Gross emissions attributable to soy deforestation. Estimates are based on the total removal of the biomass stored in the native vegetation. The numbers are subsequently transformed in CO₂-eq by multiplying carbon stock per 3.67.

This metric estimates how much deforestation is associated with industrial palm oil production in a given year.  This includes the full plantation cycle, from establishment and planting, to the harvest in a particular year. To estimate the deforestation associated with industrial oil palm production in a specific year, we used a three-year lag between deforestation and planting of oil palm and the first harvest, and aggregated deforestation at the location of oil palm production across 10 years. Deforestation for industrial oil palm is derived from The TreeMap (Gaveau et al., 2022; The TreeMap, 2026). By setting these parameters we are assuming that oil palms will start to produce, and consequently start to contribute to the volume traded in a given year, in the third year after the deforestation event for new plantations. These trees will continue to produce for around 30 years; however, here we are accounting only for the first ten years of production. For example, the total area of land that was classified as forest in 2007, but had been converted to oil palm plantations by 2017, would be considered as areas of deforestation contributing to industrial oil palm deforestation in 2020.

This metric estimates the major sources of GHG emissions from palm oil plantations in Indonesia: subsidence on drained peatlands, fire on peatlands, and land use change, i.e., the clearing of vegetation (forest and non-forest) to establish oil palm plantations. The GHG sequestration from new plantations was also estimated and incorporated into the net GHG emissions reported by Trase.

Gross greenhouse gas emissions from industrial palm oil conversion (tonnes of CO2-eq) were calculated within a given year by estimating all of the CO2-eq potentially emitted due to annual oil palm land use change (The TreeMap, 2026), including both from natural forest and non-forest land covers (as defined in Gaveau et al. (2022)). This metric is calculated by multiplying the area converted each year to palm oil by an estimate of the carbon stock (above and below ground, excluding soil carbon stocks) contained in each hectare of converted land. Above-ground woody biomass density (carbon stocks per hectare) was derived from Harris et al. (2021). Below-ground biomass density was estimated based on Harris et al.’s above-ground biomass density and the ratio of below-ground to above-ground biomass derived from Spawn et al. (2020). To convert biomass to carbon density, we used conversion factors of 0.47 t C t-1 for aboveground and 0.39 t C-1 for belowground, respectively (Diop et al., 2016). Carbon is converted into CO₂-eq emissions by multiplying by a factor of 44/12 (or 3.67), based on the ratio between the molecular weight of C and CO₂ following IPCC (2014).

Net greenhouse gas emissions from industrial oil palm were calculated as the difference between gross emissions from oil palm plantation expansion and carbon sequestration in new industrial oil palm plantations. Carbon sequestration was estimated by multiplying the area of new industrial oil palm plantations in the target year by the biomass carbon stock of industrial oil palm plantations per hectare. We used values from Khasanah et al. (2015) for above-ground biomass and applied a 1:4 root-to-shoot ratio from Guillaume et al. (2018) to account for below-ground biomass. Carbon stocks were converted to CO₂-equivalent emissions using a factor of 3.67 (IPCC, 2014). Net greenhouse gas emissions from land-use change were calculated as the difference between gross emissions from oil palm plantation expansion and carbon sequestration in new industrial oil palm plantations in the target year. Values that are negative at the pixel-level are thresholded to 0 before calculating aggregate statistics.

GHG emissions from fires on peat soils in a given year were estimated based on the overlap between burned areas (Ministry of Forestry, 2025) and peat areas (Ministry of Agriculture, 2019), multiplied by an emission factor of 824 t CO₂ eq ha⁻¹, following the average emission estimates provided by Rodríguez Vásquez et al. (2021). This value represents the average CO₂ released per hectare during a single peat fire event, derived from measurements of burn depth, bulk density, and emission factors from peatland sites in Kalimantan, Sumatra, and Malaysia. Because peat fires may result from multiple land uses, emissions were allocated to the oil palm sector according to the proportion of modified peatlands within each FFB production unit that are covered by oil palm. Modified peatlands are defined as peat soils used for agriculture, including rice paddies, oil palm, pulpwood plantations, aquaculture, and other agricultural uses, as identified by MapBiomas Indonesia Collection 4 (2025). This choice reflects the fact that fire is often used as land management to clear the land before it is used agriculturally and might then spread to neighbouring areas (Purnomo et al., 2017), whereas fires in intact peatland forests are uncommon (Marlier et al., 2015).

GHG emissions from peat subsidence in a given year were estimated based on the area of industrial oil palm plantations located on peat soils within each FFB production unit for the target year. This area was multiplied by an emission factor of 90 t CO₂ eq ha⁻¹ yr⁻¹, averaged over a 30-year plantation cycle, to account for methane, nitrous oxide, and carbon dioxide emissions resulting from peat oxidation and subsidence after drainage (Cooper et al., 2020). 

Oil palm planted area is the extent of productive industrial oil palm. It was calculated using data provided by The TreeMap (Gaveau et al., 2022; The TreeMap, 2026) , who mapped the year in which oil palm is first planted in a given location. Oil palms were assumed to start producing fresh fruit bunches (FFB) in the third year after planting and continue to produce for up to 30 years. Based on these assumptions,the productive oil palm planted area assigned is the total area planted  three years before the target year. For example, the productive oil palm planted area in 2018 is represented by the total area mapped as oil palm plantation in 2015 by The TreeMap (The TreeMap, 2026).

Oil palm on peatlands refers to the total area of oil palm (Gaveau et al., 2022; The TreeMap, 2026) planted on peatlands. Peatland areas were derived from Indonesia's official peatland map, which was published by the Indonesian Center for Agricultural Land Resources Research and Development within the Ministry of Agriculture in 2019 (Ministry of Agriculture, 2019). 

Percent of export volume traded under a No Deforestation, No Peat, and No Exploitation (NDPE) or Company Zero Deforestation Commitment (ZDC), including zero-deforestation and deforestation- and conversion-free commitments (Global Canopy, 2025; Thomsen, 2025). Company commitments are assessed for exporter-trader groups handling 90% of total palm oil exports, following Forest 500 guidance

We estimated local fresh fruit bunch (FFB) production by combining provincial statistics on palm oil production (Badan Pusat Statistik (BPS), 2025) with maps of local oil palm planted areas, both under industrial (Gaveau et al., 2022; The TreeMap, 2026) and smallholder (Descals et al., 2026) management. We multiplied the estimated oil palm oil area (ha) by the Crude Palm Oil (CPO) productivity (tonne/ha) published by BPS (Badan Pusat Statistik (BPS), 2025) for smallholders and industrial plantations

The soy area in a given year that overlaps with deforestation in the five years before soy detection.

This metric represents the total deforestation of natural forests, occurring in a single year, within a specific concession. To calculate this metric, Gaveau et al. (2022) combines natural forests (as defined above) with the annual tree cover loss dataset developed by the University of Maryland (Hansen et al., 2013).  The tree cover loss dataset measures the removal of trees (tree height >5 m) if the canopy cover falls below 30%. It does not distinguish between removal of natural forests and planted trees. Collaborators at TreeMap then conduct additional photo interpretation and analysis to filter out errors. The same data sources are used on The TreeMap’s mapping platform The TreeMap’s Nusantara Atlas (2025). The metric is available on Trase Supply Chain’s Sankey diagram for active concessions trading in the target year and for all concessions on the platform’s map tool.

This metric represents total deforestation for conversion to pulpwood plantations in a single year, per concession. A subset of the annual deforestation metric, it is restricted to only those areas identified as natural forest one year before the target year and consistent with patterns of expansion of pulpwood plantations in the target year, according to the classification of Gaveau et al. (2022). 

This metric represents total deforestation for conversion to pulpwood plantations in a single year, per concession occurring on peatlands (Ministry of Agriculture, 2019). 

This metric measures annual deforestation for conversion to pulpwood plantation (Gaveau et al., 2022; The TreeMap, 2025), beginning in the year after a concession was first established. Thus, this metric follows the same definition as Deforestation for pulp per year but uses the concession’s establishment year as the baseline, regardless of later ownership changes. 

This metric estimates how much historical deforestation is associated with the wood pulp supply chain. The metric specifically focuses on areas where the establishment of pulpwood plantations are estimated to be mature enough to be harvested in the target year, considering the year of deforestation detection for pulpwood expansion. We use a retrospective approach, aggregating the deforestation across ten years, with a six-year lag between the first detection of deforestation and the year of harvest (also the year of export). By setting these parameters we are assuming that the plantations will be ready for harvesting, and consequently start to contribute to the volume traded in the sixth year after the deforestation event for new plantations. For example, the total area of land that was classified as forest in 2003, but had been converted to pulpwood plantations by 2014, would be considered as areas of deforestation directly connected to the wood pulp trade in 2020.

This metric estimates the major sources of GHG emissions from pulpwood plantations in Indonesia: subsidence on drained peatlands, fire on peatlands, and land use change, i.e., the clearing of vegetation (forest and non-forest) to establish pulpwood plantations. The GHG sequestration from new plantations was also estimated and incorporated into the net GHG emissions reported by Trase.

GHG emissions from industrial wood pulp land conversion (tonnes of CO₂-eq) were calculated by estimating all of the CO₂-eq potentially emitted due to annual land use change to pulpwood plantation, including both from natural forest and non-forest land cover. (as defined in Gaveau et al. (2022)), estimated within a given year and concession. This metric is calculated by multiplying the area converted each year to pulpwood plantation by an estimate of the carbon stock contained in each hectare of converted land independent of the fact whether than converted land was before forest or non-forest as we assumed that the carbon stock density map will account for the difference in carbon stocks between forest and non-forest.. Carbon stocks represent the total carbon stored in above- and below-ground biomass, but does not include carbon stocks in the soil. We create local estimates of above-ground woody biomass density (carbon stocks per hectare) using data from Harris  et al. (2021). Below-ground biomass density was estimated based on Harris et al. (2021)’s above-ground biomass density and the ratio of below-ground to above-ground biomass derived from Spawn et al. (2010). To convert biomass to carbon density, we used conversion factors of 0.47 t C t-1 for aboveground and 0.39 t C -1 for belowground, respectively (Diop et al., 2016). Carbon is converted into CO₂-eq emissions by multiplying by a factor of 44/12 ( or 3.67), based on the ratio between the molecular weight of C and CO₂ following IPCC (2014).

Net greenhouse gas emissions from industrial pulpwood plantation were calculated as the difference between gross emissions from pulpwood plantation expansion and carbon sequestration in new industrial pulpwood plantations. Carbon sequestration was estimated by multiplying the area of new industrial oil palm plantations in the target year by the biomass carbon stock of industrial pulpwood plantations across a plantation's full rotation. Based on research by Deshmukh et al., (2023), we assume that fully mature pulp plantations will, on average, store 42.4 tC/ha. We then adjust this maximum to estimate the average biomass within the plantation over its full rotation, as described by CFA & Quantis, (2022). This adjustment reflects the fact that perennial plantation systems are cyclical - while the plantation will contain lesser or greater stocks of carbon in individual years, the average across the full rotation can serve as a useful estimate of the carbon impact of land use change into long-term plantation forestry. Carbon stocks were converted to CO₂-equivalent emissions using a factor of 3.67 (IPCC, 2014).

The metric represents the (greenhouse gas) emissions from peat subsidence (tonnes of CO₂-eq) in a single year per concession, assigned to the target year. The emissions are calculated based on pulpwood planted area on peat (more details above) inside the concession for the target year. This area is multiplied by an emission factor of 73.4 tCO₂-eq, indicating a constant 20 tonnes of C loss for every hectare and year due to subsidence after peat drainage, as described by Carlson, Goodman and May-Tobin, (2015) and as recommended by the IPCC's Wetlands Supplement (2013).

Production of soy per jurisdiction for a given year.

This metric represents the GHG emissions from peat burning (tonnes of CO₂-eq) in a single year per concession. The emissions are calculated based on the area of overlap between the burned areas and peat areas (more details above) multiplied by 824 tonnes of CO₂-eq following the average emission factor estimates provided by Rodríguez Vásquez et al., (2021). The value represents an average amount of CO₂-eq per hectare that is released into the atmosphere in a unique event of fire on peat. Rodríguez Vásquez et al., (2021) based their estimate on other studies  from peatlands sites in Kalimantan, Sumatra, and Malaysia where depth of burn, bulk density, and average emissions factors were measured.

Pulpwood supply was mapped using annual industrial wood supply plans (Rencana Pemenuhan Bahan Baku Industri, RPBBI) submitted by Indonesia’s six active pulp mills and one wood chip mill to the Ministry of Forestry (MoF). These reports list suppliers, supplier types, and volumes in cubic metres. Two suppliers holding Land Conversion Permits (Izin Pemanfaatan Kayu, IPK) were classified as sourcing from natural forests. About 96% of pulpwood between 2015–2024 originated from industrial wood fibre plantation concessions (Hutan Tanaman Industri, HTI). Concession boundaries were identified from MoF’s Forest Utilization License dataset (2022–2025). Additional concessions linked to the Royal Golden Eagle (RGE) Group were included due to potential pulpwood expansion. Imports, traders, and community suppliers were aggregated as reported in the RPBBI. 

Pulpwood plantation area represents the extent of productive industrial pulpwood, most commonly Acacia mangium and Acacia crassicarpa. It was calculated using data provided by Gaveau et al. (2022), who mapped the annual expansion of industrial pulpwood plantations by visually inspecting a time series of cloud-free Landsat composites from 2000 to 2024. They also referred to indicative maps of pulpwood concessions and reviewed various online and press reports to identify the locations of these plantations. Because the mapped expansion year reflects the establishment rather than the mature stage contributing to trade flows, we applied a six-year lag between the detected expansion and the calculation of active pulpwood plantation area. For example, deforestation detected in 2016 for pulpwood production is interpreted as a productive plantation in 2022. The boundaries of the plantations were delineated at a 1:50,000 scale based on expert photo interpretation. Beyond spectral information, the visual interpretation also incorporated the structure of the landscape. Pulpwood plantations typically appear as long linear or rectangular blocks located near harvesting trails, with lower density than that of other forestry crops such as oil palm. The final mapped area includes recently harvested, immature, and mature stands, as well as failed plantations. Scattered community plantations were not reported, as they are more difficult to detect through remote sensing. 

Pulpwood planted area on peat is the area of peatland within  concessions that has been planted with pulpwood species, as defined by Gaveau et al.(2022), i.e. a subset of the total peat area. Trase identifies peatlands using Indonesia's official peatland map, which was produced by the Indonesian Centre for Agricultural Land Resources Research and Development (Balai Besar Penelitian dan Pengembangan Sumberdaya Lahan Pertanian in Indonesian) within the Ministry of Agriculture (Indonesian Center for Agricultural Land Resources Research and Development, Ministry of Agriculture (MoA), Government of Indonesia, 2019).

The concession area of the industrial timber plantations (HTI - Hutan Tanaman Industri) based on maps from Indonesia’s Ministry of Forestry.

This metric estimates the area burned within each concession in the target year. To measure burned area, we rely upon the burned area dataset from MoF. The dataset  has been published yearly since 2015 by MoF and is mapped at 1:250,000 scale on a monthly basis by visual interpretation of the extent of forest and land fires from Landsat-8, Landsat-9 and Sentinel-2 satellite imagery. The visually identified burned areas are then validated by experts using hotspot data from the Terra-Aqua MODIS Satellite, NOAA 20, S-NPP and Landsat-8 as well as field reports (Endrawanti et al., 2018). It is worth noting that recent research indicates that this data might underestimate the total burned area. For example, Gaveau et al. (2021), present evidence that the 2019 burned area was 90% greater than the area reported by MoF.

Burned peat refers to peat areas within pulpwood concession boundaries that have been affected by fire, as identified from burned area data (Ministry of Forestry, 2025)

Natural forest area is the total area of remaining forest derived from Gaveau et al.(2022). Gaveau et al. (2022) combine primary forest cover in 2000 (Margono et al., 2014), and the global forest change data developed at the University of Maryland (Hansen et al., 2013). This definition of ‘forest’ excludes young regenerating forests, agroforests, scrublands, tree plantations, agricultural lands, grasslands, and non-vegetated areas (Gaveau et al., 2022). 

Peatland areas were derived from Indonesia's official peat map, produced by the Indonesian Center for Agricultural Land Resources Research and Development within the Ministry of Agriculture (2019).

Gross emissions attributable to deforestation. Estimates are based on the total biomass removal stored in the native vegetation. The numbers are subsequently transformed in CO₂-eq by multiplying carbon stock per 3.67.

The territorial deforestation data is a composite dataset, integrating information from Prodes Amazon (INPE, 2025), Prodes Cerrado (INPE, 2025), and MapBiomas deforestation (MapBiomas, 2025; Souza et al., 2020) data for all other biomes. Deforestation is defined as the conversion of primary native vegetation. 

Total deforestation per jurisdiction for a given year.

Cattle deforestation per ton is a measure of the deforested area (pasture deforestation) associated with the production of beef for a given year, expressed in hectares per ton of carcass. Pasture deforestation was calculated by overlapping pasture maps with the annual territorial deforestation five years prior to the target year for pasture, including the year the pasture was detected.

Gross emissions associated with the production of beef (pasture deforestation) for a given year, expressed in hectares per ton of carcass. Pasture deforestation was calculated by overlapping pasture maps with the annual territorial deforestation five years prior to the target year for pasture, including the year the pasture was detected.

Size of the cattle herd for a given year.

Area used for pasture in hectares. Available only for the Chaco region.

Percentage of the beef that is exported by a trader with a zero deforestation commitment.

Total production of corn (first and second crop).

Average corn yield per jurisdiction in a given year.

The percentual volume of corn produced in a specific jurisdiction that was traded by a company signing zero deforestation commitments in a given year.

The soy area in the target year that overlaps with deforestation five years before soy detection.

Gross greenhouse gas (GHG) emissions from territorial  deforestation, expressed in tonnes of CO₂-equivalent (t CO₂-eq) are determined by calculating the emissions resulting from the removal (land conversion) of native vegetation carbon stocks. Carbon stocks include the above and below ground biomass including necromass. The carbon stocks are based on the values provided by SEEG v12 (SEEG, 2024). Carbon is converted into CO₂-eq by multiplying the values by 3.67, based on the ratio between the molecular weight of C and CO₂.

Soy planted in areas recently deforested per tonne of soy produced in a given jurisdiction.

Emissions attributable to soy deforestation per tonne of soy produced in a given jurisdiction. Estimated based on the balance between the carbon stored in the native vegetation and the soy fields in CO₂-eq.

Percentage of the soy that is exported by a trader with a zero deforestation commitment.

Production of soy per jurisdiction in a given year.

Area of land used to grow soybeans per jurisdiction in a given year.

Average soy yield per jurisdiction in a given year.

Pasture deforestation is quantified as the total area of deforestation converted to pasture within a five-year allocation period. We calculate this by spatially overlapping annual pasture maps from MapBiomas with annual territorial deforestation data. This analysis identifies all deforestation that occurred within the five years prior to, and including, the year pasture was detected, yielding a cumulative estimate of deforestation linked to pasture land use. The territorial deforestation data is a composite dataset, integrating information from Prodes Amazon (INPE, 2025), Prodes Cerrado (INPE, 2025), and MapBiomas deforestation (MapBiomas, 2025; Souza et al., 2020) data for all other biomes.

Cattle deforestation estimates the amount of deforestation associated with the production of beef for a given year. Cattle deforestation combines the area deforested for pastures and cattle production, assuming a 5 year life cycle from birth to slaughter, accounting for the average pasture deforestation during the 5 year lifespan.

Impact indicators