set without replacing the underlying production system.¹²¹¹

Innovation, entrepreneurship, and renewal remains at score zero as an independent lens. Canal blocks, wells, wet-peat crops, remote monitoring, and community fire arrangements are documented innovations, but four-project cost work and village cases cannot establish a scalable national outcome. Novel infrastructure is not renewal unless hydrology, livelihood, and ecological function persist.¹⁰¹⁴¹⁵

Governance, stewardship, and accountability is defining across permits, tenure, PHUs, concession duties, fire attribution, restoration, successor ministries, and transboundary harm. Stewardship requires authority and liability to follow water and operational control, while accountability requires maps and monitoring usable by affected communities—not only by agencies and concession holders.²¹¹¹⁶

Culture, informal organization, trust, and voice has supporting weight, correcting the imported zero. Resident testimony, contested causal narratives, village collaboration, military and police dominance, and gendered restoration participation directly concern whose voice is credible and whose relationships govern action. Bounded studies do not establish one culture across Indonesia's peatlands.^49101520

Executive attention, information, and organizational sensing has supporting weight, also correcting the imported zero. National fire-loss ledgers, PHU maps, water thresholds, agency targets, live fire monitoring, and post-BRGM policy reports route selected signals to senior public and corporate authority. Missing beneficial-ownership, permit, tenure, and performance data still prevent one accountable operational picture.^511161820

Organizational ignorance has supporting weight because incompatible maps, withheld concession boundaries, modular responsibility, disputed causation, and model-specific ledgers make it possible to know fire without knowing the responsible organization or the full harm. This is an editorial classification of produced gaps, not a claim that every official, firm, researcher, or resident knows or ignores the same evidence.²³⁴²⁰

The declared links to delegation and governance are conceptual lenses grounded in those mechanisms. The benefit-for-all-life link is an ethical audit of how commodity, employment, health, land, habitat, climate, and future effects are counted. None of the links asserts direct influence or a classification adopted by the cited institutions.²⁰

What the evidence cannot yet close

The public record reviewed here lacks a complete, current, machine-readable reconciliation of concession ownership, beneficial ownership, permits, PHUs, community and Indigenous territories, canals, water levels, fires, sanctions, and restoration performance. Community evidence comes from valuable but bounded case studies; it does not represent Indonesia's many peatland peoples or provide consistent Indigenous-language testimony. Corporate responses are absent in some rights investigations. Official restoration and fire statistics measure activities and mapped area, not always durable hydrological and ecological function. Those gaps prevent a source-complete account.

Paths into deeper study

• Overlay one PHU's canals, water levels, concession and beneficial ownership, customary claims, villages, hotspots, burned-area verification, buyers, sanctions, and restoration duties.
• Reconcile the 2015 burned-area, emissions, exposure, mortality, and economic estimates without treating distinct models as one ledger.
• Follow one community's consent, compensation, livelihood, patrol, and restoration authority before and after a concession or canal block.
• Audit post-BRGM budgets, staff, monitoring stations, concession restoration, and ecological outcomes across several dry seasons.