Assessing Cocoa Landuse Change and Exposure to Artisanal and Small-Scale Gold Mining Using Satellite Imagery

MCML Authors

Xiaoxiang Zhu

Prof. Dr.

Core PI

Data Science in Earth Observation

Abstract

Cocoa and gold are economically vital commodities that continue to drive the expansion of agricultural and mining frontiers. However, integrated assessments that jointly examine landuse patterns, cocoa exposure, and associated environmental and economic risks remain limited. This study assesses the spatial risk posed by artisanal and small-scale gold mining (ASGM) to cocoa landscapes by combining landcover reference data with spatial, environmental, and socioeconomic analyses. In parallel, we evaluate machine learning and deep learning models to map cocoa and ASGM land use from satellite imagery as an independent demonstration of current mapping capabilities. Among the tested models, the Prithvi-EO-2.0 geospatial foundation model achieved the highest accuracy in delineating cocoa and ASGM areas, outperforming U-Net and a random forest baseline. Using independently derived landuse reference layers, spatial analysis between 2016 and 2022 indicates that direct cocoa-to-ASGM conversion is subtle. However, substantial portions of cocoa landscapes coincide with zones of elevated ASGM-related mercury emissions. A cocoa risk index integrating proximity to ASGM sites, proximity to observed cocoa–ASGM transition areas, and terrain slope identifies more than 80% of cocoa area as falling within high-risk zones. These findings emphasize priority landscapes where mining pressure and environmental exposure converge, underscoring the need for targeted regulation and environmentally sustainable ASGM practices to protect cocoa production.

article OZK+26

---

IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing

19. Mar. 2026.

---

Authors

S. Ofori-Ampofo • A. Zappacosta • R. S. Kuzu • P. Schauer • M. Willberg • X. Zhu

---

Links

DOI

---

Research Area

 C3 | Physics and Geo Sciences