Making Deforestation-Free Supply Chain Monitoring a Reality in Smallholder Cocoa Systems

Cacao plant

SNV and Touton SA are proud to share news of progress made on a robust and cost-effective deforestation monitoring system during a workshop recently organised by Cocoa & Forests Initiative (CFI) in Accra, Ghana.

The system was developed and presented by remote sensing company Satelligence in collaboration with Ghanaian GIS experts.

With satellite data from the European Sentinel-1 and 2 satellites and Artificial Intelligence it is now possible to accurately differentiate monoculture cocoa, cocoa agroforest with various shade tree densities and forest, detect deforestation associated with cocoa and to pin point illegal cocoa farms in forest reserves in near real-time (NRT) so that appropriate remediation action can be put in place. It has also become possible to easily monitor renovation and rehabilitation (R&R) actions in cocoa landscapes.

So far, companies that have committed to eliminating deforestation in their supply chains – notably through their engagement in the Cocoa and Forests Initiative, were struggling to demonstrate measurable reduction in deforestation in their supply chains due to lack of appropriate and cost-effective forest monitoring systems. This complexity is due to the fact that cocoa is an understorey forest crop with a structure similar to many forest trees, which makes it difficult to delineate it from other forest types.

Processed Sentinel-2 input, containing modified Copernicus Sentinel data [2018]

Processed Sentinel-2 input, containing modified Copernicus Sentinel data [2018]

Classification result by Satelligence

Classification result by Satelligence

Our approach builds on a method developed in Ghana (*1) to detect and distinguish cocoa plantation from forests and other vegetation types. The combined use of 10m detail radar imagery that can better capture forest structure, and new machine learning techniques enabled us to achieve higher accuracy than previously possible. We engaged local knowledge including field visits and active support by national experts.

Persistent cloud cover and haze problem over West Africa was addressed by using Sentinel-1 radar that can ‘see’ through clouds and haze; and also by cloud removal and a tedious haze correction that cleaned up all usable pixels of Sentinel-2 collected every 5 days throughout the entire year 2018.

Farm data with seven land use classes produced by Satelligence

Farm data with seven land use classes produced by Satelligence

The system was trialled in the Juabeso-Bia Hotspot Intervention Area of the Ghana Cocoa REDD+ Programme with an estimated area of 243,500ha. The landscape hosts two forest reserves (Krokosua hills and Bia North forest reserves) and one National Park (Bia National Park).

As a next step, protocols for validating NRT deforestation alert systems and feedback mechanisms will be developed. This will be piloted by supply chain company actors in partnership with relevant institutional structures of Ghana’s Forestry Commission in close collaboration with community resource management units in order to generate appropriate response measures.


(*1)Daniel Tutu Benefoh et al (2018). Assessing land-use typologies and change intensities in a structurally complex Ghanaian cocoa landscape. Applied Geography (2018) 99:109–119; Kwabena Asubonteng et al(2018). Effects of Tree-crop Farming on Land-cover Transitions in a Mosaic Landscape in the Eastern Region of Ghana. Environmental Management (2018) 62:529–547

For further information, please contact: