USING HIGH-RESOLUTION SATELLITE IMAGERY TO ASSESS LAND-USE AND LAND-COVER CHANGE ASSOCIATED WITH ARTISANAL SMALL-SCALE GOLDMINING IN GHANA

Abstract

As artisanal and small-scale gold mining (ASGM) expands globally in response to the growing demand for gold, there is a need to monitor and quantify the rapid land-use and land-cover changes associated with ASGM activity, particularly mining and agriculture, as they are often complementary livelihoods in communities. The use of remote sensing, with its frequent collection times and the availability of freely-accessible high-resolution imagery, is a powerful tool for analyzing land-use and land-cover change (LULCC) associated with ASGM. In this study, we utilized an interdisciplinary, mixed-methods approach, including image classification, community mapping, and interviews to determine how much land was converted to ASGM and the dominant land-use and land-cover transitions in two climatically and geologically distinct communities in Ghana between 2015/2016 and 2024. We used high-resolution (10 meter) multispectral Sentinel-2 imagery for our image classification. We also compared three different supervised classification methods, Random Forest (RF), Support Vector Machine (SVM), and Spectral Angle Mapper (SAM) to determine the best classifier for ASGM. Our results show that RF had the highest overall accuracy for almost all of the classifications, at 85% for 2015 and 87% for 2024 for the Southern site and 85% for 2016 for the Northern site. The exception was SVM outperforming RF for 2024 for the Northern site with an 86% overall accuracy. We report a 105% increase in mining for the Southern site and a 0.14% increase in mining for the Northern site. There was a surprising increase in agriculture at both sites, with a 2.4% increase in the South and a 45% increase in the North. The interdisciplinary approach used in this study provides a framework for integrating remote sensing methods and community-based social science methods to understand the complex LULCC associated with ASGM in these farming-mining communities.