As the European Union Deforestation Regulation (EUDR) deadline approaches, coffee and cocoa supply chains face growing pressure to demonstrate deforestation-free sourcing. With compliance required by December 2026 for large operators and June 2027 for qualifying small enterprises, traders are focused on mitigating sourcing risks and safeguarding market access.
Yet, a major risk remains: inaccurate deforestation risk classifications, particularly “false positives,” where plots are incorrectly classified as having deforestation risk due to limitations in mapping or data accuracy. The direct consequences could fall on small-scale farmers, who may struggle to demonstrate compliance and risk being excluded from international markets, even when their production meets deforestation-free requirements.
To support supply-chain actors, Solidaridad designed an approach that reduces inaccurate deforestation risk classifications, particularly “false positives,” on small-scale plots managed under agroforestry systems, while combining different datasets to provide the auditable evidence required for EUDR due diligence and responsible sourcing.
Stronger evidence for fairer supply chains
Standard global geospatial tools can struggle to distinguish small-scale farms and agroforestry systems from natural forests, secondary forests, pastures and transitional areas occurring in close proximity, generating severe “false positive” deforestation alerts that threaten to unfairly exclude sustainable farmers from European markets. Protect Forests & Farmers (PFF) is a project implemented by Solidaridad and supported by the Postcode Lottery to address and prevent this exclusion.
During its first phase, the project focused on developing and testing a scalable, cost-effective approach adapted to the realities of agroforestry systems and small-scale agriculture. With this approach, more accurate data on forest, non-forest and deforestation is produced, so farmers who have protected forests and produced sustainably are not penalised by inaccurate maps.
Caption: The approach developed by Solidaridad combines geospatial analysis, machine learning algorithms, high-resolution satellite imagery and field validated information to produce highly accurate forest/non forest and deforestation layers—reducing false positives and enabling verifiable EUDR compliance
The PFF team selected two pilot areas of approximately 10,000 km² each, in Honduras and Peru and calculated that widely used global datasets generate false-positive rates of up to 31% in Peru and 78% in Honduras, in the case of the Hansen layer. The layers developed by the team substantially reduced these errors, bringing false-positive rates down to approximately 4–11%, while overall accuracy reached around 90% in the pilot areas.
With these promising results, the approach was further applied across approximately 80,000 km² of production landscapes in Peru and Honduras, and tested across 10,000 km² in Brazil. Building on these pilots, the PFF team consolidated a scalable approach for generating more accurate land-use information where it is most needed, across larger areas and different contexts.
Auditable EUDR evidence through convergence of evidence
The team also identified an opportunity to create additional value for supply-chain actors by translating this improved land-use information into stronger, more reliable evidence for deforestation risk assessments. To achieve this, the project adopted a “convergence of evidence” approach, combining multiple sources of information – including global datasets, local cartography and PFF-generated layers – to build a more robust and traceable assessment. Rather than relying on a single dataset, the approach cross-checks evidence from different sources to determine whether they point to the same conclusion. This helps overcome the limitations of individual datasets, reduce false positives and provide a more objective and reliable assessment of deforestation risk across different landscapes.
The analysis is complemented by an environmental legality assessment, which checks each plot against official information on land-use restrictions and compliance, including protected areas, indigenous lands and available records of environmental penalties.
This work has been developed in close engagement with key local stakeholders in both countries, ensuring that it addresses gaps in existing data and tools and responds to country-specific needs and technical requirements. Working with local technical experts and institutions, including MIDAGRI in Peru and IHCAFE in Honduras, has helped ensure that the solution is relevant, fit for purpose and aligned with national and sectoral needs. In addition, Solidaridad is engaging with European operators to better understand their needs and identify gaps in current solutions. These insights are helping us to tailor the solution to provide practical support to companies in meeting their EUDR due diligence requirements.
Phase 2 will build on these results by integrating the convergence-of-evidence analysis into a broader supply-chain data management platform called ‘Aura+‘. This solution will make reliable and actionable deforestation risk information more accessible and scalable for companies and farmer organizations while reducing the cost and complexity of managing farm-level data. More importantly, results will be delivered as auditable evidence that companies can use to support their EUDR due diligence processes and responsible sourcing.
In the future, this solution will also include a module to help companies and farmer organizations quantify farm-level carbon footprints and develop decarbonization pathways.
From better data to better opportunities for farmers
Better geographic information is only part of the solution. The project also works directly with small-scale farmers to strengthen their readiness for changing market requirements while supporting more resilient and productive farming practices.
In the first phase of the project, 1,164 small farmers in Peru and Honduras received orientation about upcoming EUDR requirements and 815 received technical training on deforestation-free farming practices. In Peru, these activities reached producers from the Acopagro, Cuencas del Huallaga and Valle Grande cooperatives in the San Martín region, as well as the Mareriaki and Coopagro cooperatives in the Junín region. In Honduras, 729 producers across nine partner organizations (including ASOPROSAN, CAFESMO, COMICAOL, COMICLOL and COPRACNIL) participated in awareness sessions and intensive technical training. As a result, 715 producers (managing 967 agroforestry plots) across the two countries have so far secured access to the European market.
The project demonstrates that meeting deforestation-free requirements can go hand-in-hand with higher farm profitability. In Honduras, small-scale farmers managing less than 1.5 hectares typically produce around 821 kg/ha. Farmers adopting the project’s Climate-Smart Agriculture package, including shade management, soil conservation and targeted pruning, more than doubled productivity, reaching 1,687 kg/ha. More broadly, farmers implementing multiple regenerative practices achieved a 37% increase in productivity.
These results point to a powerful opportunity: deforestation free production does not have to be an additional burden for farmers. It can become a driver for better farm performance. By combining market readiness with climate-smart and regenerative practices, farmers can meet evolving deforestation-free requirements while increasing productivity, strengthening livelihoods and building long-term resilience.