Understand do Pasto ao Prato’s data

Frequently Asked Questions

• How does dPaP map where the cattle comes from and define each plant's supply shed?

  The app relies on a document that is required by law in Brazil: the Animal Transit Permit (GTA). Every time a batch of cattle is moved from one farm to another, or from a farm to a meatpacking plant, that movement has to be recorded in a GTA. It works like the cattle's "invoice", and it is required by sanitary regulation.

  From the GTAs we calculate what share of the cattle entering each plant comes from each Brazilian municipality. The more cattle a municipality sends to a plant, the greater its weight in that plant's supply chain. The set of municipalities and their weights is what we call the plant's supply shed. This accounts for both the direct and indirect sourcing of cattle by the slaughterhouse.

  Even though GTAs are essential documents for reconstructing Brazil's cattle chain, whether for sanitary or other purposes, not every Brazilian state publishes this data. For meatpacking plants without a GTA trail, dPaP uses two different approaches depending on the type of facility. For units that only process or repack beef (but not slaughtering animals), the app uses a broader estimate: the average origin of cattle slaughtered in federally inspected plants in the same state. For the other plants, we use a machine learning model to estimate the supply shed.

• Why does the methodology look at risk across the meatpacker’s ‘supply shed’ instead of the scale of each rural property?

  dPaP looks at the deforestation occurring across the supply shed (the volumes of cattle sourced from each municipality), and not at each individual farm. This is because the supply shed approach is more complete than property-level monitoring. Property-level monitoring misses 71% of deforestation and 50% of the pasture area in the Brazilian Amazon because of inconsistencies in animal movement records and property-level data (farm names, owners).

• What life-cycle and moving-average time frames are applied to the indicators?

  • 5-year life cycle: Deforestation and fire measurements accumulate the 5 years prior to the production base year, which is the animal's lifespan until slaughter.
  • 5-year moving average: The calculation uses a 5-year moving average to smooth out atypical spikes.

  Deforestation exposure: instead of treating deforestation as an event tied to a specific animal, we consider deforestation to be shared by all the cattle that share the purchasing zone. Deforestation is transferred proportionally along the beef chain according to the traded volumes. This attribution of deforestation is called "mass balance".

• What are the five socio-environmental and sanitary indicators assessed by dPaP?

  dPaP monitors and ranks meatpacking plants across five analytical dimensions:

  • Socio-environmental Commitments and Compliance: Identifies the existence of agreements (such as TACs with the Federal Prosecution Service) and multi-sector protocols (Boi na Linha and the Cerrado Protocol), and measures compliance audited in a public, independent way.
  • Deforestation: Measures the conversion of native vegetation in the purchasing zone (in hectares per thousand tonnes of beef produced).
  • Fires: Assesses the number of heat spots recorded in pasture areas within the purchasing zone.
  • Slave-like Labor: Crosses the CPF/CNPJ in the GTAs of direct and indirect suppliers with the Registry of Employers that subjected workers to slavery-like conditions (the "Dirty List") published by the Ministry of Labor and Employment.
  • Sanitation and Animal Welfare: Assesses the history of fines for sanitary infractions compared with plants of similar size under the same inspection instance (SIF, SIE, SIM or SISBI).
• How does the scoring and final rating system shown in the app work?

  The final rating assigned to products in the app (ranging from Very good to Very bad and Unknown) results from crossing the deforestation risk in the purchasing zone with the plant's mitigation capacity:

  • Mitigation and Transparency: If a plant holds public commitments, adopts sector protocols and publishes independent audit results showing high compliance, its rating is positive. Regardless of the risk assigned to its purchasing zone, the plant demonstrates control over the environmental impact of its suppliers.
  • Lack of Diligence and High Risk: If the plant operates in a region with high deforestation or fire rates and holds no public commitments under sector protocols, runs no transparent audits or, even with audits, shows low compliance, it takes on the risk assigned to its purchasing zone. These plants receive unfavorable ratings, depending on the degree of risk assigned to their purchasing zone.
  • Unknown: Category assigned when there is no public audit data or not enough official information about that plant's cattle movements (GTA) to map its chain.
• What are the main challenges dPaP faces in building the data?

  Building dPaP's indicators runs into structural challenges of transparency and public data governance in Brazil. The GTA (Animal Transit Guide) is managed in a decentralized way by each state, with no standardization of format, frequency or access conditions, and in most states it is not made available in a public, systematic way.

  Moreover, in most states the GTA carries no link to the Rural Environmental Registry (CAR), which prevents directly connecting a cattle movement to a specific georeferenced property. Even when that link is pursued through other means, access to the CAR's nominative data is limited, making identification difficult. Added to this is the temporary interruption or restriction of public access to the Slave Labor Dirty List, another source used to monitor irregularities in the chain.

  Finally, informal slaughter, which by definition escapes any inspection system, remains structurally invisible to this public data and can only be inferred indirectly. Together, these barriers show that the data infrastructure needed for effective environmental governance of the beef chain already partly exists, but is fragmented, incomplete or restricted. This gap limits both public enforcement and voluntary transparency initiatives such as dPaP itself.