Cocoa predictive deforestation risk (COL, CIV, ECU)

This dataset presents the results of a predictive deforestation risk assessment for Colombia, Ecuador, and Côte d’Ivoire.
The results were produced using the cocoa_deforestation_risk tools (https://ciatgit.ciat.cgiar.org/Data_driven_sustainability_public/terra-…), an extension of the Forest Foresight (https://forestforesight.atlassian.net/wiki) package by WWF, developed in collaboration with the World Cocoa Foundation. This dataset was used to analyse how new layers, higher spatial resolutions, and different prediction periods affect prediction accuracy.
Methodology:Key Forest Foresight inputs were recomputed using country-specific rather than global forest maps. Additional metrics, such as metapopulation and forest-height data, were also included. Variants using longer prediction periods and higher spatial resolutions were prepared.
Forest Foresight was then run to compute a commodity-agnostic deforestation prediction. This prediction was subsequently correlated with existing cocoa datasets and cocoa prices to construct a commodity-specific deforestation risk prediction for cocoa.
The dataset presents the analysis results for prediction periods of 3, 6, 12, and 24 months at spatial resolutions of 400, 200, 100, and 30 metres.