Developing an integrated methodology for assessing risk and impact of illegal alluvial mining-related environmental stressors in the Pra river basin, Ghana.

Galamsey is a local Ghanaian term that describes small-scale illegal gold mining, primarily in rivers. This thesis investigates illegal alluvial mining-related risks and impacts in the Pra basin of Ghana using a spatially distributed GIS-based approach. Land use, river and geomorphol-ogy and environmental pollution were investigated using exclusively public data. Land use was assessed based on Sentinel-2 imagery from Dec. 2018 - Jan. 2020. Iso Cluster Unsupervised Classification in ArcGIS was conducted using a Modified Normalized Di˙erence Water Index (MNDWI)-based image stack to optimally detect mines. In total, 0.95% of the basin are cov-ered in open mining pits, resulting in approx. 2% of the basin if mining-related bareland is additionally considered. The proposed detection method showed high eÿciency compared to supervised classification, and an overall accuracy of 91.2%. River and geomorphological im-pacts of mining were assessed using the Universal Soil Loss Equation (USLE), channel stream power based on mean discharge and slope, as well as actual landscape and riverbed eleva-tion change derived from a DEM of Di˙erence between 2000 and 2015. USLE predicted low rates of erosion across the basin, but was shown to be an insuÿcient tool to model the dom-inant processes of gully and channel erosion which are exacerbated by alluvial mining. Mean landscape elevation change in the basin over 15 years was -0.87 m, with mining areas experi-encing a significantly elevated mean loss of -3.96 m in this period. All main channels showed significant erosion, with the largest riverbed elevation change taking place in the Lower Pra reach. Stream power indicated concordant behaviour, but requires further on-site verification before it can be used as an indicator. Environmental pollution was assessed in a structured literature review that identified over 2400 heavy metal concentration datapoints of 12 di˙er-ent metals in 168 locations and all environmental compartments with exception of ambient air. 36% of datapoints exceeded WHO recommended limits for the respective environmental compartment. Widespread heavy metal pollution was evident throughout the basin even in locations far away from mining sites, despite no or minimal other anthropogenic and geogenic sources in the basin. Groundwater was found to be heavily polluted with mercury, with 82.1%of measurements exceeding WHO recommended limits. Risk predictions were made based on existing surface water abstraction points in the basin using Pearson Correlation Analysis. Various potential predictors of pollution were investigated. Upstream fraction of mining land use and local channel slope were identified as the only significant predictors of pollution risk. Pollution pathway tracing was conducted to better understand the mechanisms of pollution transport throughout the basin. The analysis was conducted using results from the environ-mental pollution survey, additional literature review, a specifically developed pollutant routing algorithm, and 2D floodplain analysis. Overall, river-driven processes including contamination via floodplain inundation and aquifer recharge from polluted surface waters were shown to be the dominant pollution transfer processes in the basin, thereby providing a foundation for targeted future research and remediation strategies