Sustainability Policy for Smallholder Farmers: An Opportunity for Digital Inclusion

At the Alliance, technological innovation for climate-resilient farming is a growing area of work: not only to provide farmers with the latest weather predictions and tailored recommendations on varieties and practices (within our work on climate information services), but also for farmers to estimate yields and monitor vegetation cover. While these tools were initially developed for farmers to prepare for market and monitor land health, these tools could also support farmers in overcoming the challenge of demonstrating their coffee plantations in areas with dense tree cover. 

“To create more inclusive EUDR-compliant value chains, we can bring together lessons learned from the Alliance's work across the value chains of other products subject to EUDR, finding ways to secure producers’ market access and deliver on the ‘improved livelihoods’ objective that is indeed present in the regulation.”

According to Bunn, there are two urgent challenges to address for fair and accurate compliance: demonstrating production in dense forest areas and ensuring accurate yield reporting to ensure that farmers are not forced to illegally combine their harvests. 

To distinguish compliant from non-compliant farms, plot-level georeferencing data is put through monitoring systems to establish whether the plot had forest cover in 2020 (the baseline year), with crops grown on land deforested since 2020 being blocked from imports to the EU. However, this method of evaluating compliance creates three issues: First, many organizations – including governments, NASA, local NGOs, universities, etc. – each produce land cover data using different tools on different dates that generate varying results, meaning that there is no one point of reference to identify deforestation; second, the lack of capacity of distinguishing agroforestry plantations from forest puts both smallholder farmers’ livelihoods and these sustainable farming systems at risk; and third, existing land cover datasets are likely to be biased against smallholders in favor of large producers, as large-scale producers have easier access to digital tools that demonstrate their production.  

To overcome the challenge of having one reliable dataset that detects small-scale plantations, the Alliance’s ‘Terra-i’ tool can contribute by using satellite data to evaluate remote sensing systems. Terra-i is a monitoring system that detects changes in vegetation cover in near-real time. Gradually being upscaled across the tropics, Terra-i tool has been applied in Peru, Colombia, Honduras and Vietnam, where the tool’s ability to monitor anthropogenic land-use change has been adopted by national actors to increase sustainability.  

As explained by Bunn:

“Unlike private sector tools, Terra-i's methods are inclusive of smallholder farmers, clearly documented, and developed together with local organizations. Terra-i demonstrates the urgent need to create smallholder-inclusive reference data, which can be used as a standard to evaluate other tools. Building capacity among national stakeholders to enable them to build land cover datasets which reflect their own regulations, local agro-ecological and social conditions should be another area of focus. This way, smallholders and their institutions can be empowered to contest exclusionary foreign data.” 

Furthermore, there are concerns that the EUDR will not actually curb deforestation, because to overcome the challenge of demonstrating their plantations, some farming communities collaborate by identifying a compliant plot on land cover maps, and various nearby farmers combine their harvests and exporting their joint production as if all the coffee was produced on the one compliant site. To overcome this issue, common practice is to create records of producers’ sales over several years to detect sudden jumps in production. Alternatively, for example for certified coffee, yield estimates are calculated by manually counting cherries on the trees. This work is laborious, and estimates are often miscalculated due to human error, or deliberately avoiding cumbersome counts of well-bearing plants. 

However, digital tools can reduce the time investment and human error in estimating yields, one example being the Alliance’s ‘Croppie’ mobile app: an accessible tool that makes counting easier and keeps a record of results, thus allowing farmers with no digital records to demonstrate their production and therefore overcome challenges with EUDR compliance. 

Croppie is a picture-based mobile app that uses AI to estimate yields based on images uploaded by the farmers together with local geographical and farming systems information. The app calculates expected outputs by counting coffee cherries on the images alongside data such as the number of productive trees and the size of the plantations. If all coffee-producing farmers could provide yield estimates – including smallholders with shade-covered plantations – all farmers would be able to prove compliance and export their harvests independently, rather than relying on collaboration with nearby plots to sell their produce. 

To secure farmers’ livelihoods in the face of changing markets and regulations, the Alliance’s Croppie team collaborates with cooperatives to reach farmers, share the tool and improve sales planning; works with financial institutions to facilitate farmers’ access to credit; and with exporters to increase the traceability of coffee value chains.