rogerpanel/Cameroon-Saphire-analysis

Research analysis of Cameroon Saphire # Cameroon-Saphire-analysis Reproducible remote-sensing assessment of the **environmental footprint of artisanal sapphire mining and trade in Cameroon's Adamawa region** — land-cover change, land degradation, and sediment/turbidity dynamics — using **only publicly available datasets** and a **DPSIR** (Drivers–Pressures–State–Impact–Response) framework. This repository is the analysis codebase accompanying the manuscript: > *Environmental footprint of artisanal sapphire mining and trade in Cameroon's > Adamawa region: land degradation, sediment dynamics, and prospects for > sustainable rehabilitation.* (in preparation) It is intended to be fully inspectable by reviewers, and reproducible **with or without** a Google Earth Engine account. --- ## Scientific scope Sapphire (corundum) mining differs fundamentally from gold ASGM: it is **mercury-free**, so the environmental signature is **physical** — vegetation clearance, soil disturbance, pit/tailings creation, and downstream sedimentation — rather than heavy-metal contamination. The analysis quantifies that physical footprint at three Adamawa/North/NW placers (**Mayo Kewol**, **Marma/Tibati**, **Mayo Kila**) from Sentinel-2 (2016–2025), with a Landsat baseline for context. Headline indicators (DPSIR *State/Impact*): - **Disturbed/mined area trajectory** (Random Forest land-cover classification) - **Forest loss** timing vs mining onset (Hansen Global Forest Change) - **Relative surface-water turbidity** (NDTI) upstream vs downstream, controlled for rainfall (CHIRPS) - **Bias-adjusted areas with 95 % confidence intervals** (Olofsson et al. 2014) --- ## Repository layout ``` config/ sites.yaml (study-site AOIs) + datasets.yaml (GEE asset registry) src/ analysis modules (numbered = pipeline order) 00_feasibility_check.py visual footprint confirmation (run first) 01_build_composites.py cloud-masked dry-season Sentinel-2 composites 02_spectral_indices.py NDVI, NDWI, MNDWI, NDTI, BSI 03_ …