Calculating Supply Chain Deforestation Emissions for Brazil’s Meatpacking Sector: A New Corporate Accounting Methodology

Calculating Supply Chain Deforestation Emissions for Brazil’s Meatpacking Sector: A New Corporate Accounting Methodology | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Calculating Supply Chain Deforestation Emissions for Brazil’s Meatpacking Sector: A New Corporate Accounting Methodology AMINTAS BRANDAO JR, SARAH Klopatek, LISA RAUSCH, FABIO DIAS, JACOB MUNGER, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9138992/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 5 You are reading this latest preprint version Abstract Background: The Greenhouse Gas Protocol (GHGP) distinguishes Scope 1 emissions, which occur directly at the facility or company, from Scope 2 emissions linked to electricity use, and Scope 3 emissions associated with other supply-chain inputs. Estimating Scope 3 deforestation-related emissions in Brazil’s cattle sector remains challenging due to the fragmented nature of supply chains and the lack of national farm-to-gate traceability. Most companies cannot identify indirect suppliers, where deforestation is increasingly concentrated, making Scope 3 land-use change emissions difficult to quantify. Instead, they use top-down statistical land-use change models based on national, state or municipal emission factors which are usually limited regarding variation of companies’ carbon footprint. To address this gap, we developed a spatially explicit method aligned with the GHGP’s Land Sector and Removals Guidance, using data already available to slaughterhouses to estimate cattle supply zones and compare alternative input datasets and methods. Results: Emissions estimates varied significantly depending on the deforestation dataset and spatial attribution method. Global Forest Watch deforestation data produced gross emissions nearly 100% higher and discounted emissions about 120% higher than PRODES data. Supply zones based on locations of direct suppliers produced emissions up to 25% lower than fixed-radius buffers around slaughterhouses. When adjusted for market share and temporal discounting, total emissions decreased by 85–97%. While supply zones improved geographic precision relative to fixed-radius buffers and municipal, state or country estimates, they still include both suppliers and non-suppliers and do not fully capture company-specific mitigation actions. Conclusions: This study presents a practical, replicable method for estimating deforestation-related Scope 3 emissions using currently available data for Brazil’s cattle sector. The approach is consistent with GHGP recommendations and supports corporate reporting. Supply zones represent an important advance over traditional top-down statistical land-use change methods. Improving accuracy of Scope 3 emissions from deforestation however, will require expanded traceability capable of identifying all animals and all tiers of cattle suppliers. Scope 3 Deforestation Greenhouse Gas Protocol (GHGP) Cattle supply chains Land-use change emissions spatially explicit accounting Figures Figure 1 Figure 2 Figure 3 1. BACKGROUND The increasing awareness of forest loss associated with agricultural commodities has intensified regulatory oversight of global supply chains. Companies are now under mounting pressure to meet non-deforestation commitments, to ensure the products they source adhere to local laws governing land use, and to comply with laws governing global trade of commodities, such as the European Union Deforestation Regulation (EUDR, 2025). In this context, quantifying the links between supply chains, deforestation, and associated greenhouse gas (GHG) emissions has become an increasingly urgent task for the private sector (Panwar et al., 2023). The Greenhouse Gas Protocol (GHGP), developed through a partnership between the World Resources Institute (WRI) and the World Business Council for Sustainable Development (WBCSD), is the most widely adopted framework for quantifying emissions across three scopes: direct emissions from owned or controlled sources (Scope 1), emissions from purchased electricity, heat, and cooling (Scope 2), and other indirect value-chain emissions (Scope 3) including those from land use change (GHGP 2011; GHGP 2013). In Brazil, accounting for Scope 3 land-use change emissions linked to cattle production faces significant structural barriers. The supply chain is highly fragmented, with animals often moved across multiple properties before slaughter, complicating efforts to trace land-use impacts. Different slaughterhouses frequently buy from the same groups of suppliers, making it difficult to attribute deforestation to specific facilities. Additionally, limited transparency beyond direct suppliers, those selling cattle to slaughterhouses, enables deforestation linked to indirect suppliers, which sell cattle earlier in the supply chain, to persist, despite zero-deforestation agreements (Walker et al., 2013 ; Gibbs et al., 2015 ; Alix-Garcia and Gibbs, 2017 ; Rajão et al., 2020 ; Skidmore et al., 2021 ; West et al., 2022 ; Brandão Jr. et al., 2023 ; Skidmore et al., 2024 ). Beyond supply chain complexity, legal and regulatory restrictions on data sharing, the lack of harmonized reporting standards, and concerns over business confidentiality further hinder effective Scope 3 emissions accounting (Stenzel and Waichman, 2023 ). Providing meatpacking companies with robust tools to accurately measure and manage emissions is critical. The expansion of cattle pastures in Brazil is linked to more than 80% of all deforestation since 1985 (Souza Jr. et al., 2020 ). In 2024–2025, Brazil accounted for 19% of the global beef production (USDA, 2025), while deforestation remained the dominant source of emissions–responsible for 46% of the national greenhouse gas output the previous year, with nearly one-third concentrated in the Amazon (SEEG, 2023). Full supply chain traceability systems, such as individual ear tags, could help to generate the animal-level emissions data needed to precisely quantify company-specific Scope 3 emissions. Brazil however, still lacks a national farm to slaughter cattle traceability system. Building such a system would require substantial digital and physical infrastructure, clear regulatory frameworks and stable funding in addition to effective enforcement, which means it is likely to take many years to fully materialize. To date, SISBOV which is a voluntary system implemented in 2002 is the only system that has made meaningful progress in this area, yet it covers only 1% of the country’s ranches (Rausch et al., 2020 ). At the same time, subnational initiatives offer more immediate pathways. The state of Santa Catarina in southern Brazil has implemented a state-level individual traceability program (Rocci, 2025), though logistical and institutional constraints make scaling to the national level difficult (Cavalcante et al., 2018). Pará, recently delayed its traceability rollout from 2026 to 2031 citing limited international market interest in individually traceable cattle from the state (Júnior & Russo, 2025 ). In Mato Grosso, the Passaporte Verde Led by IMAC, links socio-environmental traceability to market access, targeting farms with recent illegal deforestation and supporting remediation over a 48-month period (Mato Grosso, 2025). At the same time, Brazil’s meatpacking sector continues to move towards interim solutions to calculate Scope 3 emissions using available data and guidance. Corporate reporting of land-use emissions relies on the Greenhouse Gas Protocol’s Land Sector and Removals Guidance (2023), which outlines three main approaches: statistical land use change (sLUC), direct land use change (dLUC), and jurisdictional direct land use change (jdLUC) (Fitts et al. 2025). The sLUC approach is applied when companies have access to only jurisdiction-level sourcing data. The dLUC approach requires traceability at the level of land management unit, such as a farm or plot. The jdLUC method combines jurisdiction-level sourcing with crop distribution spatial data. Among these, sLUC is the most widely adopted due to its lower data requirements. In contrast, dLUC and jdLUC are rarely implemented, largely due to restricted access to detailed sourcing information and the absence of standardized procedures. Although these methods were developed for broad sectoral application, many target industries still lack the robust empirical foundations needed to apply the (Hettler and Graf-Vlachy 2023). Here we develop an approach that bridges dLUC and jdLUC, using cattle supply zones to estimate deforestation-linked emissions for slaughterhouses. This approach relies on slaughterhouse geolocation, around which a zone containing a high concentration of suppliers– the supply zone– can be estimated (Brandão Jr. et al., 2023 ), enabling the assessment of company market share. Previous studies have proposed various methods to delineate these zones, including drawing buffers around slaughterhouse coordinates (Alix-Garcia and Gibbs, 2017 ), estimating maximum purchasing distances (Barreto et al., 2017 ), and using trip length distributions (Santos and Costa, 2018). In this paper, we compare the simplest approach to delineate supply zones and a specific method developed for the Brazilian cattle sector by Brandão Jr. et al ( 2023 ) that delineates zones using spatial proximity between slaughterhouses and locations of its direct suppliers. Analyzing detailed animal movement data, Brandão Jr. et al found that over 80% of indirect-to-direct transactions occurred within 100 kilometers of a direct supplier, with a median distance of only 20 kilometers. These patterns allow for the construction of slaughterhouse supply zones based on direct supplier locations. Supply zones capture the geographic footprint of procurement and much of the associated deforestation, especially in landscapes where cattle ranching dominates land use. Here we designed our model to estimate Scope 3 emissions from deforestation at the slaughterhouse level using spatial proxies for sourcing dynamics. Our goal was to provide companies with a scalable, bottom-up framework for estimating emissions even in data-scarce contexts. We focused on federally inspected meatpacking plants (SIFs), which represent most slaughter operations in the state of Pará, averaging 70% of the annual cattle slaughter from 2013 to 2023 according to property-level animal movement data ( Figure S1 ). We focus on Para because of its uniquely complete and accessible data on cattle movements and property boundaries, but similar analyses could be done nation-wide using company locations, direct supplier locations (when available), and estimated median distances between indirect suppliers and direct suppliers. We estimated emissions for Pará’s SIF slaughterhouses using two main scenarios to reflect varying levels of corporate information of data access: (1) slaughterhouses locations only, (2) supply zones defined by the spatial footprint of direct suppliers using the method by Brandão Jr. et al. ( 2023 ). Applying these scenarios from 2004 to 2023, we assessed how data granularity affected supplier attribution, emissions trajectories, deforestation baselines, and market share calculations. Unlike conventional sLUC approaches that apply uniform deforestation rates, our method incorporates spatial and temporal variation, aligning emissions estimates more closely with actual land-use dynamics and procurement behavior (Schmidt et al., 2022; CDP, 2020 ). This work contributes to the transition encouraged by the GHGP and the Science Based Targets initiative toward geographically explicit, supplier-linked accounting (GHGP, 2022; SBTi, 2024). As companies face increasing pressure and regulatory mandates to disclose supply-chain Scope 3 emissions, such as under the EU Corporate Sustainability Reporting Directive (CSRD, 2025), our supply zone approach offers a practical and scientifically grounded solution. Its implementation represents a significant advancement in improving the accuracy and transparency of GHG inventories for corporations sourcing beef from Brazil. 2. METHODS We developed a spatially explicit framework to estimate deforestation-related greenhouse gas emissions attributable to individual slaughterhouses in Brazil, consistent with Scope 3 guidance under the Greenhouse Gas Protocol. Drawing on two decades of deforestation data from 2004 to 2023, combined with shorter time series on cattle movements, rural property boundaries, and spatially explicit carbon stocks, we implemented a five-step analytical process. This process included delineation of slaughterhouse supply zones; identification of pasture-driven deforestation within each zone; quantification of associated carbon emissions; application of a 20-year temporal discounting factor to weigh historical emissions; and allocation of emissions to slaughterhouses based on their market share within each supply zone. We evaluated the framework under three main scenarios representing increasing levels of supply chain resolution, ranging from simple geographic buffers to supply zones derived from the geolocation of direct suppliers, defined as farms delivering cattle to slaughterhouses. All emissions were standardized to 2023 as the reporting baseline. The framework was designed to be transparent, replicable, and compatible with corporate GHG reporting systems using commonly available supply chain data. Emissions associated with post-deforestation land management practices were excluded from the analysis. 2. 1 Pre-processing To build our model, we integrated property-level cattle movement records, deforestation maps, slaughterhouse locations, property boundaries, and spatially explicit carbon emission factors ( Table 1 ). The dataset spans 2004 to 2023, with traceability of cattle movements available from 2013 onward. The model was designed to align with GHGP guidance and to be replicable by companies seeking to estimate Scope 3 land-use change emissions in data-constrained contexts. Below, we describe the data sources and preprocessing steps used to construct the analytical framework. Cattle movement data between properties and to slaughterhouses. We used data from the Animal Transit Records (Guia de Trânsito Animal in Portuguese, GTA) system maintained by the Pará State Agency for Agricultural and Livestock Health (ADEPARA), which records all cattle movements between farms and to establishments such as slaughterhouses in the state of Pará. Mandatory under Brazilian law, the GTA records serve as the primary mechanism for livestock traceability and animal health surveillance. Our dataset covers the period from 2013 to 2023 and includes records for more than 204 million head of cattle transported across the state. Each record contains standardized information on the purpose of the movement, including slaughter, fattening, breeding, auction, or other uses; the number of animals; sex classification; age group; and unique identifiers for both origin and destination properties, derived from the landowner’s individual (Cadastro de Pessoas Físicas, CPF) or corporate (Cadastro Nacional da Pessoa Jurídica, CNPJ) registration. Records also report on the date of movement and the municipalities of origin and destination. Using these data, we identified direct suppliers, defined as farms that sold cattle directly to slaughterhouses for immediate processing. Direct suppliers accounted for approximately 29 million head of cattle over the study period, and their spatial distribution was used as a proxy to delineate slaughterhouse supply zones. Property boundaries. We obtained the Rural Environmental Registry (Cadastro Ambiental Rural, CAR) dataset through December 2023 from the Pará State Environmental Secretariat (SEMAS). CAR is a mandatory national registry requiring all rural properties in Brazil to report their boundaries and landowner information. The dataset includes geo-referenced property boundaries represented as polygons, together with landowner identifiers recorded as CPF for individuals and CNPJ for companies. Following the approaches described by Skidmore et al. (2020, 2021, 2022), West et al. (2022), and Brandão Jr. et al. (2023), we matched cattle movement records from the GTA to CAR properties using landowner identifiers and complementary attributes, including municipality and property name. This matching process enabled the spatial linkage of cattle movements to their corresponding rural properties. The resulting dataset identifies and geo-references the properties involved in cattle transactions, hereafter referred to as GTA properties. Federal slaughterhouses. We obtained the official registry of 22 companies operating federally inspected slaughterhouses (Serviço de Inspeção Federal, SIF) in Pará from the Brazilian Ministry of Agriculture, Livestock, and Supply (MAPA). SIF facilities are authorized to process meat for interstate and international trade and are required to comply with federal sanitary inspection standards. The registry provides company names, physical addresses, and unique SIF registration codes for each facility. We geocoded these addresses to obtain precise geographic coordinates (latitude and longitude) using the Google Maps API, with manual verification for locations that could not be automatically resolved. These facilities processed approximately 70% of all cattle slaughtered in the state, with an average of 1,861,546 head per year (Figure S1). The remaining 30% were slaughtered at state-inspected (SIE) and municipally inspected (SIM) facilities that serve local markets. To protect commercial confidentiality, all SIF slaughterhouses were aggregated into a single analytical category rather than reported individually. Deforestation. We used annual deforestation maps from PRODES, produced by Brazil’s National Institute for Space Research (INPE), based on 30 × 30 m Landsat imagery. The PRODES Brazil dataset includes cumulative deforestation in primary forests through 2007, annual deforestation from 2008 to 2023, forest and non-forest vegetation classes, and areas covered by water or cloud. To extend temporal coverage to 2004–2006, we incorporated a vector product released by INPE in 2016 for the Brazilian Legal Amazon, which provides annual deforestation maps from 2001 onward. From this product, we generated a mask of cumulative deforestation through 2003 and converted the 2004–2006 annual deforestation layers from a vector to a raster format to match the resolution and spatial alignment of the 2023 PRODES Brazil raster. To ensure consistency across the time series, we excluded all cells outside the PRODES deforestation extent through 2007. The resulting dataset is a continuous raster representing cumulative deforestation through 2003 and annual deforestation from 2004 to 2023. For comparison, we also used the Global Forest Watch dataset (Global Forest Watch; Hansen et al., 2013), which provides annual tree cover loss at 30 m resolution for 2004–2023 based on automated time-series analysis of Landsat imagery. Whereas PRODES relies on expert visual interpretation to identify clear-cut deforestation in primary forest, Global Forest Watch applies spectral change detection to quantify all forms of tree cover loss globally, irrespective of forest type or disturbance driver. This comparison allowed us to assess the consistency and robustness of deforestation estimates across monitoring systems with distinct methodological approaches. To estimate the share of deforestation attributable to pasture expansion, we used land cover transition data from MapBiomas Collection 10 (MapBiomas) as a proxy. Specifically, we calculated the proportion of transitions from forest (class 1) to pasture (class 15) relative to transitions from forest to farming (class 14) and applied this ratio to the total deforested area within each supply zone to estimate the contribution of pasture expansion to overall deforestation. Carbon emission factors . We incorporated spatially explicit carbon emission factor maps from Harris et al. (2021). These maps provide annual estimates of aboveground biomass carbon, belowground biomass carbon, and soil organic carbon stocks in a harmonized time series from 2004 to 2023, with all components mapped at 30-meter resolution. Each carbon stock layer was reprojected and spatially aligned with the PRODES deforestation maps to enable pixel-level calculations of total carbon density. To calculate emission factors, we summed the three carbon pools and converted carbon to carbon dioxide equivalent using the standard molecular weight ratio of 44 to 12, representing the mass of CO₂ relative to elemental carbon. Emission factors are expressed in metric tons of carbon dioxide equivalent per hectare (tCO₂e/ha). For example, a pixel with 150 tC/ha in aboveground biomass, 30 tC/ha in belowground biomass, and 50 tC/ha in soil organic carbon yields a total carbon density of 230 tC/ha, equivalent to 843.3 tCO₂e/ha (230 × 44 ÷ 12). This spatially explicit approach allowed us to estimate carbon emissions from deforestation events based on the specific carbon stocks present at each location and year, accounting for spatial heterogeneity in forest carbon density across Pará. Table 1. Main datasets used in this study. Data Period Source Slaughterhouse location 2013-2023 MAPA Cattle transaction records 2013-2023 Pará’s State Sanitation Agency Property maps 2023 Para’s State Environmental Agency Annual Deforestation maps 2004-2023 INPE Emission factors (tCO2e/ha/year) 2004-2023 Global Forest Watch (GFW) based on Harris et al. (2021) 2.2 Estimating Baseline Emissions Our five-step approach to estimate deforestation and GHG emissions linked to cattle supply chains ( Table 2 ) accounts for different levels of supply chain data available for each slaughterhouse. Define spatial boundary: We delineated the spatial extent of supply zones for SIFs based on the available data for each state. These zones were defined either as generalized geographic buffers around slaughterhouse locations or as areas surrounding direct suppliers. Calculate deforestation: We quantified deforestation within each supply zone using PRODES data as the primary source. Additionally, we conducted a comparative analysis with GFW data to assess consistency across monitoring approaches. Calculate GHG emissions: We converted deforestation into GHG emissions using a committed emissions approach. This calculation incorporated pixel-level carbon stock values from aboveground, belowground, and soil carbon pools, based on the model by Harris et al. (2021). Apply linear discounting: We applied a linear discounting factor to distribute emissions over a 20-year period, in accordance with GHGP guidelines for land-use change emissions reporting. Apply market share: We allocated emissions to individual slaughterhouses using either market share proxies or direct traceability data, depending on the scenario and the level of supply chain resolution available. We applied the five-step approach under two main scenarios, with each scenario representing a distinct level of data availability and spatial precision ( Table 2 ): Scenario 1: Buffer-Based Approach. This scenario applies when the geographic locations of slaughterhouses are the only available information, without access to supplier-level data. Supply zones were therefore defined using fixed 200 kilometers radius buffers around each facility (based on insights from Brandao Jr et al., 2023), and all deforestation emissions within these buffers were attributed to the corresponding slaughterhouses. We tested two variations: 1a: Emissions were distributed equally among all slaughterhouses operating within each buffer zone, without market share adjustment. 1b: Emissions were allocated proportionally to the market share of SIFs within each supply zone. Market share was calculated as the ratio of cattle slaughtered by SIF slaughterhouses within each supply zone to the total number of cattle slaughtered by all slaughterhouses operating in that zone, including those under federal, state, and municipal inspection. Scenario 2: Supplier-Based Approach. This scenario incorporates supplier data to construct spatially explicit supply zones, following the methodology of Brandão Jr. et al. (2023), which uses spatial autocorrelation to identify contiguous sourcing regions based on direct supplier concentrations. Two temporal variants were considered: Scenario 2a: Based on suppliers active in 2023, representing the most recent supply chain configuration. As in Scenario 1b, emissions were allocated proportionally based on market share. Scenario 2b: Includes all suppliers active between 2013 and 2023, capturing a full decade of sourcing dynamics. As in Scenario 1b, emissions were allocated proportionally based on market share. Table 2. Five-step methodology to estimate deforestation and GHG emissions linked to cattle supply chains in Pará applied under distinct scenarios depending on the supply chain data available. Step Scenario 1: Buffer-Based Approach Scenario 2: Supplier-Based Approach 1a 1b 2a 2b Step 1: Define spatial boundary 200 km buffer 200 km buffer Direct Suppliers (2023) as proxy Direct Suppliers (2013–2023) as proxy Step 2: Calculate deforestation. PRODES (GFW for comparison) PRODES (GFW for comparison) PRODES (GFW for comparison) PRODES (GFW for comparison) Step 3: Calculate GHG emissions. Pixel-level carbon stocks Pixel-level carbon stocks Pixel-level carbon stocks Pixel-level carbon stocks Step 4: Apply linear discounting 20-year linear discounting 20-year linear discounting 20-year linear discounting 20-year linear discounting Step 5: Apply market share No market share Market share Market share Market share Step 1: Define spatial boundary We defined spatial boundaries around each slaughterhouse to identify areas with deforestation. The approach varied based on available data. Scenario 1: Buffer-Based Approach When supplier data were unavailable, we created a 200 km radius circles around each slaughterhouse. This distance reflects typical cattle transport distances in Pará (Brandão Jr. et al., 2023). All deforestation within these circles was considered potentially linked to cattle supply chains. 1a: Slaughterhouse Buffer No Market Share. Assigns 100% of emissions within the buffer to that slaughterhouse after the linear discount. 1b: Slaughterhouse Buffer with Market Share. Distributes emissions proportionally among all slaughterhouses operating in the buffer based on their relative cattle slaughter volumes (market share) after the linear discount. Scenario 2: Supplier-Based Approach When supplier data were available from GTA traceability records, we identified the actual properties that sold cattle to each slaughterhouse. We then connected these supplier locations into contiguous zones using spatial autocorrelation (Brandão Jr. et al., 2023), creating irregular boundaries that follow real cattle sourcing patterns rather than simple geometric circles (Figure 1). 2a: Direct Supplier's Zone 2023. This analysis includes only suppliers active in 2023, providing a snapshot of the current supply chain geography while applying both the linear discount and market share assumption. 2a: Direct Supplier's Zone 2004-23. This approach includes all suppliers active between 2004 and 2023, generating 20 annual supply zone maps. These maps were overlaid, and each area was labeled according to the most recent year of supplier activity. Because comprehensive GTA data only became available starting in 2013, we assumed that the 2013 supply zone configuration remained constant for the earlier period from 2004 to 2012. The analysis applies both the linear discount and market share assumptions throughout. Step 2: Calculate deforestation. Annual deforestation was quantified by intersecting PRODES data with each slaughterhouse’s supply zone. For each year between 2004 and 2023, we calculated the deforested area within supply zones under each scenario. All spatial operations were performed using a projected coordinate system appropriate for area calculations (EPSG:5880, Brazil Polyconic). We also estimated deforestation using data from GFW to enable comparison with an alternative monitoring system commonly used by companies in emissions reporting (Lucia et al., 2025). While PRODES maps clear-cut deforestation defined as the complete removal of native forest cover relative to a 1988 baseline, GFW captures a broader spectrum of canopy loss. This includes both full clearing and partial degradation, in which forest cover is reduced but not entirely removed. Such losses may result from logging, fire, seasonal dynamics, or natural disturbances and reflect intermediate transitions between forested and non-forested conditions. The two systems also differ in methodological consistency over time. PRODES has applied a stable and continuous methodology since 2008, enabling reliable year-to-year comparisons of deforestation trends (INPE, 2025). In contrast, GFW implemented a major methodological update in 2015 (Weisse and Potapov, 2021), which may affect the comparability of long-term trend analyses. In both deforestation estimates, we applied a proportional adjustment to isolate the share attributable to cattle pasture expansion. This factor was derived from MapBiomas land-cover transition data by calculating the proportion of transitions from forest to pasture relative to total transitions from forest to agricultural land and applying this ratio to the total deforested area within each supply zone. Step 3: Calculate GHG emissions. We estimated annual land-use change emissions by intersecting deforestation data with spatially explicit emission factor maps from Harris et al. (2021). Emissions were calculated as the product of deforested areas (activity data) and emission factors, which quantify GHGs released per hectare (GHGP, 2013). Adopting a committed emissions approach, we assumed that the full carbon stock of each deforested pixel was released in the year of clearing. This method aligns with current recommendations for gross emissions accounting and avoids uncertain assumptions about post-clearing carbon dynamics. Emission factors from Harris et al. (2021) incorporate aboveground and belowground biomass and soil organic carbon, expressed in metric tons of CO₂ equivalent per hectare per year for 2002–2023. To ensure spatial consistency, we reprojected and resampled these layers to match the resolution and coordinate system of PRODES and GFW data. For each slaughterhouse and year (2004–2023), we overlaid annual deforestation maps with the corresponding emission factor layer. Emissions were calculated by multiplying the area of each deforested pixel by its specific carbon value and summing results within each sourcing zone. This procedure was applied across all sourcing scenarios, generating an annual time series of gross emissions per plant at varying levels of supply chain resolution. We did not account for forest regrowth, carbon removals, or post-clearing land use due to limited data availability. All spatial analyses were conducted using the Google Earth Engine (GEE) platform. Step 4: Apply linear discounting We applied a 20-year linear discounting function to the gross emissions calculated in Step 3, following GHGP recommendations. Each deforestation event was weighted based on its age relative to the reporting year, with a 0.5% annual reduction. For instance, emissions from 2023 were counted at 9.75% and from 2004 at 0.25%. Events prior to 2004 were excluded from the inventory. As an example, a 2010 deforestation event generating 1000 tCO₂e in gross emissions would contribute only 32.5 tCO₂e (3.25% of 1000 tCO2e) to the reported emissions for 2023. This discount was applied uniformly across all scenarios. Step 5: Apply market share In the final step, we allocated the discounted emissions to individual slaughterhouses according to the structure defined in each scenario. While the total emissions from deforestation remained constant within each sourcing zone, the method of allocation varied depending on data availability and supply chain resolution. For the period from 2013 to 2023, we used actual market shares calculated from GTA slaughter records. For 2004 to 2012, when comprehensive GTA data were not available, we assumed market shares remained constant at their 2013 values (Figure S4). Scenario 1a: Slaughterhouse Buffer No Market Share. In the absence of supplier-level data, we attributed 100% of emissions within each buffer zone to the SIF slaughterhouse located in that zone, assuming full responsibility for all deforestation after applying the linear discount. For example, if a buffer zone generated 32.5 tCO₂e of discounted emissions in 2010, the entire amount would be allocated to that SIF facility. This scenario represents the simplest allocation method, where the SIF slaughterhouses were assigned all emissions within its 200 km radius buffer. Scenario 1b: Slaughterhouse Buffer with Market Share. Using the same 200 km buffer zones as Scenario 1a, we applied market share adjustments to allocate emissions proportionally among all slaughterhouses operating within each zone. Using the same example, if SIF slaughterhouses accounted for 67% of total cattle slaughter in that zone in 2010, we would allocate 67% of 32.5 tCO₂e, or 21.8 tCO₂e, to the SIF facility. The remaining 33% (10.7 tCO₂e) would be attributed to non-SIF slaughterhouses operating in the same buffer zone. This approach recognizes that multiple facilities may source cattle from overlapping geographic areas. Scenario 2a Direct Supplier's Zone 2023 and Scenario 2b Direct Supplier's Zone 2004-23. When direct supplier data were available to define spatially explicit sourcing zones, we calculated each SIF slaughterhouse's proportional share of total cattle slaughter within its supply zone for each year. This proportion was then used to allocate emissions among all slaughterhouses operating in that zone. The calculation method was identical to Scenario 1b: if SIF market share was 67% in 2010, we allocated 67% of emissions (21.8 tCO₂e from the 32.5 tCO₂e total) to SIF facilities. The key difference from Sc1 is that supply zones were defined by actual supplier distributions rather than fixed 200 km buffers, resulting in more spatially accurate zones. Scenario 2a used only suppliers active in 2023, while Scenario 2b incorporated all suppliers from 2004 to 2023, capturing temporal dynamics in sourcing patterns. 3. RESULTS 3.1 Spatial and temporal patterns of SIF supply zones The total spatial extent of supply zones varied substantially across methods, ranging from 43.7 Mha in Scenario 2a to 74.9 Mha in Scenarios 1a and 1b (Fig. 1 ; Table 3 ). Scenarios 1a and 1b, which relied on uniform 200 km radius buffers around each slaughterhouse, produced the largest estimated supply zones, totaling 74.9 Mha. In contrast, Scenario 2a delineated supply zones using spatially explicit sourcing areas based exclusively on direct suppliers active in 2023, resulting in a total area of 43.7 Mha, a 42% reduction relative to buffer-based zones. Expanding the temporal scope, Scenario 2b incorporated all direct suppliers active between 2013 and 2023, increasing the total supply zone area to 60.1 Mha, a 38% increase relative to Scenario 2a. Despite this expansion, Scenario 2b remained 20% smaller than the buffer-based zones. Although buffer-based zones were substantially larger in total area than supplier-based zones, the extent of pasture included within each zone was remarkably similar. In Scenarios 1a and 1b, pasture covered 20.7 Mha, corresponding to 28% of the total zone area, while natural vegetation accounted for 49.9 Mha (67%). Despite its smaller overall extent, Scenario 2a captured a nearly equivalent pasture area of 20.1 Mha, representing 46% of the zone, while natural vegetation declined sharply to 21.2 Mha (49%), a 57% reduction relative to the buffer-based estimate. In Scenario 2b, pasture expanded modestly to 21.8 Mha (36%), whereas natural vegetation increased to 34.6 Mha (58%). Relative to Scenario 2a, this corresponds to an 8% increase in pasture area and a 63% increase in natural vegetation. Together, these results suggest that supplier-based zones more precisely capture active cattle production areas while excluding distant forest regions with limited connection to current slaughterhouse supply chains. At the same time, they indicate that buffer-based zones encompass pasture areas comparable to those identified through supplier-based approaches, and that much of the additional area captured under the multi-year supplier scenario corresponds to forested areas, rather than current supplier activity. Table 3 Estimated supply zone areas for SIF slaughterhouses by scenario Scenario Supplier Type Period Delimitation Criterion Estimated Zone Area (Mha) Pasture 2023 Area (Mha) Natural Vegetation 2023 (Mha) 1a and 1b All (Direct and Indirect) 2023 Fixed 200 km buffer 74.9 20.7 49.9 2a Direct 2023 Brandao Jr et al. (2023) method to define the zones based on direct suppliers 43.7 20.1 21.2 2b Direct 2013–2023 60.1 21.8 34.6 3.1 PRODES-based emissions Buffer zones centered on slaughterhouses (Scenarios 1a and 1b) and zones defined by the spatial distribution of direct suppliers (Scenarios 2a and 2b) yielded total gross emissions exceeding 3.7 billion metric tons of CO₂ equivalent between 2004 and 2023 (Fig. 2 , upper panel). Scenario 2b, which included all identified direct suppliers over the full 20-year period, produced the highest estimate at 4,202 Mt. Scenario 2a, focusing only on 2023 direct suppliers, yielded 3.7 billion metric tons of CO₂, representing a 12% reduction compared to Scenario 2b. Scenarios 1a and 1b, both using slaughterhouse buffers, produced identical gross emissions of 3.9 billion metric tons of CO₂. The temporal trends in Fig. 2 (upper panel) showed that all four scenarios followed similar trajectories throughout the 2004 to 2023 period. The lines were nearly overlapping, indicating consistent patterns across different spatial definitions. Gross emissions declined sharply from 2004 to 2012, dropping from approximately 400 Mt to around 100 Mt, reflecting the well-documented reduction in Amazon deforestation during this period. From 2012 to 2015, emissions stabilized at relatively low levels around 100 Mt. After 2015, a gradual increase began, accelerating notably from 2018 onward and reaching peak values of around 280 Mt in 2020 and 2021. Following this peak, emissions declined moderately toward 2023, dropping to approximately 150 Mt by the reporting year cutoff. Zones from buffer-based and supplier-based approaches showed convergence in results. Scenarios 1a and 1b, which used fixed-radius buffers around slaughterhouses, produced identical gross emissions of 3,938 Mt, as evidenced by completely overlapping lines in Fig. 2 . Scenarios 2a and 2b, which applied spatially explicit sourcing zones based on direct supplier locations as described by Brandão Jr. et al. ( 2023 ), ranged from 3,717 to 4,202 Mt. The close alignment of all scenario lines throughout the time series indicated that the difference between the lowest estimate, Scenario 2a at 3,717 Mt, and the highest, Scenario 2b at 4,202 Mt, was only 13%. This suggests that the method used to define supply zones, whether simple geographic buffers or more detailed spatial autocorrelation techniques, had minimal impact on temporal patterns or total emissions. The difference between buffer-based scenarios (1a and 1b) and supplier-based scenarios (2a and 2b) was even smaller, ranging from 6% when comparing Scenarios 1a or 1b to Scenario 2a, to less than 1% when compared to Scenario 2b. Our assessment of the distribution of deforestation compared with the location of suppliers and slaughterhouses suggests that most direct suppliers are concentrated in regions where deforestation remained active, rather than in areas where forest cover had already been depleted. When emissions were adjusted for temporal discounting and market share (Fig. 2 , lower panel), all estimates dropped substantially, while maintaining distinct but aligned scenario-specific patterns. The discounted emissions panel showed that all scenarios began near zero in 2004 and increased gradually through 2015, remaining below 10 Mt. From 2016 onward, emissions accelerated more rapidly, with visible separation between scenarios. Scenarios 1a and 1b, based on slaughterhouse buffers, showed the highest discounted emissions. Scenario 1a peaked around 23 Mt in 2020 and 2021, while Scenario 1b reached approximately 19 Mt during the same period. Scenarios 2a and 2b, based on direct supplier zones, showed lower peaks, reaching approximately 16 to 17 Mt in 2020 and 2021. These adjustments were applied using linear discount factors based on GHGP guidelines and incorporated actual market share data from GTA slaughter records. In Scenario 1a, gross emissions of 3,938 Mt were reduced to 180 Mt of discounted CO₂e, representing a 95% reduction. The effect of market share adjustment was evident when comparing Scenarios 1a and 1b. Although both had identical gross emissions of 3,938 Mt, applying market share allocation reduced emissions from 180 Mt in Scenario 1a to 128 Mt in Scenario 1b, a 29% reduction. This demonstrates that distributing emissions proportionally based on actual slaughter volumes significantly reduced individual slaughterhouse responsibility. Scenario 2a showed discounted emissions of 117 Mt, representing a 97% reduction from its gross value of 3,717 Mt. Scenario 2b, which began with gross emissions of 4,202 Mt, had discounted emissions of 133 Mt, also reflecting a 97% reduction. Comparing Scenarios 2a and 2b revealed the effect of extending the temporal scope. Moving from a single-year analysis in Scenario 2a to a 20-year analysis in Scenario 2b increased gross emissions by 13% and discounted emissions by 14%, demonstrating proportional consistency. Among all scenarios, Scenario 2a yielded the lowest discounted emissions at 117 Mt, while Scenario 1a yielded the highest at 180 Mt, a difference of 54%. This range reflects the combined influence of spatial definition, temporal scope, and market share allocation. 3.2 Emission estimates from GFW versus PRODES datasets Using GFW data consistently yielded higher emissions estimates than PRODES across all four scenarios, with gross emissions up to 101–102% higher and discounted emissions up to 123–127% higher (Table 3 ). For example, in Scenario 2b (Direct Supplier Zone 2004–23), where PRODES reported 4.2 billion tons, GFW reported approximately 8.4 billion tons of CO₂e, representing 101% more gross emissions and 123% more discounted emissions. This increase reflects core methodological differences between PRODES and GFW strategies to map deforestation. While PRODES detects only clear-cut deforestation requiring full removal of native vegetation using a Brazil-specific algorithm, GFW includes a broader range of forest disturbances such as degradation, selective logging, and certain natural events. GFW also introduced methodological updates after 2015 like the incorporation of Landsat 8 images that may have further widened the gap, unlike PRODES that maintained a consistent methodological approach. Comparing temporal trends between PRODES (Fig. 2 ) and GFW (Fig. 3 ) shows systematic differences in how each dataset detects and quantifies deforestation-driven emissions. Both datasets showed declining trends from 2004 to 2012 (PRODES: ~400 Mt to ~ 100 Mt; GFW: ~550 Mt to ~ 300 Mt, consistently double PRODES values) and stabilization from 2012–2015 (PRODES: 100–150 Mt; GFW: 250–300 Mt) but diverged substantially after 2015. PRODES exhibited a gradual increase from ~ 100 Mt to ~ 280 Mt by 2020–2021 with a smooth trajectory, while GFW showed a sharper, earlier spike reaching ~ 850 Mt around 2016–2017 in Scenario 1a, three times PRODES' value and coinciding with GFW's methodological updates that enabled broader forest disturbance detection. Both declined toward 2023, with PRODES dropping to ~ 150 Mt and GFW falling steeply to 400–450 Mt while maintaining roughly double PRODES values. The discounted emissions (accounting for linear and market share adjustments) showed more contrasts: PRODES displayed gradual increases from near zero to ~ 5 Mt by 2015, accelerating to peaks of 16–23 Mt (varying by scenario) around 2020–2021 in a broad bell-shaped curve, then declining moderately to 10–12 Mt by 2023. GFW discounted emissions (Fig. 3 , lower panel) displayed much more dramatic peaks and variability. From 2004 to 2015, increases were similar to PRODES, rising to ~ 10–15 Mt. However, between 2016 and 2018, we observed sharp spikes that far exceeded any PRODES values: Scenario 1a reached over 55 Mt around 2017, Scenario 1b peaked around 45 Mt, and Scenarios 2a and 2b reached approximately 30–35 Mt. These peaks were roughly 2.5 to 3 times higher than PRODES peaks occurring three to four years later. GFW curves showed sharp, concentrated peaks rather than sustained elevation. After 2018, steep declines dropped to near zero by 2023, while PRODES declined gradually to approximately 5–10 Mt. Table 3 Comparison of PRODES and GFW-based deforestation emissions (2004–2023) across the supply chain scenarios. Values are reported as gross and discounted MtCO₂e, with percentage differences showing how much higher GFW estimates are relative to PRODES. Scenario PRODES (MtCO2e) GFW (MtCO2e) Difference (%) Gross Discounted Gross Discounted Gross Discounted Sc1a: Slaughterhouse Buffer no Market Share 3,937.53 179.73 7,944.09 408.36 102% 127% Sc1b: Slaughterhouse Buffer with Market Share 3,937.53 128.07 7,944.09 287.45 102% 124% Sc2a: Direct Supplier's Zone 2023 3,717.07 116.66 7,520.27 263.77 102% 126% Sc2b: Direct Supplier's Zone 2004-23 4,202.21 133.28 8,457.10 296.78 101% 123% 4. DISCUSSION 4.1. Data availability and the limits of top-down approaches. Corporate climate accounting continues to face substantial data constraints, particularly in sectors characterized by complex supply chains and limited traceability, such as Brazilian cattle production (Zu Ermgassen et al., 2020 ; Garrett et al., 2021). Most companies lack georeferenced supplier information, and none can track animal movements continuously from birth to slaughter. As discussed in the Scope 3 literature, high transaction costs associated with searching, coordinating, and monitoring suppliers often outweigh the perceived benefits of disclosure, unless firms can anticipate a clear competitive advantage (Patchell, 2018 ). Data privacy concerns and fragmented information systems further constrain the collection of farm-level land-use change data (Stenzel and Waichman, 2023 ). Furthermore, many suppliers in general are reluctant to share information due to concerns over confidentiality, competitive exposure, and regulatory risk (Zu Ermgassen et al., 2020 ). As a result, companies remain reliant on proxies rather than precise, farm-level traceability, which weakens incentives for transparency, since firms cannot credibly claim performance above sectoral baselines. In response to these constraints, many companies adopt top-down approaches based on aggregated land-use change statistics. For example, the global emissions analysis tool developed by Fitts et al. (2025), which is operationally simple, but approaches operate at broad spatial scales, typically countries or states, and therefore overlook the geographic heterogeneity of sourcing. This limitation is consequential, because deforestation-related emissions vary substantially across locations and over time. Moreover, global emission factors often fail to reflect local production conditions. For instance, Estevam and Assad (2025) show that IPCC default emission factors tend to exceed Brazilian averages for the cattle sector. Our supply-zone approach directly addresses these limitations by estimating deforestation-related emissions with greater spatial specificity. By linking slaughterhouse operations to surrounding sourcing areas, the framework captures spatial patterns of deforestation more accurately than purely top-down methods. Both simplified buffer-based zones and zones derived from actual supplier locations produced similar emissions estimates in our analysis. This convergence reflects the spatial clustering of deforestation, slaughterhouses, and direct suppliers in accessible regions, as well as the fact that the 200 km buffers were calibrated using empirically observed distances between slaughterhouses and their suppliers. However, accurately allocating emissions also requires accounting for slaughterhouse market share within each supply zone. Without this adjustment, emissions from shared sourcing areas may be fully attributed to a single facility, leading to systematic overestimation. This issue has been highlighted in public assessments, such as a recent investigation by Greenpeace ( 2025 ), which relied on simplified sourcing assumptions that likely overstated company-level deforestation responsibility in the Amazon. In contrast, our framework explicitly incorporates market share, reducing the risk of double counting and misattribution. Our supply-zone framework offers a scalable and relatively low-cost pathway toward more accurate Scope 3 accounting. By narrowing system boundaries from the national to the regional level, it approximates jurisdictional approaches to direct land-use change accounting while remaining compatible with corporate data realities (Fitts et al., 2025). The framework also allows for methodological flexibility depending on data availability. When only slaughterhouse locations and market shares are known, Scenario 1b provides the most appropriate balance by allocating emissions proportionally and avoiding overestimation in shared zones. Scenario 1a, which applies uniform buffers without market share adjustment, tends to overestimate emissions in regions accessible to multiple slaughterhouses. When georeferenced data on direct suppliers are available, Scenario 2a offers higher spatial precision by restricting sourcing areas to active suppliers in the reporting year. Extending sourcing zones to include historical suppliers can artificially inflate sourcing areas and is often infeasible for companies lacking long-term supplier records. 4.2. Implications of using PRODES vs GFW for emissions accounting. The comparison between PRODES and GFW illustrates how methodological choices can substantially influence deforestation-related emissions estimates. Although both datasets capture deforestation in the Brazilian Amazon, across all evaluated scenarios, GFW values were approximately 100% higher than the PRODES results. These discrepancies arise from fundamental differences in scope, detection logic, and classification criteria. PRODES focuses on clear-cut deforestation of native vegetation, identified through expert automatic detection and visual interpretation within a jurisdictionally consistent and legally grounded framework. GFW, in contrast, detects a broader spectrum of forest disturbances, including fire, degradation, and selective logging, using automated global algorithms based on spectral change detection (Sims and Goldman, 2025). The pronounced increase in GFW-based estimates between 2016 and 2018, which is not observed in PRODES, reflects methodological updates that expanded sensitivity to partial canopy loss rather than a sudden shift in clear-cut deforestation dynamics. These methodological differences have direct implications for how companies quantify and report supply-chain emissions (Fitts et al., 2025b ). While GFW offers global coverage and is valuable in contexts where national monitoring systems are weak or inconsistent, PRODES provides more stable, transparent, and legally grounded estimates in Brazil, where deforestation definitions are formally codified within the national monitoring framework. As a result, PRODES is widely used by companies operating in the Brazilian Amazon to assess supplier compliance with zero-illegal-deforestation commitments (Gibbs et al., 2015 ; Raoni et al., 2020). For companies sourcing cattle in Brazil, reliance on PRODES enhances coherence with national regulations and strengthens the credibility of emissions reporting. PRODES also adheres to the principle of conservativeness by prioritizing legal clarity, institutional validation, and methodological consistency over time. Although it does not capture all forms of forest disturbance detected by GFW, this exclusion reflects a deliberate methodological choice aligned with enforcement and regulatory metrics. Companies that apply GFW-based estimates in Brazil without appropriate contextual adjustments therefore risk systematically overestimating deforestation-related emissions. 4.3. Model limitations and recommendations This study introduces a scalable, spatially explicit model to estimate deforestation-related emissions from cattle supply chains, serving as a bridge between statistical land-use change and direct land use methods (Fitts et al., 2025). By linking emissions to supply zones and applying region-specific carbon factors, the approach improves both geographic precision and methodological transparency. Future improvements, however, must address the "double counting" inherent in Scope 3 accounting, where emissions are shared across complex supply networks. As Hertwich & Wood ( 2018 ) note, Scope 3 emissions are essentially the Scope 1 emissions of upstream actors; acknowledging this shared responsibility is vital for identifying leverage points for mitigation rather than simply allocating blame. To refine the allocation of emissions in overlapping supply zones, methodologies such as product expansion or shared responsibility allocation should be applied to distribute burdens equitably based on market share or land footprint (Fitts et al., 2025). Despite its strengths, the model has important limitations. First, it does not incorporate forest regeneration, afforestation, or transitions from pasture to cropland. These land-use dynamics may affect long-term emissions but remain untracked in national datasets like PRODES. Second, the approach does not differentiate primary from secondary forests, which according to Imazon have been deforested at nearly identical rates; between 2019 and 2023, both primary and secondary forests lost slightly over 2 million hectares each (Imazon, 2025). Third, and most critically, supply-zone methods aggregate suppliers to multiple slaughterhouses, which means that it does not differentiate deforestation levels from low and high-performers, or from slaughterhouses with stringent monitoring deforestation protocols or no protocols at all. Additional work is needed to develop approaches that account for company actions to reduce deforestation within their specific supply chain. Spatial heterogeneity introduces further complexity. In Pará, for example, deforestation has moved westward over the past two decades. This means a slaughterhouse operating today in a low-deforestation zone may have contributed to forest loss earlier in the supply chain’s history. When emissions are temporally discounted, older impacts become less visible, even if they remain ecologically relevant. Future improvements should focus on integrating dynamic supply chain data, particularly cattle movement records. This would allow emissions to be allocated based on actual trade flows rather than geographic proxies. For indirect suppliers, such data could reveal laundering patterns and strengthen traceability beyond first-tier sourcing. For direct suppliers, emissions could be reduced by favoring full-cycle ranches with CAR registration and stable animal flows. These enhancements would improve both precision and integrity in emissions accounting. Supply zones are not a substitute for traceability, but they are an actionable step forward. They offer companies a transparent, repeatable method to estimate land-use emissions using currently available data. Until full traceability becomes standard, spatial models like this can help bridge the gap between feasibility and accountability. 5. CONCLUSIONS Our method provides a practical and scalable solution for estimating Scope 3 deforestation emissions in cattle supply chains in Brazil using data that are readily available. This represents a critical step forward for companies seeking to align with climate disclosure frameworks and meet regulatory expectations. As corporate climate targets increasingly require geographically explicit emissions data, our approach enables meatpackers to generate credible estimates without waiting for perfect traceability systems to be in place. Especially, if companies can apply Scenario 1b or Scenario 2a, depending on the data available. However, our results also underscore the limitations of current methodologies and the urgent need for more precise tools to fully capture the extent of land-use change emissions. Our zone-based methods offer feasible proxies in data-constrained contexts, but they remain approximate. While farm to slaughter traceability of cattle introduces higher upfront costs and may require stronger government oversight, it also enables companies to make robust, verifiable zero-deforestation claims. Abbreviations GHG Greenhouse Gas GHGP Greenhouse Gas Protocol CO₂e Carbon Dioxide Equivalent CDP Carbon Disclosure Project EUDR European Union Deforestation Regulation CSRD Corporate Sustainability Reporting Directive SIF Serviço de Inspeção Federal (Federal Inspection Service) SIE Serviço de Inspeção Estadual (State Inspection Service) SIM Serviço de Inspeção Municipal (Municipal Inspection Service) CAR Cadastro Ambiental Rural (Rural Environmental Registry) CPF Cadastro de Pessoas Físicas (Brazilian Individual Taxpayer Registry) CNPJ Cadastro Nacional da Pessoa Jurídica (Brazilian Corporate Taxpayer Registry) GTA Guia de Trânsito Animal (Animal Transit Guide) INPE Instituto Nacional de Pesquisas Espaciais (Brazilian National Institute for Space Research) MAPA Ministério da Agricultura, Pecuária e Abastecimento (Brazilian Ministry of Agriculture, Livestock, and Supply) ADEPARA Agência de Defesa Agropecuária do Estado do Pará (Pará's State Agricultural and Livestock Health Agency) SEMA. Secretaria de Estado de Meio Ambiente e Sustentabilidade (Pará’s Environmental Secretariat) GFW Global Forest Watch GLEAM-i Global Livestock Environmental Assessment Model – interactive ALCI. Agricultural Life Cycle Inventory Generator SBTi Science Based Targets initiative Mt Million metric tons Mha Million hectares tCO₂e/ha Tons of CO₂ equivalent per hectare Declarations Availability of data and materials The data and materials used in this study are available from the corresponding author upon request, subject to data-sharing restrictions. Competing interests Lisa Rausch and Hollly Gibbs have an ongoing consulting relationship with the nonprofit National Wildlife Federation. The National Wildlife Federation did not provide editorial oversight over the contents of the manuscript. Fabio Martins Guerra Nunes Dias is employed by the meat industry, but this association does not compromise the objectivity or integrity of the study’s results. The other authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Funding His work was supported by the Gordon and Betty Moore Foundation. Consent to Publish Consent to Publish declaration: not applicable Ethics and Consent to Participate Ethics and Consent to Participate declarations: not applicable Authors' contributions A.B.J. conceived the study, designed the methodology, analyzed the data, and wrote the original draft. S.C.K. contributed to study design and interpretation and reviewed and revised the manuscript. F.D., J.M., L.R., and S.S.L. provided substantive comments and revisions on the manuscript. H.K.G. supervised the research, and acquired funding. All authors reviewed and approved the final version of the manuscript. Acknowledgements We would like to thank Malena Candido for reviewing different early versions of the manuscript. References Alix-Garcia, J., & Gibbs, H. K. (2017). Forest conservation effects of Brazil's zero deforestation cattle agreements undermined by leakage. Global Environmental Change, 47, 201–217. Barreto, P., Pereira, R., Brandão Jr., A., & Baima, S. (2017). Will meat-packing plants help halt deforestation in the Amazon? Imazon & Instituto Centro de Vida. Brandão Jr., A., Rausch, L., Gibbs, H. K., & West, T. A. (2023). Spatially explicit assessment of cattle ranching and deforestation in the Brazilian Amazon. Environmental Research Letters, 18(4), 044030. CDP. (2020). Missing the Mark: The state of corporate deforestation action. CDP Worldwide. European Commission. (2021). Proposal for a regulation on the making available on the Union market and the export from the Union of certain commodities and products associated with deforestation and forest degradation (COM(2021) 706 final). Fitts, L., James, O., Gibbs, D., Goldman, E., Harris, N., Ramlow, M., Sloat, L., Wilson, C., Winchester, C., Ernstoff, A., & Hohmann, T. (2025a). Geospatial methods for corporate GHG accounting of deforestation and land occupation. World Resources Institute. https://www.wri.org/research/geospatial-methods-corporate-ghg-accounting-deforestation-and-land-occupation Fitts, L., James, O., Gibbs, D., Goldman, E., Harris, N., Ramlow, M., Sloat, L., Wielgosz, B., Wilson, C., Winchester, C., & Ernstoff, A. (2025b). Statistical land use change emissions from deforestation and land occupation for crops. World Resources Institute. https://www.wri.org/research/statistical-land-use-change-emissions-deforestation-and-land-occupation-crops Gibbs, H. K., Rausch, L., Munger, J., Schelly, I., Morton, D. C., Noojipady, P., & Walker, N. F. (2015). Brazil's Soy Moratorium. Science, 347(6220), 377–378. Greenhouse Gas Protocol (GHGP). (2011). Corporate value chain (Scope 3) accounting and reporting standard. World Resources Institute & World Business Council for Sustainable Development. https://ghgprotocol.org/sites/default/files/standards/Corporate-Value-Chain-Accounting-Reporing-Standard_041613_2.pdf Greenhouse Gas Protocol (GHGP). (2013). Technical guidance for calculating Scope 3 emissions. World Resources Institute & World Business Council for Sustainable Development. Greenhouse Gas Protocol (GHGP). (2026). Land Sector and Removals Standard. World Resources Institute & World Business Council for Sustainable Development. https://ghgprotocol.org/land-sector-and-removals-standard Greenpeace. (2025). JBS: Cooking the Planet. Case study: JBS' supply chain linked to cattle raised illegally on Indigenous Land. Greenpeace International. Harris, N. L., Gibbs, D. A., Baccini, A., Birdsey, R. A., De Bruin, S., Farina, M., Fatoyinbo, L., Hansen, M. C., Herold, M., Houghton, R. A., & Potapov, P. V. (2021). Global maps of twenty-first century forest carbon fluxes. Nature Climate Change, 11(3), 234–240. Hertwich, E. G., & Wood, R. (2018). The growing importance of scope 3 greenhouse gas emissions from industry. Environmental Research Letters, 13(10), 104013. Hettler, M., & Graf-Vlachy, L. (2024). Corporate scope 3 carbon emission reporting as an enabler of supply chain decarbonization: A systematic review and comprehensive research agenda. Business Strategy and the Environment, 33(2), 263–282. Júnior, D., & Russo, A. (2025, December 2). Com críticas ao mercado internacional, Pará adia início da obrigatoriedade da rastreabilidade bovina. Estadão Agro. https://agro.estadao.com.br/pecuaria/com-criticas-ao-mercado-internacional-para-adia-obrigatoriedade-de-rastreabilidade-bovina. Panwar, R. (2023). Business and biodiversity: Achieving the 2050 vision for biodiversity conservation through transformative business practices. Biodiversity and Conservation, 32(11), 3607–3613. Patchell, J. (2018). Can the implications of the GHGP's scope 3 standard be realized? Journal of Cleaner Production, 185, 941–958. Rajão, R., Soares-Filho, B., Nunes, F., Börner, J., Machado, L., Assis, D., & Oliveira, A. (2020). The rotten apples of Brazil's agribusiness. Science, 369(6501), 246–248. Rausch, L., Munger, J., & Gibbs, H. K. (2020). Amazon deforestation linked to European imports via specially licensed ranches and their suppliers. Gibbs Land Use and Environment Report Series. https://gibbs-lab.wisc.edu/assets/SISBOV_report_March_2020.pdf. Science Based Targets Initiative (SBTi). (2024). Aligning corporate value chains to global climate goals: Scope 3 discussion paper. Skidmore, M. E., Moffette, F., Rausch, L., & Gibbs, H. K. (2020). Characterizing compliance in cattle supply chains: What factors encourage deforestation-free production in the Brazilian Amazon? Evidensia. https://www.evidensia.eco/resources/937/download/ Skidmore, M. E., Moffette, F., Rausch, L., Christie, M., Munger, J., & Gibbs, H. K. (2021). Cattle ranchers and deforestation in the Brazilian Amazon: Production, location, and policies. Global Environmental Change, 68, 102280. Skidmore, M. E., Sims, K. M., Rausch, L. L., & Gibbs, H. K. (2022). Sustainable intensification in the Brazilian cattle industry: The role for reduced slaughter age. Environmental Research Letters, 17(6), 064026. Skidmore, M. E., Barrozo, M., Rausch, L., & Gibbs, H. K. (2024). Cattle supply chain fidelity in the Brazilian Amazon: Key considerations for deforestation monitoring. Environmental Research Letters, 19(11). Souza Jr., C. M., Shimbo, J. Z., Rosa, M. R., et al. (2020). Reconstructing three decades of land use and land cover changes in Brazilian biomes with Landsat archive and Earth Engine. Remote Sensing, 12(17), 2735. Stenzel, A., & Waichman, I. (2023). Supply-chain data sharing for scope 3 emissions. npj Climate Action, 2(1), 7. Walker, N. F., Patel, S. A., & Kalif, K. A. B. (2013). From Amazon pasture to the high street: Deforestation and the Brazilian cattle product supply chain. Tropical Conservation Science, 6(3), 446–467. Weisse, M., & Potapov, P. (2021). Assessing trends in tree cover loss over 20 years of data. Global Forest Watch. https://www.globalforestwatch.org/blog/data-and-tools/tree-cover-loss-satellite-data-trend-analysis/ West, T. A., Munger, J., & Gibbs, H. K. (2022). Regional variation in zero-deforestation cattle commitments and their potential impact in the Brazilian Amazon. Environmental Research Letters, 17(3), 034014. zu Ermgassen, E. K., Godar, J., Lathuillière, M. J., Löfgren, P., Gardner, T., Vasconcelos, A., & Meyfroidt, P. (2020). The origin, supply chain, and deforestation risk of Brazil's beef exports. Proceedings of the National Academy of Sciences, 117(50), 31770–31779. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Reviewers agreed at journal 11 May, 2026 Reviewers invited by journal 06 Apr, 2026 Editor assigned by journal 26 Mar, 2026 Submission checks completed at journal 26 Mar, 2026 First submitted to journal 16 Mar, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9138992","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":620415246,"identity":"b1fed205-8e0a-4bf3-b7f8-0f9465ca1809","order_by":0,"name":"AMINTAS BRANDAO JR","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAzUlEQVRIiWNgGAWjYBACNvn+B4d5f0jI8UukP/7wg+FfncEBAlr4Jc4YPu7tsTGWnJHYJtnDcCCZoBbJGTnGxmfb0hI33EhsY2ZhOJC4gZAWgxtnzKTPth1O3AzSwkCUlvv9z6RzgVq23U5s/szA8IcYW863wbQ0SBNli/2NnOO/bdv+J26eTawWgxs5adKyQFs2yD9skCbe+7I9h40l7jwEBrLBgWTJBkJa7veYSfP+OCzHPwMUlRX/6vgJ6MAwgUT1o2AUjIJRMAqwAwAyC12C4bzwRwAAAABJRU5ErkJggg==","orcid":"","institution":"University of Wisconsin-Madison","correspondingAuthor":true,"prefix":"","firstName":"AMINTAS","middleName":"","lastName":"BRANDAO","suffix":"JR"},{"id":620415249,"identity":"aa5c1cc4-8062-4dfc-8978-f59815318276","order_by":1,"name":"SARAH Klopatek","email":"","orcid":"","institution":"University of Wisconsin-Madison","correspondingAuthor":false,"prefix":"","firstName":"SARAH","middleName":"","lastName":"Klopatek","suffix":""},{"id":620415251,"identity":"3f5fb1d0-ff6e-4d70-8814-fb08aeb3fb79","order_by":2,"name":"LISA RAUSCH","email":"","orcid":"","institution":"University of Wisconsin-Madison","correspondingAuthor":false,"prefix":"","firstName":"LISA","middleName":"","lastName":"RAUSCH","suffix":""},{"id":620415252,"identity":"44a4532f-3fe9-41f0-9c15-ac843874a920","order_by":3,"name":"FABIO DIAS","email":"","orcid":"","institution":"University of São Paulo","correspondingAuthor":false,"prefix":"","firstName":"FABIO","middleName":"","lastName":"DIAS","suffix":""},{"id":620415255,"identity":"468e31f1-9cfb-44f6-ac49-29de817c5f40","order_by":4,"name":"JACOB MUNGER","email":"","orcid":"","institution":"University of Wisconsin-Madison","correspondingAuthor":false,"prefix":"","firstName":"JACOB","middleName":"","lastName":"MUNGER","suffix":""},{"id":620415256,"identity":"ed2cbde3-c8be-4a05-b0eb-6ae2d289b4ac","order_by":5,"name":"SETH Spawn-Lee","email":"","orcid":"","institution":"The Nature Conservancy: Arlington","correspondingAuthor":false,"prefix":"","firstName":"SETH","middleName":"","lastName":"Spawn-Lee","suffix":""},{"id":620415257,"identity":"281c1e7c-0381-41c2-9010-9a75edf6b384","order_by":6,"name":"HOLLY K. GIBBS","email":"","orcid":"","institution":"University of Wisconsin-Madison","correspondingAuthor":false,"prefix":"","firstName":"HOLLY","middleName":"K.","lastName":"GIBBS","suffix":""}],"badges":[],"createdAt":"2026-03-16 14:23:14","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9138992/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9138992/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":106803810,"identity":"6f385c5f-696d-4795-bad5-7ff5fe0ed793","added_by":"auto","created_at":"2026-04-13 15:03:34","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":200331,"visible":true,"origin":"","legend":"\u003cp\u003eScenarios representing increasing levels of supply chain information for cattle suppliers linked to federal-inspected (SIF) slaughterhouses in Pará.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-9138992/v1/ae7d253939de0286d559b792.png"},{"id":106803808,"identity":"ac78dc55-e193-4f22-998b-aeb660fb20f6","added_by":"auto","created_at":"2026-04-13 15:03:34","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":196438,"visible":true,"origin":"","legend":"\u003cp\u003eAnnual gross and discounted emissions from PRODES-based deforestation across four supply chain scenarios (2004–2023).\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-9138992/v1/0dcd9554859c9d6e2a8fffc7.png"},{"id":106803809,"identity":"aae9da58-b7cd-47f3-a196-b12e595329b7","added_by":"auto","created_at":"2026-04-13 15:03:34","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":191703,"visible":true,"origin":"","legend":"\u003cp\u003eGross and discounted GFW-based deforestation emissions (2004–2023) linked to federally inspected slaughterhouses in Pará under four supply chain scenarios.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-9138992/v1/65634adb6746ed284c51a6dd.png"},{"id":106961434,"identity":"367665d1-48aa-46d1-aae3-a0d1b0145494","added_by":"auto","created_at":"2026-04-15 09:25:33","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1468599,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9138992/v1/3dd2eb1a-06e5-4267-98cd-eec824161ce9.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Calculating Supply Chain Deforestation Emissions for Brazil’s Meatpacking Sector: A New Corporate Accounting Methodology","fulltext":[{"header":"1. BACKGROUND","content":"\u003cp\u003eThe increasing awareness of forest loss associated with agricultural commodities has intensified regulatory oversight of global supply chains. Companies are now under mounting pressure to meet non-deforestation commitments, to ensure the products they source adhere to local laws governing land use, and to comply with laws governing global trade of commodities, such as the European Union Deforestation Regulation (EUDR, 2025). In this context, quantifying the links between supply chains, deforestation, and associated greenhouse gas (GHG) emissions has become an increasingly urgent task for the private sector (Panwar et al., 2023).\u003c/p\u003e \u003cp\u003eThe Greenhouse Gas Protocol (GHGP), developed through a partnership between the World Resources Institute (WRI) and the World Business Council for Sustainable Development (WBCSD), is the most widely adopted framework for quantifying emissions across three scopes: direct emissions from owned or controlled sources (Scope 1), emissions from purchased electricity, heat, and cooling (Scope 2), and other indirect value-chain emissions (Scope 3) including those from land use change (GHGP 2011; GHGP 2013).\u003c/p\u003e \u003cp\u003eIn Brazil, accounting for Scope 3 land-use change emissions linked to cattle production faces significant structural barriers. The supply chain is highly fragmented, with animals often moved across multiple properties before slaughter, complicating efforts to trace land-use impacts. Different slaughterhouses frequently buy from the same groups of suppliers, making it difficult to attribute deforestation to specific facilities. Additionally, limited transparency beyond direct suppliers, those selling cattle to slaughterhouses, enables deforestation linked to indirect suppliers, which sell cattle earlier in the supply chain, to persist, despite zero-deforestation agreements (Walker et al., \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; Gibbs et al., \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Alix-Garcia and Gibbs, \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Raj\u0026atilde;o et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Skidmore et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; West et al., \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Brand\u0026atilde;o Jr. et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Skidmore et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Beyond supply chain complexity, legal and regulatory restrictions on data sharing, the lack of harmonized reporting standards, and concerns over business confidentiality further hinder effective Scope 3 emissions accounting (Stenzel and Waichman, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eProviding meatpacking companies with robust tools to accurately measure and manage emissions is critical. The expansion of cattle pastures in Brazil is linked to more than 80% of all deforestation since 1985 (Souza Jr. et al., \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). In 2024\u0026ndash;2025, Brazil accounted for 19% of the global beef production (USDA, 2025), while deforestation remained the dominant source of emissions\u0026ndash;responsible for 46% of the national greenhouse gas output the previous year, with nearly one-third concentrated in the Amazon (SEEG, 2023).\u003c/p\u003e \u003cp\u003eFull supply chain traceability systems, such as individual ear tags, could help to generate the animal-level emissions data needed to precisely quantify company-specific Scope 3 emissions. Brazil however, still lacks a national farm to slaughter cattle traceability system. Building such a system would require substantial digital and physical infrastructure, clear regulatory frameworks and stable funding in addition to effective enforcement, which means it is likely to take many years to fully materialize. To date, SISBOV which is a voluntary system implemented in 2002 is the only system that has made meaningful progress in this area, yet it covers only 1% of the country\u0026rsquo;s ranches (Rausch et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). At the same time, subnational initiatives offer more immediate pathways. The state of Santa Catarina in southern Brazil has implemented a state-level individual traceability program (Rocci, 2025), though logistical and institutional constraints make scaling to the national level difficult (Cavalcante et al., 2018). Par\u0026aacute;, recently delayed its traceability rollout from 2026 to 2031 citing limited international market interest in individually traceable cattle from the state (J\u0026uacute;nior \u0026amp; Russo, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). In Mato Grosso, the Passaporte Verde Led by IMAC, links socio-environmental traceability to market access, targeting farms with recent illegal deforestation and supporting remediation over a 48-month period (Mato Grosso, 2025).\u003c/p\u003e \u003cp\u003eAt the same time, Brazil\u0026rsquo;s meatpacking sector continues to move towards interim solutions to calculate Scope 3 emissions using available data and guidance. Corporate reporting of land-use emissions relies on the Greenhouse Gas Protocol\u0026rsquo;s Land Sector and Removals Guidance (2023), which outlines three main approaches: statistical land use change (sLUC), direct land use change (dLUC), and jurisdictional direct land use change (jdLUC) (Fitts et al. 2025). The sLUC approach is applied when companies have access to only jurisdiction-level sourcing data. The dLUC approach requires traceability at the level of land management unit, such as a farm or plot. The jdLUC method combines jurisdiction-level sourcing with crop distribution spatial data. Among these, sLUC is the most widely adopted due to its lower data requirements. In contrast, dLUC and jdLUC are rarely implemented, largely due to restricted access to detailed sourcing information and the absence of standardized procedures. Although these methods were developed for broad sectoral application, many target industries still lack the robust empirical foundations needed to apply the (Hettler and Graf-Vlachy 2023).\u003c/p\u003e \u003cp\u003eHere we develop an approach that bridges dLUC and jdLUC, using cattle supply zones to estimate deforestation-linked emissions for slaughterhouses. This approach relies on slaughterhouse geolocation, around which a zone containing a high concentration of suppliers\u0026ndash; the supply zone\u0026ndash; can be estimated (Brand\u0026atilde;o Jr. et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), enabling the assessment of company market share. Previous studies have proposed various methods to delineate these zones, including drawing buffers around slaughterhouse coordinates (Alix-Garcia and Gibbs, \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2017\u003c/span\u003e), estimating maximum purchasing distances (Barreto et al., \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2017\u003c/span\u003e), and using trip length distributions (Santos and Costa, 2018). In this paper, we compare the simplest approach to delineate supply zones and a specific method developed for the Brazilian cattle sector by Brand\u0026atilde;o Jr. et al (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) that delineates zones using spatial proximity between slaughterhouses and locations of its direct suppliers. Analyzing detailed animal movement data, Brand\u0026atilde;o Jr. et al found that over 80% of indirect-to-direct transactions occurred within 100 kilometers of a direct supplier, with a median distance of only 20 kilometers. These patterns allow for the construction of slaughterhouse supply zones based on direct supplier locations. Supply zones capture the geographic footprint of procurement and much of the associated deforestation, especially in landscapes where cattle ranching dominates land use.\u003c/p\u003e \u003cp\u003eHere we designed our model to estimate Scope 3 emissions from deforestation at the slaughterhouse level using spatial proxies for sourcing dynamics. Our goal was to provide companies with a scalable, bottom-up framework for estimating emissions even in data-scarce contexts. We focused on federally inspected meatpacking plants (SIFs), which represent most slaughter operations in the state of Par\u0026aacute;, averaging 70% of the annual cattle slaughter from 2013 to 2023 according to property-level animal movement data (\u003cb\u003eFigure S1\u003c/b\u003e). We focus on Para because of its uniquely complete and accessible data on cattle movements and property boundaries, but similar analyses could be done nation-wide using company locations, direct supplier locations (when available), and estimated median distances between indirect suppliers and direct suppliers.\u003c/p\u003e \u003cp\u003eWe estimated emissions for Par\u0026aacute;\u0026rsquo;s SIF slaughterhouses using two main scenarios to reflect varying levels of corporate information of data access: (1) slaughterhouses locations only, (2) supply zones defined by the spatial footprint of direct suppliers using the method by Brand\u0026atilde;o Jr. et al. (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Applying these scenarios from 2004 to 2023, we assessed how data granularity affected supplier attribution, emissions trajectories, deforestation baselines, and market share calculations. Unlike conventional sLUC approaches that apply uniform deforestation rates, our method incorporates spatial and temporal variation, aligning emissions estimates more closely with actual land-use dynamics and procurement behavior (Schmidt et al., 2022; CDP, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). This work contributes to the transition encouraged by the GHGP and the Science Based Targets initiative toward geographically explicit, supplier-linked accounting (GHGP, 2022; SBTi, 2024). As companies face increasing pressure and regulatory mandates to disclose supply-chain Scope 3 emissions, such as under the EU Corporate Sustainability Reporting Directive (CSRD, 2025), our supply zone approach offers a practical and scientifically grounded solution. Its implementation represents a significant advancement in improving the accuracy and transparency of GHG inventories for corporations sourcing beef from Brazil.\u003c/p\u003e"},{"header":"2. METHODS","content":"\u003cp\u003eWe developed a spatially explicit framework to estimate deforestation-related greenhouse gas emissions attributable to individual slaughterhouses in Brazil, consistent with Scope 3 guidance under the Greenhouse Gas Protocol. Drawing on two decades of deforestation data from 2004 to 2023, combined with shorter time series on cattle movements, rural property boundaries, and spatially explicit carbon stocks, we implemented a five-step analytical process. This process included delineation of slaughterhouse supply zones; identification of pasture-driven deforestation within each zone; quantification of associated carbon emissions; application of a 20-year temporal discounting factor to weigh historical emissions; and allocation of emissions to slaughterhouses based on their market share within each supply zone.\u003c/p\u003e\n\u003cp\u003eWe evaluated the framework under three main scenarios representing increasing levels of supply chain resolution, ranging from simple geographic buffers to supply zones derived from the geolocation of direct suppliers, defined as farms delivering cattle to slaughterhouses. All emissions were standardized to 2023 as the reporting baseline. The framework was designed to be transparent, replicable, and compatible with corporate GHG reporting systems using commonly available supply chain data. Emissions associated with post-deforestation land management practices were excluded from the analysis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2. 1 Pre-processing\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo build our model, we integrated property-level cattle movement records, deforestation maps, slaughterhouse locations, property boundaries, and spatially explicit carbon emission factors (\u003cstrong\u003eTable 1\u003c/strong\u003e). The dataset spans 2004 to 2023, with traceability of cattle movements available from 2013 onward. The model was designed to align with GHGP guidance and to be replicable by companies seeking to estimate Scope 3 land-use change emissions in data-constrained contexts. Below, we describe the data sources and preprocessing steps used to construct the analytical framework.\u003c/p\u003e\n\u003cul\u003e\n \u003cli\u003e\u003cu\u003eCattle movement data between properties and to slaughterhouses.\u003c/u\u003e We used data from the Animal Transit Records (Guia de Tr\u0026acirc;nsito Animal in Portuguese, GTA) system maintained by the Par\u0026aacute; State Agency for Agricultural and Livestock Health (ADEPARA), which records all cattle movements between farms and to establishments such as slaughterhouses in the state of Par\u0026aacute;. Mandatory under Brazilian law, the GTA records serve as the primary mechanism for livestock traceability and animal health surveillance. Our dataset covers the period from 2013 to 2023 and includes records for more than 204 million head of cattle transported across the state. Each record contains standardized information on the purpose of the movement, including slaughter, fattening, breeding, auction, or other uses; the number of animals; sex classification; age group; and unique identifiers for both origin and destination properties, derived from the landowner\u0026rsquo;s individual (Cadastro de Pessoas F\u0026iacute;sicas, CPF) or corporate (Cadastro Nacional da Pessoa Jur\u0026iacute;dica, CNPJ) registration. Records also report on the date of movement and the municipalities of origin and destination. Using these data, we identified direct suppliers, defined as farms that sold cattle directly to slaughterhouses for immediate processing. Direct suppliers accounted for approximately 29 million head of cattle over the study period, and their spatial distribution was used as a proxy to delineate slaughterhouse supply zones.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003e\u003cu\u003eProperty boundaries.\u003c/u\u003e We obtained the Rural Environmental Registry (Cadastro Ambiental Rural, CAR) dataset through December 2023 from the Par\u0026aacute; State Environmental Secretariat (SEMAS). CAR is a mandatory national registry requiring all rural properties in Brazil to report their boundaries and landowner information. The dataset includes geo-referenced property boundaries represented as polygons, together with landowner identifiers recorded as CPF for individuals and CNPJ for companies. Following the approaches described by Skidmore et al. (2020, 2021, 2022), West et al. (2022), and Brand\u0026atilde;o Jr. et al. (2023), we matched cattle movement records from the GTA to CAR properties using landowner identifiers and complementary attributes, including municipality and property name. This matching process enabled the spatial linkage of cattle movements to their corresponding rural properties. The resulting dataset identifies and geo-references the properties involved in cattle transactions, hereafter referred to as GTA properties.\u003c/li\u003e\n \u003cli\u003e\u003cu\u003eFederal slaughterhouses.\u003c/u\u003e We obtained the official registry of 22 companies operating federally inspected slaughterhouses (Servi\u0026ccedil;o de Inspe\u0026ccedil;\u0026atilde;o Federal, SIF) in Par\u0026aacute; from the Brazilian Ministry of Agriculture, Livestock, and Supply (MAPA). SIF facilities are authorized to process meat for interstate and international trade and are required to comply with federal sanitary inspection standards. The registry provides company names, physical addresses, and unique SIF registration codes for each facility. We geocoded these addresses to obtain precise geographic coordinates (latitude and longitude) using the Google Maps API, with manual verification for locations that could not be automatically resolved. These facilities processed approximately 70% of all cattle slaughtered in the state, with an average of 1,861,546 head per year (Figure S1). The remaining 30% were slaughtered at state-inspected (SIE) and municipally inspected (SIM) facilities that serve local markets. To protect commercial confidentiality, all SIF slaughterhouses were aggregated into a single analytical category rather than reported individually.\u003c/li\u003e\n \u003cli\u003e\u003cu\u003eDeforestation.\u003c/u\u003e We used annual deforestation maps from PRODES, produced by Brazil\u0026rsquo;s National Institute for Space Research (INPE), based on 30 \u0026times; 30 m Landsat imagery. The PRODES Brazil dataset includes cumulative deforestation in primary forests through 2007, annual deforestation from 2008 to 2023, forest and non-forest vegetation classes, and areas covered by water or cloud. To extend temporal coverage to 2004\u0026ndash;2006, we incorporated a vector product released by INPE in 2016 for the Brazilian Legal Amazon, which provides annual deforestation maps from 2001 onward. From this product, we generated a mask of cumulative deforestation through 2003 and converted the 2004\u0026ndash;2006 annual deforestation layers from a vector to a raster format to match the resolution and spatial alignment of the 2023 PRODES Brazil raster. To ensure consistency across the time series, we excluded all cells outside the PRODES deforestation extent through 2007. The resulting dataset is a continuous raster representing cumulative deforestation through 2003 and annual deforestation from 2004 to 2023. For comparison, we also used the Global Forest Watch dataset (Global Forest Watch; Hansen et al., 2013), which provides annual tree cover loss at 30 m resolution for 2004\u0026ndash;2023 based on automated time-series analysis of Landsat imagery. Whereas PRODES relies on expert visual interpretation to identify clear-cut deforestation in primary forest, Global Forest Watch applies spectral change detection to quantify all forms of tree cover loss globally, irrespective of forest type or disturbance driver. This comparison allowed us to assess the consistency and robustness of deforestation estimates across monitoring systems with distinct methodological approaches. To estimate the share of deforestation attributable to pasture expansion, we used land cover transition data from MapBiomas Collection 10 (MapBiomas) as a proxy. Specifically, we calculated the proportion of transitions from forest (class 1) to pasture (class 15) relative to transitions from forest to farming (class 14) and applied this ratio to the total deforested area within each supply zone to estimate the contribution of pasture expansion to overall deforestation.\u003c/li\u003e\n \u003cli\u003e\u003cu\u003eCarbon emission factors\u003c/u\u003e. We incorporated spatially explicit carbon emission factor maps from Harris et al. (2021). These maps provide annual estimates of aboveground biomass carbon, belowground biomass carbon, and soil organic carbon stocks in a harmonized time series from 2004 to 2023, with all components mapped at 30-meter resolution. Each carbon stock layer was reprojected and spatially aligned with the PRODES deforestation maps to enable pixel-level calculations of total carbon density. To calculate emission factors, we summed the three carbon pools and converted carbon to carbon dioxide equivalent using the standard molecular weight ratio of 44 to 12, representing the mass of CO₂ relative to elemental carbon. Emission factors are expressed in metric tons of carbon dioxide equivalent per hectare (tCO₂e/ha). For example, a pixel with 150 tC/ha in aboveground biomass, 30 tC/ha in belowground biomass, and 50 tC/ha in soil organic carbon yields a total carbon density of 230 tC/ha, equivalent to 843.3 tCO₂e/ha (230 \u0026times; 44 \u0026divide; 12). This spatially explicit approach allowed us to estimate carbon emissions from deforestation events based on the specific carbon stocks present at each location and year, accounting for spatial heterogeneity in forest carbon density across Par\u0026aacute;.\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003e\u003cstrong\u003eTable 1.\u003c/strong\u003e Main datasets used in this study.\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"564\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 189px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eData\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 186px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePeriod\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 189px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSource\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 189px;\"\u003e\n \u003cp\u003eSlaughterhouse location\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 186px;\"\u003e\n \u003cp\u003e2013-2023\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 189px;\"\u003e\n \u003cp\u003eMAPA\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 189px;\"\u003e\n \u003cp\u003eCattle transaction records\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 186px;\"\u003e\n \u003cp\u003e2013-2023\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 189px;\"\u003e\n \u003cp\u003ePar\u0026aacute;\u0026rsquo;s State Sanitation Agency\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 189px;\"\u003e\n \u003cp\u003eProperty maps\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 186px;\"\u003e\n \u003cp\u003e2023\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 189px;\"\u003e\n \u003cp\u003ePara\u0026rsquo;s State Environmental Agency\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 189px;\"\u003e\n \u003cp\u003eAnnual Deforestation maps\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 186px;\"\u003e\n \u003cp\u003e2004-2023\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 189px;\"\u003e\n \u003cp\u003eINPE\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 189px;\"\u003e\n \u003cp\u003eEmission factors (tCO2e/ha/year)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 186px;\"\u003e\n \u003cp\u003e2004-2023\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 189px;\"\u003e\n \u003cp\u003eGlobal Forest Watch (GFW) based on Harris et al. (2021)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003e2.2 Estimating Baseline Emissions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eOur five-step approach to estimate deforestation and GHG emissions linked to cattle supply chains (\u003cstrong\u003eTable 2\u003c/strong\u003e) accounts for different levels of supply chain data available for each slaughterhouse.\u003c/p\u003e\n\u003col start=\"1\" type=\"1\"\u003e\n \u003cli\u003eDefine spatial boundary: We delineated the spatial extent of supply zones for SIFs based on the available data for each state. These zones were defined either as generalized geographic buffers around slaughterhouse locations or as areas surrounding direct suppliers.\u003c/li\u003e\n \u003cli\u003eCalculate deforestation: We quantified deforestation within each supply zone using PRODES data as the primary source. Additionally, we conducted a comparative analysis with GFW data to assess consistency across monitoring approaches.\u003c/li\u003e\n \u003cli\u003eCalculate GHG emissions: We converted deforestation into GHG emissions using a committed emissions approach. This calculation incorporated pixel-level carbon stock values from aboveground, belowground, and soil carbon pools, based on the model by Harris et al. (2021).\u003c/li\u003e\n \u003cli\u003eApply linear discounting: We applied a linear discounting factor to distribute emissions over a 20-year period, in accordance with GHGP guidelines for land-use change emissions reporting.\u003c/li\u003e\n \u003cli\u003eApply market share: We allocated emissions to individual slaughterhouses using either market share proxies or direct traceability data, depending on the scenario and the level of supply chain resolution available.\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003eWe applied the five-step approach under two main scenarios, with each scenario representing a distinct level of data availability and spatial precision (\u003cstrong\u003eTable 2\u003c/strong\u003e):\u003c/p\u003e\n\u003cul\u003e\n \u003cli\u003eScenario 1: Buffer-Based Approach. This scenario applies when the geographic locations of slaughterhouses are the only available information, without access to supplier-level data. Supply zones were therefore defined using fixed 200 kilometers radius buffers around each facility (based on insights from Brandao Jr et al., 2023), and all deforestation emissions within these buffers were attributed to the corresponding slaughterhouses. We tested two variations:\u003cul\u003e\n \u003cli\u003e1a: Emissions were distributed equally among all slaughterhouses operating within each buffer zone, without market share adjustment.\u003c/li\u003e\n \u003cli\u003e1b: Emissions were allocated proportionally to the market share of SIFs within each supply zone. Market share was calculated as the ratio of cattle slaughtered by SIF slaughterhouses within each supply zone to the total number of cattle slaughtered by all slaughterhouses operating in that zone, including those under federal, state, and municipal inspection.\u003c/li\u003e\n \u003c/ul\u003e\n \u003c/li\u003e\n \u003cli\u003eScenario 2: Supplier-Based Approach. This scenario incorporates supplier data to construct spatially explicit supply zones, following the methodology of Brand\u0026atilde;o Jr. et al. (2023), which uses spatial autocorrelation to identify contiguous sourcing regions based on direct supplier concentrations. Two temporal variants were considered:\u003cul\u003e\n \u003cli\u003eScenario 2a: Based on suppliers active in 2023, representing the most recent supply chain configuration. As in Scenario 1b, emissions were allocated proportionally based on market share.\u003c/li\u003e\n \u003cli\u003eScenario 2b: Includes all suppliers active between 2013 and 2023, capturing a full decade of sourcing dynamics. As in Scenario 1b, emissions were allocated proportionally based on market share.\u003c/li\u003e\n \u003c/ul\u003e\n \u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2.\u003c/strong\u003e Five-step methodology to estimate deforestation and GHG emissions linked to cattle supply chains in Par\u0026aacute; applied under distinct scenarios depending on the supply chain data available.\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"623\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eStep\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 261px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eScenario 1: Buffer-Based Approach\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 260px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eScenario 2: Supplier-Based Approach\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp\u003e1a\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 129px;\"\u003e\n \u003cp\u003e1b\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 130px;\"\u003e\n \u003cp\u003e2a\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 130px;\"\u003e\n \u003cp\u003e2b\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eStep 1: Define spatial boundary\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp\u003e200 km buffer\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 129px;\"\u003e\n \u003cp\u003e200 km buffer\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 130px;\"\u003e\n \u003cp\u003eDirect Suppliers (2023) as proxy\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 130px;\"\u003e\n \u003cp\u003eDirect Suppliers (2013\u0026ndash;2023) as proxy\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eStep 2: Calculate deforestation.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp\u003ePRODES (GFW for comparison)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 129px;\"\u003e\n \u003cp\u003ePRODES (GFW for comparison)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 130px;\"\u003e\n \u003cp\u003ePRODES (GFW for comparison)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 130px;\"\u003e\n \u003cp\u003ePRODES (GFW for comparison)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eStep 3: Calculate GHG emissions.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp\u003ePixel-level carbon stocks\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 129px;\"\u003e\n \u003cp\u003ePixel-level carbon stocks\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 130px;\"\u003e\n \u003cp\u003ePixel-level carbon stocks\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 130px;\"\u003e\n \u003cp\u003ePixel-level carbon stocks\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eStep 4: Apply linear discounting\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp\u003e20-year linear discounting\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 129px;\"\u003e\n \u003cp\u003e20-year linear discounting\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 130px;\"\u003e\n \u003cp\u003e20-year linear discounting\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 130px;\"\u003e\n \u003cp\u003e20-year linear discounting\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eStep 5: Apply market share\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp\u003eNo market share\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 129px;\"\u003e\n \u003cp\u003eMarket share\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 130px;\"\u003e\n \u003cp\u003eMarket share\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 130px;\"\u003e\n \u003cp\u003eMarket share\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003eStep 1: Define spatial boundary\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe defined spatial boundaries around each slaughterhouse to identify areas with deforestation. The approach varied based on available data.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eScenario 1: Buffer-Based Approach\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWhen supplier data were unavailable, we created a 200 km radius circles around each slaughterhouse. This distance reflects typical cattle transport distances in Par\u0026aacute; (Brand\u0026atilde;o Jr. et al., 2023). All deforestation within these circles was considered potentially linked to cattle supply chains.\u003c/p\u003e\n\u003cul\u003e\n \u003cli\u003e1a: Slaughterhouse Buffer No Market Share. Assigns 100% of emissions within the buffer to that slaughterhouse after the linear discount.\u003c/li\u003e\n \u003cli\u003e1b: Slaughterhouse Buffer with Market Share. Distributes emissions proportionally among all slaughterhouses operating in the buffer based on their relative cattle slaughter volumes (market share) after the linear discount.\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003e\u003cstrong\u003eScenario 2: Supplier-Based Approach\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWhen supplier data were available from GTA traceability records, we identified the actual properties that sold cattle to each slaughterhouse. We then connected these supplier locations into contiguous zones using spatial autocorrelation (Brand\u0026atilde;o Jr. et al., 2023), creating irregular boundaries that follow real cattle sourcing patterns rather than simple geometric circles (Figure 1).\u003c/p\u003e\n\u003cul\u003e\n \u003cli\u003e2a: Direct Supplier\u0026apos;s Zone 2023. This analysis includes only suppliers active in 2023, providing a snapshot of the current supply chain geography while applying both the linear discount and market share assumption.\u003c/li\u003e\n \u003cli\u003e2a: Direct Supplier\u0026apos;s Zone 2004-23. This approach includes all suppliers active between 2004 and 2023, generating 20 annual supply zone maps. These maps were overlaid, and each area was labeled according to the most recent year of supplier activity. Because comprehensive GTA data only became available starting in 2013, we assumed that the 2013 supply zone configuration remained constant for the earlier period from 2004 to 2012. The analysis applies both the linear discount and market share assumptions throughout.\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003e\u003cstrong\u003eStep 2: Calculate deforestation.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAnnual deforestation was quantified by intersecting PRODES data with each slaughterhouse\u0026rsquo;s supply zone. For each year between 2004 and 2023, we calculated the deforested area within supply zones under each scenario. All spatial operations were performed using a projected coordinate system appropriate for area calculations (EPSG:5880, Brazil Polyconic).\u003c/p\u003e\n\u003cp\u003eWe also estimated deforestation using data from GFW to enable comparison with an alternative monitoring system commonly used by companies in emissions reporting (Lucia et al., 2025). While PRODES maps clear-cut deforestation defined as the complete removal of native forest cover relative to a 1988 baseline, GFW captures a broader spectrum of canopy loss. This includes both full clearing and partial degradation, in which forest cover is reduced but not entirely removed. Such losses may result from logging, fire, seasonal dynamics, or natural disturbances and reflect intermediate transitions between forested and non-forested conditions.\u003c/p\u003e\n\u003cp\u003eThe two systems also differ in methodological consistency over time. PRODES has applied a stable and continuous methodology since 2008, enabling reliable year-to-year comparisons of deforestation trends (INPE, 2025). In contrast, GFW implemented a major methodological update in 2015 (Weisse and Potapov, 2021), which may affect the comparability of long-term trend analyses.\u003c/p\u003e\n\u003cp\u003eIn both deforestation estimates, we applied a proportional adjustment to isolate the share attributable to cattle pasture expansion. This factor was derived from MapBiomas land-cover transition data by calculating the proportion of transitions from forest to pasture relative to total transitions from forest to agricultural land and applying this ratio to the total deforested area within each supply zone.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStep 3: Calculate GHG emissions.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe estimated annual land-use change emissions by intersecting deforestation data with spatially explicit emission factor maps from Harris et al. (2021). Emissions were calculated as the product of deforested areas (activity data) and emission factors, which quantify GHGs released per hectare (GHGP, 2013). Adopting a committed emissions approach, we assumed that the full carbon stock of each deforested pixel was released in the year of clearing. This method aligns with current recommendations for gross emissions accounting and avoids uncertain assumptions about post-clearing carbon dynamics.\u003c/p\u003e\n\u003cp\u003eEmission factors from Harris et al. (2021) incorporate aboveground and belowground biomass and soil organic carbon, expressed in metric tons of CO₂ equivalent per hectare per year for 2002\u0026ndash;2023. To ensure spatial consistency, we reprojected and resampled these layers to match the resolution and coordinate system of PRODES and GFW data.\u003c/p\u003e\n\u003cp\u003eFor each slaughterhouse and year (2004\u0026ndash;2023), we overlaid annual deforestation maps with the corresponding emission factor layer. Emissions were calculated by multiplying the area of each deforested pixel by its specific carbon value and summing results within each sourcing zone. This procedure was applied across all sourcing scenarios, generating an annual time series of gross emissions per plant at varying levels of supply chain resolution. We did not account for forest regrowth, carbon removals, or post-clearing land use due to limited data availability. All spatial analyses were conducted using the Google Earth Engine (GEE) platform.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStep 4: Apply linear discounting\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe applied a 20-year linear discounting function to the gross emissions calculated in Step 3, following GHGP recommendations. Each deforestation event was weighted based on its age relative to the reporting year, with a 0.5% annual reduction. For instance, emissions from 2023 were counted at 9.75% and from 2004 at 0.25%. Events prior to 2004 were excluded from the inventory. As an example, a 2010 deforestation event generating 1000 tCO₂e in gross emissions would contribute only 32.5 tCO₂e (3.25% of 1000 tCO2e) to the reported emissions for 2023. This discount was applied uniformly across all \u0026nbsp;scenarios.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStep 5: Apply market share\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn the final step, we allocated the discounted emissions to individual slaughterhouses according to the structure defined in each scenario. While the total emissions from deforestation remained constant within each sourcing zone, the method of allocation varied depending on data availability and supply chain resolution. For the period from 2013 to 2023, we used actual market shares calculated from GTA slaughter records. For 2004 to 2012, when comprehensive GTA data were not available, we assumed market shares remained constant at their 2013 values (Figure S4).\u003c/p\u003e\n\u003cul\u003e\n \u003cli\u003eScenario 1a: Slaughterhouse Buffer No Market Share. In the absence of supplier-level data, we attributed 100% of emissions within each buffer zone to the SIF slaughterhouse located in that zone, assuming full responsibility for all deforestation after applying the linear discount. For example, if a buffer zone generated 32.5 tCO₂e of discounted emissions in 2010, the entire amount would be allocated to that SIF facility. This scenario represents the simplest allocation method, where the SIF slaughterhouses were assigned all emissions within its 200 km radius buffer. \u0026nbsp; \u0026nbsp;\u003c/li\u003e\n \u003cli\u003eScenario 1b: Slaughterhouse Buffer with Market Share. Using the same 200 km buffer zones as Scenario 1a, we applied market share adjustments to allocate emissions proportionally among all slaughterhouses operating within each zone. Using the same example, if SIF slaughterhouses accounted for 67% of total cattle slaughter in that zone in 2010, we would allocate 67% of 32.5 tCO₂e, or 21.8 tCO₂e, to the SIF facility. The remaining 33% (10.7 tCO₂e) would be attributed to non-SIF slaughterhouses operating in the same buffer zone. This approach recognizes that multiple facilities may source cattle from overlapping geographic areas.\u003c/li\u003e\n \u003cli\u003eScenario 2a Direct Supplier\u0026apos;s Zone 2023 and Scenario 2b Direct Supplier\u0026apos;s Zone 2004-23. When direct supplier data were available to define spatially explicit sourcing zones, we calculated each SIF slaughterhouse\u0026apos;s proportional share of total cattle slaughter within its supply zone for each year. This proportion was then used to allocate emissions among all slaughterhouses operating in that zone. The calculation method was identical to Scenario 1b: if SIF market share was 67% in 2010, we allocated 67% of emissions (21.8 tCO₂e from the 32.5 tCO₂e total) to SIF facilities. The key difference from Sc1 is that supply zones were defined by actual supplier distributions rather than fixed 200 km buffers, resulting in more spatially accurate zones. Scenario 2a used only suppliers active in 2023, while Scenario 2b incorporated all suppliers from 2004 to 2023, capturing temporal dynamics in sourcing patterns.\u003c/li\u003e\n\u003c/ul\u003e"},{"header":"3. RESULTS","content":"\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Spatial and temporal patterns of SIF supply zones\u003c/h2\u003e \u003cp\u003eThe total spatial extent of supply zones varied substantially across methods, ranging from 43.7 Mha in Scenario 2a to 74.9 Mha in Scenarios 1a and 1b (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e; Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Scenarios 1a and 1b, which relied on uniform 200 km radius buffers around each slaughterhouse, produced the largest estimated supply zones, totaling 74.9 Mha. In contrast, Scenario 2a delineated supply zones using spatially explicit sourcing areas based exclusively on direct suppliers active in 2023, resulting in a total area of 43.7 Mha, a 42% reduction relative to buffer-based zones. Expanding the temporal scope, Scenario 2b incorporated all direct suppliers active between 2013 and 2023, increasing the total supply zone area to 60.1 Mha, a 38% increase relative to Scenario 2a. Despite this expansion, Scenario 2b remained 20% smaller than the buffer-based zones.\u003c/p\u003e \u003cp\u003eAlthough buffer-based zones were substantially larger in total area than supplier-based zones, the extent of pasture included within each zone was remarkably similar. In Scenarios 1a and 1b, pasture covered 20.7 Mha, corresponding to 28% of the total zone area, while natural vegetation accounted for 49.9 Mha (67%). Despite its smaller overall extent, Scenario 2a captured a nearly equivalent pasture area of 20.1 Mha, representing 46% of the zone, while natural vegetation declined sharply to 21.2 Mha (49%), a 57% reduction relative to the buffer-based estimate. In Scenario 2b, pasture expanded modestly to 21.8 Mha (36%), whereas natural vegetation increased to 34.6 Mha (58%). Relative to Scenario 2a, this corresponds to an 8% increase in pasture area and a 63% increase in natural vegetation.\u003c/p\u003e \u003cp\u003eTogether, these results suggest that supplier-based zones more precisely capture active cattle production areas while excluding distant forest regions with limited connection to current slaughterhouse supply chains. At the same time, they indicate that buffer-based zones encompass pasture areas comparable to those identified through supplier-based approaches, and that much of the additional area captured under the multi-year supplier scenario corresponds to forested areas, rather than current supplier activity.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eEstimated supply zone areas for SIF slaughterhouses by scenario\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eScenario\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSupplier Type\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePeriod\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eDelimitation Criterion\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eEstimated Zone Area (Mha)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003ePasture 2023 Area (Mha)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eNatural Vegetation 2023 (Mha)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1a and 1b\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAll (Direct and Indirect)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eFixed 200 km buffer\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e74.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e20.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e49.9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2a\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDirect\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eBrandao Jr et al. (2023) method to define the zones based on direct suppliers\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e43.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e20.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e21.2\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2b\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDirect\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2013\u0026ndash;2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e60.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e21.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e34.6\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e3.1 PRODES-based emissions\u003c/h2\u003e \u003cp\u003eBuffer zones centered on slaughterhouses (Scenarios 1a and 1b) and zones defined by the spatial distribution of direct suppliers (Scenarios 2a and 2b) yielded total gross emissions exceeding 3.7\u0026nbsp;billion metric tons of CO₂ equivalent between 2004 and 2023 (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, upper panel). Scenario 2b, which included all identified direct suppliers over the full 20-year period, produced the highest estimate at 4,202 Mt. Scenario 2a, focusing only on 2023 direct suppliers, yielded 3.7\u0026nbsp;billion metric tons of CO₂, representing a 12% reduction compared to Scenario 2b. Scenarios 1a and 1b, both using slaughterhouse buffers, produced identical gross emissions of 3.9\u0026nbsp;billion metric tons of CO₂.\u003c/p\u003e \u003cp\u003eThe temporal trends in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e (upper panel) showed that all four scenarios followed similar trajectories throughout the 2004 to 2023 period. The lines were nearly overlapping, indicating consistent patterns across different spatial definitions. Gross emissions declined sharply from 2004 to 2012, dropping from approximately 400 Mt to around 100 Mt, reflecting the well-documented reduction in Amazon deforestation during this period. From 2012 to 2015, emissions stabilized at relatively low levels around 100 Mt. After 2015, a gradual increase began, accelerating notably from 2018 onward and reaching peak values of around 280 Mt in 2020 and 2021. Following this peak, emissions declined moderately toward 2023, dropping to approximately 150 Mt by the reporting year cutoff.\u003c/p\u003e \u003cp\u003eZones from buffer-based and supplier-based approaches showed convergence in results. Scenarios 1a and 1b, which used fixed-radius buffers around slaughterhouses, produced identical gross emissions of 3,938 Mt, as evidenced by completely overlapping lines in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. Scenarios 2a and 2b, which applied spatially explicit sourcing zones based on direct supplier locations as described by Brand\u0026atilde;o Jr. et al. (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), ranged from 3,717 to 4,202 Mt. The close alignment of all scenario lines throughout the time series indicated that the difference between the lowest estimate, Scenario 2a at 3,717 Mt, and the highest, Scenario 2b at 4,202 Mt, was only 13%. This suggests that the method used to define supply zones, whether simple geographic buffers or more detailed spatial autocorrelation techniques, had minimal impact on temporal patterns or total emissions. The difference between buffer-based scenarios (1a and 1b) and supplier-based scenarios (2a and 2b) was even smaller, ranging from 6% when comparing Scenarios 1a or 1b to Scenario 2a, to less than 1% when compared to Scenario 2b. Our assessment of the distribution of deforestation compared with the location of suppliers and slaughterhouses suggests that most direct suppliers are concentrated in regions where deforestation remained active, rather than in areas where forest cover had already been depleted.\u003c/p\u003e \u003cp\u003eWhen emissions were adjusted for temporal discounting and market share (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, lower panel), all estimates dropped substantially, while maintaining distinct but aligned scenario-specific patterns. The discounted emissions panel showed that all scenarios began near zero in 2004 and increased gradually through 2015, remaining below 10 Mt. From 2016 onward, emissions accelerated more rapidly, with visible separation between scenarios. Scenarios 1a and 1b, based on slaughterhouse buffers, showed the highest discounted emissions. Scenario 1a peaked around 23 Mt in 2020 and 2021, while Scenario 1b reached approximately 19 Mt during the same period. Scenarios 2a and 2b, based on direct supplier zones, showed lower peaks, reaching approximately 16 to 17 Mt in 2020 and 2021.\u003c/p\u003e \u003cp\u003eThese adjustments were applied using linear discount factors based on GHGP guidelines and incorporated actual market share data from GTA slaughter records. In Scenario 1a, gross emissions of 3,938 Mt were reduced to 180 Mt of discounted CO₂e, representing a 95% reduction. The effect of market share adjustment was evident when comparing Scenarios 1a and 1b. Although both had identical gross emissions of 3,938 Mt, applying market share allocation reduced emissions from 180 Mt in Scenario 1a to 128 Mt in Scenario 1b, a 29% reduction. This demonstrates that distributing emissions proportionally based on actual slaughter volumes significantly reduced individual slaughterhouse responsibility.\u003c/p\u003e \u003cp\u003eScenario 2a showed discounted emissions of 117 Mt, representing a 97% reduction from its gross value of 3,717 Mt. Scenario 2b, which began with gross emissions of 4,202 Mt, had discounted emissions of 133 Mt, also reflecting a 97% reduction. Comparing Scenarios 2a and 2b revealed the effect of extending the temporal scope. Moving from a single-year analysis in Scenario 2a to a 20-year analysis in Scenario 2b increased gross emissions by 13% and discounted emissions by 14%, demonstrating proportional consistency. Among all scenarios, Scenario 2a yielded the lowest discounted emissions at 117 Mt, while Scenario 1a yielded the highest at 180 Mt, a difference of 54%. This range reflects the combined influence of spatial definition, temporal scope, and market share allocation.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Emission estimates from GFW versus PRODES datasets\u003c/h2\u003e \u003cp\u003eUsing GFW data consistently yielded higher emissions estimates than PRODES across all four scenarios, with gross emissions up to 101\u0026ndash;102% higher and discounted emissions up to 123\u0026ndash;127% higher (Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e3\u003c/span\u003e). For example, in Scenario 2b (Direct Supplier Zone 2004\u0026ndash;23), where PRODES reported 4.2\u0026nbsp;billion tons, GFW reported approximately 8.4\u0026nbsp;billion tons of CO₂e, representing 101% more gross emissions and 123% more discounted emissions. This increase reflects core methodological differences between PRODES and GFW strategies to map deforestation. While PRODES detects only clear-cut deforestation requiring full removal of native vegetation using a Brazil-specific algorithm, GFW includes a broader range of forest disturbances such as degradation, selective logging, and certain natural events. GFW also introduced methodological updates after 2015 like the incorporation of Landsat 8 images that may have further widened the gap, unlike PRODES that maintained a consistent methodological approach.\u003c/p\u003e \u003cp\u003eComparing temporal trends between PRODES (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e) and GFW (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e) shows systematic differences in how each dataset detects and quantifies deforestation-driven emissions. Both datasets showed declining trends from 2004 to 2012 (PRODES: ~400 Mt to ~\u0026thinsp;100 Mt; GFW: ~550 Mt to ~\u0026thinsp;300 Mt, consistently double PRODES values) and stabilization from 2012\u0026ndash;2015 (PRODES: 100\u0026ndash;150 Mt; GFW: 250\u0026ndash;300 Mt) but diverged substantially after 2015. PRODES exhibited a gradual increase from ~\u0026thinsp;100 Mt to ~\u0026thinsp;280 Mt by 2020\u0026ndash;2021 with a smooth trajectory, while GFW showed a sharper, earlier spike reaching\u0026thinsp;~\u0026thinsp;850 Mt around 2016\u0026ndash;2017 in Scenario 1a, three times PRODES' value and coinciding with GFW's methodological updates that enabled broader forest disturbance detection. Both declined toward 2023, with PRODES dropping to ~\u0026thinsp;150 Mt and GFW falling steeply to 400\u0026ndash;450 Mt while maintaining roughly double PRODES values. The discounted emissions (accounting for linear and market share adjustments) showed more contrasts: PRODES displayed gradual increases from near zero to ~\u0026thinsp;5 Mt by 2015, accelerating to peaks of 16\u0026ndash;23 Mt (varying by scenario) around 2020\u0026ndash;2021 in a broad bell-shaped curve, then declining moderately to 10\u0026ndash;12 Mt by 2023.\u003c/p\u003e \u003cp\u003eGFW discounted emissions (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, lower panel) displayed much more dramatic peaks and variability. From 2004 to 2015, increases were similar to PRODES, rising to ~\u0026thinsp;10\u0026ndash;15 Mt. However, between 2016 and 2018, we observed sharp spikes that far exceeded any PRODES values: Scenario 1a reached over 55 Mt around 2017, Scenario 1b peaked around 45 Mt, and Scenarios 2a and 2b reached approximately 30\u0026ndash;35 Mt. These peaks were roughly 2.5 to 3 times higher than PRODES peaks occurring three to four years later. GFW curves showed sharp, concentrated peaks rather than sustained elevation. After 2018, steep declines dropped to near zero by 2023, while PRODES declined gradually to approximately 5\u0026ndash;10 Mt.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eComparison of PRODES and GFW-based deforestation emissions (2004\u0026ndash;2023) across the supply chain scenarios. Values are reported as gross and discounted MtCO₂e, with percentage differences showing how much higher GFW estimates are relative to PRODES.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eScenario\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003ePRODES\u003c/p\u003e \u003cp\u003e(MtCO2e)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003eGFW\u003c/p\u003e \u003cp\u003e(MtCO2e)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e \u003cp\u003eDifference\u003c/p\u003e \u003cp\u003e(%)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGross\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eDiscounted\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGross\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDiscounted\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eGross\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eDiscounted\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSc1a: Slaughterhouse Buffer no Market Share\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e3,937.53\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e179.73\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e7,944.09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e408.36\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e102%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e127%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSc1b: Slaughterhouse Buffer with Market Share\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e3,937.53\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e128.07\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e7,944.09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e287.45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e102%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e124%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSc2a: Direct Supplier's Zone 2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e3,717.07\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e116.66\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e7,520.27\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e263.77\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e102%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e126%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSc2b: Direct Supplier's Zone 2004-23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4,202.21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e133.28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e8,457.10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e296.78\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e101%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e123%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"4. DISCUSSION","content":"\u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003e4.1. Data availability and the limits of top-down approaches.\u003c/h2\u003e \u003cp\u003eCorporate climate accounting continues to face substantial data constraints, particularly in sectors characterized by complex supply chains and limited traceability, such as Brazilian cattle production (Zu Ermgassen et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Garrett et al., 2021). Most companies lack georeferenced supplier information, and none can track animal movements continuously from birth to slaughter. As discussed in the Scope 3 literature, high transaction costs associated with searching, coordinating, and monitoring suppliers often outweigh the perceived benefits of disclosure, unless firms can anticipate a clear competitive advantage (Patchell, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). Data privacy concerns and fragmented information systems further constrain the collection of farm-level land-use change data (Stenzel and Waichman, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Furthermore, many suppliers in general are reluctant to share information due to concerns over confidentiality, competitive exposure, and regulatory risk (Zu Ermgassen et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). As a result, companies remain reliant on proxies rather than precise, farm-level traceability, which weakens incentives for transparency, since firms cannot credibly claim performance above sectoral baselines.\u003c/p\u003e \u003cp\u003eIn response to these constraints, many companies adopt top-down approaches based on aggregated land-use change statistics. For example, the global emissions analysis tool developed by Fitts et al. (2025), which is operationally simple, but approaches operate at broad spatial scales, typically countries or states, and therefore overlook the geographic heterogeneity of sourcing. This limitation is consequential, because deforestation-related emissions vary substantially across locations and over time. Moreover, global emission factors often fail to reflect local production conditions. For instance, Estevam and Assad (2025) show that IPCC default emission factors tend to exceed Brazilian averages for the cattle sector.\u003c/p\u003e \u003cp\u003eOur supply-zone approach directly addresses these limitations by estimating deforestation-related emissions with greater spatial specificity. By linking slaughterhouse operations to surrounding sourcing areas, the framework captures spatial patterns of deforestation more accurately than purely top-down methods. Both simplified buffer-based zones and zones derived from actual supplier locations produced similar emissions estimates in our analysis. This convergence reflects the spatial clustering of deforestation, slaughterhouses, and direct suppliers in accessible regions, as well as the fact that the 200 km buffers were calibrated using empirically observed distances between slaughterhouses and their suppliers. However, accurately allocating emissions also requires accounting for slaughterhouse market share within each supply zone. Without this adjustment, emissions from shared sourcing areas may be fully attributed to a single facility, leading to systematic overestimation. This issue has been highlighted in public assessments, such as a recent investigation by Greenpeace (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2025\u003c/span\u003e), which relied on simplified sourcing assumptions that likely overstated company-level deforestation responsibility in the Amazon. In contrast, our framework explicitly incorporates market share, reducing the risk of double counting and misattribution.\u003c/p\u003e \u003cp\u003eOur supply-zone framework offers a scalable and relatively low-cost pathway toward more accurate Scope 3 accounting. By narrowing system boundaries from the national to the regional level, it approximates jurisdictional approaches to direct land-use change accounting while remaining compatible with corporate data realities (Fitts et al., 2025). The framework also allows for methodological flexibility depending on data availability. When only slaughterhouse locations and market shares are known, Scenario 1b provides the most appropriate balance by allocating emissions proportionally and avoiding overestimation in shared zones. Scenario 1a, which applies uniform buffers without market share adjustment, tends to overestimate emissions in regions accessible to multiple slaughterhouses. When georeferenced data on direct suppliers are available, Scenario 2a offers higher spatial precision by restricting sourcing areas to active suppliers in the reporting year. Extending sourcing zones to include historical suppliers can artificially inflate sourcing areas and is often infeasible for companies lacking long-term supplier records.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e4.2. Implications of using PRODES vs GFW for emissions accounting.\u003c/h2\u003e \u003cp\u003eThe comparison between PRODES and GFW illustrates how methodological choices can substantially influence deforestation-related emissions estimates. Although both datasets capture deforestation in the Brazilian Amazon, across all evaluated scenarios, GFW values were approximately 100% higher than the PRODES results.\u003c/p\u003e \u003cp\u003eThese discrepancies arise from fundamental differences in scope, detection logic, and classification criteria. PRODES focuses on clear-cut deforestation of native vegetation, identified through expert automatic detection and visual interpretation within a jurisdictionally consistent and legally grounded framework. GFW, in contrast, detects a broader spectrum of forest disturbances, including fire, degradation, and selective logging, using automated global algorithms based on spectral change detection (Sims and Goldman, 2025). The pronounced increase in GFW-based estimates between 2016 and 2018, which is not observed in PRODES, reflects methodological updates that expanded sensitivity to partial canopy loss rather than a sudden shift in clear-cut deforestation dynamics.\u003c/p\u003e \u003cp\u003eThese methodological differences have direct implications for how companies quantify and report supply-chain emissions (Fitts et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2025b\u003c/span\u003e). While GFW offers global coverage and is valuable in contexts where national monitoring systems are weak or inconsistent, PRODES provides more stable, transparent, and legally grounded estimates in Brazil, where deforestation definitions are formally codified within the national monitoring framework. As a result, PRODES is widely used by companies operating in the Brazilian Amazon to assess supplier compliance with zero-illegal-deforestation commitments (Gibbs et al., \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Raoni et al., 2020). For companies sourcing cattle in Brazil, reliance on PRODES enhances coherence with national regulations and strengthens the credibility of emissions reporting. PRODES also adheres to the principle of conservativeness by prioritizing legal clarity, institutional validation, and methodological consistency over time. Although it does not capture all forms of forest disturbance detected by GFW, this exclusion reflects a deliberate methodological choice aligned with enforcement and regulatory metrics. Companies that apply GFW-based estimates in Brazil without appropriate contextual adjustments therefore risk systematically overestimating deforestation-related emissions.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e4.3. Model limitations and recommendations\u003c/h2\u003e \u003cp\u003eThis study introduces a scalable, spatially explicit model to estimate deforestation-related emissions from cattle supply chains, serving as a bridge between statistical land-use change and direct land use methods (Fitts et al., 2025). By linking emissions to supply zones and applying region-specific carbon factors, the approach improves both geographic precision and methodological transparency. Future improvements, however, must address the \"double counting\" inherent in Scope 3 accounting, where emissions are shared across complex supply networks. As Hertwich \u0026amp; Wood (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) note, Scope 3 emissions are essentially the Scope 1 emissions of upstream actors; acknowledging this shared responsibility is vital for identifying leverage points for mitigation rather than simply allocating blame. To refine the allocation of emissions in overlapping supply zones, methodologies such as product expansion or shared responsibility allocation should be applied to distribute burdens equitably based on market share or land footprint (Fitts et al., 2025).\u003c/p\u003e \u003cp\u003eDespite its strengths, the model has important limitations. First, it does not incorporate forest regeneration, afforestation, or transitions from pasture to cropland. These land-use dynamics may affect long-term emissions but remain untracked in national datasets like PRODES. Second, the approach does not differentiate primary from secondary forests, which according to Imazon have been deforested at nearly identical rates; between 2019 and 2023, both primary and secondary forests lost slightly over 2\u0026nbsp;million hectares each (Imazon, 2025). Third, and most critically, supply-zone methods aggregate suppliers to multiple slaughterhouses, which means that it does not differentiate deforestation levels from low and high-performers, or from slaughterhouses with stringent monitoring deforestation protocols or no protocols at all. Additional work is needed to develop approaches that account for company actions to reduce deforestation within their specific supply chain.\u003c/p\u003e \u003cp\u003eSpatial heterogeneity introduces further complexity. In Par\u0026aacute;, for example, deforestation has moved westward over the past two decades. This means a slaughterhouse operating today in a low-deforestation zone may have contributed to forest loss earlier in the supply chain\u0026rsquo;s history. When emissions are temporally discounted, older impacts become less visible, even if they remain ecologically relevant.\u003c/p\u003e \u003cp\u003eFuture improvements should focus on integrating dynamic supply chain data, particularly cattle movement records. This would allow emissions to be allocated based on actual trade flows rather than geographic proxies. For indirect suppliers, such data could reveal laundering patterns and strengthen traceability beyond first-tier sourcing. For direct suppliers, emissions could be reduced by favoring full-cycle ranches with CAR registration and stable animal flows. These enhancements would improve both precision and integrity in emissions accounting. Supply zones are not a substitute for traceability, but they are an actionable step forward. They offer companies a transparent, repeatable method to estimate land-use emissions using currently available data. Until full traceability becomes standard, spatial models like this can help bridge the gap between feasibility and accountability.\u003c/p\u003e \u003c/div\u003e"},{"header":"5. CONCLUSIONS","content":"\u003cp\u003eOur method provides a practical and scalable solution for estimating Scope 3 deforestation emissions in cattle supply chains in Brazil using data that are readily available. This represents a critical step forward for companies seeking to align with climate disclosure frameworks and meet regulatory expectations. As corporate climate targets increasingly require geographically explicit emissions data, our approach enables meatpackers to generate credible estimates without waiting for perfect traceability systems to be in place. Especially, if companies can apply Scenario 1b or Scenario 2a, depending on the data available.\u003c/p\u003e \u003cp\u003eHowever, our results also underscore the limitations of current methodologies and the urgent need for more precise tools to fully capture the extent of land-use change emissions. Our zone-based methods offer feasible proxies in data-constrained contexts, but they remain approximate. While farm to slaughter traceability of cattle introduces higher upfront costs and may require stronger government oversight, it also enables companies to make robust, verifiable zero-deforestation claims.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003eGHG\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Greenhouse Gas\u003c/p\u003e\n\u003cp\u003eGHGP\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Greenhouse Gas Protocol\u003c/p\u003e\n\u003cp\u003eCO₂e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Carbon Dioxide Equivalent\u003c/p\u003e\n\u003cp\u003eCDP\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Carbon Disclosure Project\u003c/p\u003e\n\u003cp\u003eEUDR \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;European Union Deforestation Regulation\u003c/p\u003e\n\u003cp\u003eCSRD \u0026nbsp;\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Corporate Sustainability Reporting Directive\u003c/p\u003e\n\u003cp\u003eSIF \u0026nbsp;\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Servi\u0026ccedil;o de Inspe\u0026ccedil;\u0026atilde;o Federal (Federal Inspection Service)\u003c/p\u003e\n\u003cp\u003eSIE \u0026nbsp;\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Servi\u0026ccedil;o de Inspe\u0026ccedil;\u0026atilde;o Estadual (State Inspection Service)\u003c/p\u003e\n\u003cp\u003eSIM \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Servi\u0026ccedil;o de Inspe\u0026ccedil;\u0026atilde;o Municipal (Municipal Inspection Service)\u003c/p\u003e\n\u003cp\u003eCAR\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Cadastro Ambiental Rural (Rural Environmental Registry)\u003c/p\u003e\n\u003cp\u003eCPF \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Cadastro de Pessoas F\u0026iacute;sicas (Brazilian Individual Taxpayer Registry)\u003c/p\u003e\n\u003cp\u003eCNPJ \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Cadastro Nacional da Pessoa Jur\u0026iacute;dica (Brazilian Corporate Taxpayer Registry)\u003c/p\u003e\n\u003cp\u003eGTA\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Guia de Tr\u0026acirc;nsito Animal (Animal Transit Guide)\u003c/p\u003e\n\u003cp\u003eINPE \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Instituto Nacional de Pesquisas Espaciais (Brazilian National Institute for Space Research)\u003c/p\u003e\n\u003cp\u003eMAPA \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Minist\u0026eacute;rio da Agricultura, Pecu\u0026aacute;ria e Abastecimento (Brazilian Ministry of Agriculture, Livestock, and Supply)\u003c/p\u003e\n\u003cp\u003eADEPARA\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Ag\u0026ecirc;ncia de Defesa Agropecu\u0026aacute;ria do Estado do Par\u0026aacute; (Par\u0026aacute;\u0026apos;s State Agricultural and Livestock Health Agency)\u003c/p\u003e\n\u003cp\u003eSEMA. \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Secretaria de Estado de Meio Ambiente e Sustentabilidade (Par\u0026aacute;\u0026rsquo;s Environmental Secretariat)\u003c/p\u003e\n\u003cp\u003eGFW \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Global Forest Watch\u003c/p\u003e\n\u003cp\u003eGLEAM-i \u0026nbsp; \u0026nbsp;\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Global Livestock Environmental Assessment Model \u0026ndash; interactive\u003c/p\u003e\n\u003cp\u003eALCI. \u0026nbsp;\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Agricultural Life Cycle Inventory Generator\u003c/p\u003e\n\u003cp\u003eSBTi\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Science Based Targets initiative\u003c/p\u003e\n\u003cp\u003eMt \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Million metric tons\u003c/p\u003e\n\u003cp\u003eMha \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Million hectares\u003c/p\u003e\n\u003cp\u003etCO₂e/ha \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Tons of CO₂ equivalent per hectare\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe data and materials used in this study are available from the corresponding author upon request, subject to data-sharing restrictions.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eLisa Rausch and Hollly Gibbs have an ongoing consulting relationship with the nonprofit National Wildlife Federation. The National Wildlife Federation did not provide editorial oversight over the contents of the manuscript. Fabio Martins Guerra Nunes Dias is employed by the meat industry, but this association does not compromise the objectivity or integrity of the study’s results. \u0026nbsp;The other authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eHis work was supported by the Gordon and Betty Moore Foundation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to Publish\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eConsent to Publish declaration: not applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics and Consent to Participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eEthics and Consent to Participate declarations: not applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors' contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA.B.J. conceived the study, designed the methodology, analyzed the data, and wrote the original draft. S.C.K. contributed to study design and interpretation and reviewed and revised the manuscript. F.D., J.M., L.R., and S.S.L. provided substantive comments and revisions on the manuscript. H.K.G. supervised the research, and acquired funding. All authors reviewed and approved the final version of the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe would like to thank Malena Candido for reviewing different early versions of the manuscript.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAlix-Garcia, J., \u0026amp; Gibbs, H. K. (2017). Forest conservation effects of Brazil\u0026apos;s zero deforestation cattle agreements undermined by leakage. Global Environmental Change, 47, 201\u0026ndash;217.\u003c/li\u003e\n\u003cli\u003eBarreto, P., Pereira, R., Brand\u0026atilde;o Jr., A., \u0026amp; Baima, S. (2017). Will meat-packing plants help halt deforestation in the Amazon? Imazon \u0026amp; Instituto Centro de Vida.\u003c/li\u003e\n\u003cli\u003eBrand\u0026atilde;o Jr., A., Rausch, L., Gibbs, H. K., \u0026amp; West, T. A. (2023). Spatially explicit assessment of cattle ranching and deforestation in the Brazilian Amazon. Environmental Research Letters, 18(4), 044030.\u003c/li\u003e\n\u003cli\u003eCDP. (2020). Missing the Mark: The state of corporate deforestation action. CDP Worldwide.\u003c/li\u003e\n\u003cli\u003eEuropean Commission. (2021). Proposal for a regulation on the making available on the Union market and the export from the Union of certain commodities and products associated with deforestation and forest degradation (COM(2021) 706 final).\u003c/li\u003e\n\u003cli\u003eFitts, L., James, O., Gibbs, D., Goldman, E., Harris, N., Ramlow, M., Sloat, L., Wilson, C., Winchester, C., Ernstoff, A., \u0026amp; Hohmann, T. (2025a). Geospatial methods for corporate GHG accounting of deforestation and land occupation. World Resources Institute. https://www.wri.org/research/geospatial-methods-corporate-ghg-accounting-deforestation-and-land-occupation\u003c/li\u003e\n\u003cli\u003eFitts, L., James, O., Gibbs, D., Goldman, E., Harris, N., Ramlow, M., Sloat, L., Wielgosz, B., Wilson, C., Winchester, C., \u0026amp; Ernstoff, A. (2025b). Statistical land use change emissions from deforestation and land occupation for crops. World Resources Institute. https://www.wri.org/research/statistical-land-use-change-emissions-deforestation-and-land-occupation-crops\u003c/li\u003e\n\u003cli\u003eGibbs, H. K., Rausch, L., Munger, J., Schelly, I., Morton, D. C., Noojipady, P., \u0026amp; Walker, N. F. (2015). Brazil\u0026apos;s Soy Moratorium. Science, 347(6220), 377\u0026ndash;378.\u003c/li\u003e\n\u003cli\u003eGreenhouse Gas Protocol (GHGP). (2011). Corporate value chain (Scope 3) accounting and reporting standard. World Resources Institute \u0026amp; World Business Council for Sustainable Development. https://ghgprotocol.org/sites/default/files/standards/Corporate-Value-Chain-Accounting-Reporing-Standard_041613_2.pdf\u003c/li\u003e\n\u003cli\u003eGreenhouse Gas Protocol (GHGP). (2013). Technical guidance for calculating Scope 3 emissions. World Resources Institute \u0026amp; World Business Council for Sustainable Development.\u003c/li\u003e\n\u003cli\u003eGreenhouse Gas Protocol (GHGP). (2026). Land Sector and Removals Standard. World Resources Institute \u0026amp; World Business Council for Sustainable Development. https://ghgprotocol.org/land-sector-and-removals-standard\u003c/li\u003e\n\u003cli\u003eGreenpeace. (2025). JBS: Cooking the Planet. Case study: JBS\u0026apos; supply chain linked to cattle raised illegally on Indigenous Land. Greenpeace International.\u003c/li\u003e\n\u003cli\u003eHarris, N. L., Gibbs, D. A., Baccini, A., Birdsey, R. A., De Bruin, S., Farina, M., Fatoyinbo, L., Hansen, M. C., Herold, M., Houghton, R. A., \u0026amp; Potapov, P. V. (2021). Global maps of twenty-first century forest carbon fluxes. Nature Climate Change, 11(3), 234\u0026ndash;240.\u003c/li\u003e\n\u003cli\u003eHertwich, E. G., \u0026amp; Wood, R. (2018). The growing importance of scope 3 greenhouse gas emissions from industry. Environmental Research Letters, 13(10), 104013.\u003c/li\u003e\n\u003cli\u003eHettler, M., \u0026amp; Graf-Vlachy, L. (2024). Corporate scope 3 carbon emission reporting as an enabler of supply chain decarbonization: A systematic review and comprehensive research agenda. Business Strategy and the Environment, 33(2), 263\u0026ndash;282.\u003c/li\u003e\n\u003cli\u003eJ\u0026uacute;nior, D., \u0026amp; Russo, A. (2025, December 2). Com cr\u0026iacute;ticas ao mercado internacional, Par\u0026aacute; adia in\u0026iacute;cio da obrigatoriedade da rastreabilidade bovina. Estad\u0026atilde;o Agro. https://agro.estadao.com.br/pecuaria/com-criticas-ao-mercado-internacional-para-adia-obrigatoriedade-de-rastreabilidade-bovina.\u003c/li\u003e\n\u003cli\u003ePanwar, R. (2023). Business and biodiversity: Achieving the 2050 vision for biodiversity conservation through transformative business practices. Biodiversity and Conservation, 32(11), 3607\u0026ndash;3613.\u003c/li\u003e\n\u003cli\u003ePatchell, J. (2018). Can the implications of the GHGP\u0026apos;s scope 3 standard be realized? Journal of Cleaner Production, 185, 941\u0026ndash;958.\u003c/li\u003e\n\u003cli\u003eRaj\u0026atilde;o, R., Soares-Filho, B., Nunes, F., B\u0026ouml;rner, J., Machado, L., Assis, D., \u0026amp; Oliveira, A. (2020). The rotten apples of Brazil\u0026apos;s agribusiness. Science, 369(6501), 246\u0026ndash;248.\u003c/li\u003e\n\u003cli\u003eRausch, L., Munger, J., \u0026amp; Gibbs, H. K. (2020). Amazon deforestation linked to European imports via specially licensed ranches and their suppliers. Gibbs Land Use and Environment Report Series. https://gibbs-lab.wisc.edu/assets/SISBOV_report_March_2020.pdf.\u003c/li\u003e\n\u003cli\u003eScience Based Targets Initiative (SBTi). (2024). Aligning corporate value chains to global climate goals: Scope 3 discussion paper.\u003c/li\u003e\n\u003cli\u003eSkidmore, M. E., Moffette, F., Rausch, L., \u0026amp; Gibbs, H. K. (2020). Characterizing compliance in cattle supply chains: What factors encourage deforestation-free production in the Brazilian Amazon? Evidensia. https://www.evidensia.eco/resources/937/download/\u003c/li\u003e\n\u003cli\u003eSkidmore, M. E., Moffette, F., Rausch, L., Christie, M., Munger, J., \u0026amp; Gibbs, H. K. (2021). Cattle ranchers and deforestation in the Brazilian Amazon: Production, location, and policies. Global Environmental Change, 68, 102280.\u003c/li\u003e\n\u003cli\u003eSkidmore, M. E., Sims, K. M., Rausch, L. L., \u0026amp; Gibbs, H. K. (2022). Sustainable intensification in the Brazilian cattle industry: The role for reduced slaughter age. Environmental Research Letters, 17(6), 064026.\u003c/li\u003e\n\u003cli\u003eSkidmore, M. E., Barrozo, M., Rausch, L., \u0026amp; Gibbs, H. K. (2024). Cattle supply chain fidelity in the Brazilian Amazon: Key considerations for deforestation monitoring. Environmental Research Letters, 19(11).\u003c/li\u003e\n\u003cli\u003eSouza Jr., C. M., Shimbo, J. Z., Rosa, M. R., et al. (2020). Reconstructing three decades of land use and land cover changes in Brazilian biomes with Landsat archive and Earth Engine. Remote Sensing, 12(17), 2735.\u003c/li\u003e\n\u003cli\u003eStenzel, A., \u0026amp; Waichman, I. (2023). Supply-chain data sharing for scope 3 emissions. npj Climate Action, 2(1), 7.\u003c/li\u003e\n\u003cli\u003eWalker, N. F., Patel, S. A., \u0026amp; Kalif, K. A. B. (2013). From Amazon pasture to the high street: Deforestation and the Brazilian cattle product supply chain. Tropical Conservation Science, 6(3), 446\u0026ndash;467.\u003c/li\u003e\n\u003cli\u003eWeisse, M., \u0026amp; Potapov, P. (2021). Assessing trends in tree cover loss over 20 years of data. Global Forest Watch. https://www.globalforestwatch.org/blog/data-and-tools/tree-cover-loss-satellite-data-trend-analysis/\u003c/li\u003e\n\u003cli\u003eWest, T. A., Munger, J., \u0026amp; Gibbs, H. K. (2022). Regional variation in zero-deforestation cattle commitments and their potential impact in the Brazilian Amazon. Environmental Research Letters, 17(3), 034014.\u003c/li\u003e\n\u003cli\u003ezu Ermgassen, E. K., Godar, J., Lathuilli\u0026egrave;re, M. J., L\u0026ouml;fgren, P., Gardner, T., Vasconcelos, A., \u0026amp; Meyfroidt, P. (2020). The origin, supply chain, and deforestation risk of Brazil\u0026apos;s beef exports. Proceedings of the National Academy of Sciences, 117(50), 31770\u0026ndash;31779.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"carbon-balance-and-management","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"cbam","sideBox":"Learn more about [Carbon Balance and Management](https://cbmjournal.biomedcentral.com/)","snPcode":"13021","submissionUrl":"https://submission.nature.com/new-submission/13021/3","title":"Carbon Balance and Management","twitterHandle":"@BioMedCentral","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Scope 3, Deforestation, Greenhouse Gas Protocol (GHGP), Cattle supply chains, Land-use change emissions, spatially explicit accounting","lastPublishedDoi":"10.21203/rs.3.rs-9138992/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9138992/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eBackground: The Greenhouse Gas Protocol (GHGP) distinguishes Scope 1 emissions, which occur directly at the facility or company, from Scope 2 emissions linked to electricity use, and Scope 3 emissions associated with other supply-chain inputs. Estimating Scope 3 deforestation-related emissions in Brazil’s cattle sector remains challenging due to the fragmented nature of supply chains and the lack of national farm-to-gate traceability. Most companies cannot identify indirect suppliers, where deforestation is increasingly concentrated, making Scope 3 land-use change emissions difficult to quantify. Instead, they use top-down statistical land-use change models based on national, state or municipal emission factors which are usually limited regarding variation of companies’ carbon footprint. To address this gap, we developed a spatially explicit method aligned with the GHGP’s Land Sector and Removals Guidance, using data already available to slaughterhouses to estimate cattle supply zones and compare alternative input datasets and methods.\u003c/p\u003e\n\u003cp\u003eResults: Emissions estimates varied significantly depending on the deforestation dataset and spatial attribution method. Global Forest Watch deforestation data produced gross emissions nearly 100% higher and discounted emissions about 120% higher than PRODES data. Supply zones based on locations of direct suppliers produced emissions up to 25% lower than fixed-radius buffers around slaughterhouses. When adjusted for market share and temporal discounting, total emissions decreased by 85–97%. While supply zones improved geographic precision relative to fixed-radius buffers and municipal, state or country estimates, they still include both suppliers and non-suppliers and do not fully capture company-specific mitigation actions.\u003c/p\u003e\n\u003cp\u003eConclusions: This study presents a practical, replicable method for estimating deforestation-related Scope 3 emissions using currently available data for Brazil’s cattle sector. The approach is consistent with GHGP recommendations and supports corporate reporting. Supply zones represent an important advance over traditional top-down statistical land-use change methods. \u0026nbsp;Improving accuracy of Scope 3 emissions from deforestation however, will require expanded traceability capable of identifying all animals and all tiers of cattle suppliers.\u003c/p\u003e","manuscriptTitle":"Calculating Supply Chain Deforestation Emissions for Brazil’s Meatpacking Sector: A New Corporate Accounting Methodology","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-13 15:03:30","doi":"10.21203/rs.3.rs-9138992/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewerAgreed","content":"282116016081467701936552204123757395106","date":"2026-05-11T14:58:17+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-04-07T03:44:16+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-03-26T04:24:45+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-03-26T04:24:30+00:00","index":"","fulltext":""},{"type":"submitted","content":"Carbon Balance and Management","date":"2026-03-16T14:06:45+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"carbon-balance-and-management","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"cbam","sideBox":"Learn more about [Carbon Balance and Management](https://cbmjournal.biomedcentral.com/)","snPcode":"13021","submissionUrl":"https://submission.nature.com/new-submission/13021/3","title":"Carbon Balance and Management","twitterHandle":"@BioMedCentral","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"189ffdc5-9eb2-4e8d-b2ab-225aafa58138","owner":[],"postedDate":"April 13th, 2026","published":true,"recentEditorialEvents":[{"type":"reviewerAgreed","content":"282116016081467701936552204123757395106","date":"2026-05-11T14:58:17+00:00","index":51,"fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-04-13T15:03:30+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-13 15:03:30","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9138992","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9138992","identity":"rs-9138992","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}