Remote sensing detection of mining sites at regional scale: a case study in Sardinia

The study of mining sites represents multidisciplinary research at the intersection of earth sciences, environmental monitoring, and technological innovation. Mining sites, whether active or abandoned, are focal points of human activity with profound impacts on landscapes and ecosystems. In various historical mining regions across Europe, where deposits exploited for centuries are now deemed depleted, recent attempts have emerged to assess the residual potential of these areas in relation to Critical Raw Materials (CRMs) resources. Sardinia, as a prominent illustration of such regions, has played a pivotal role as Italy's main mining center. Specifically, the districts surrounding Iglesias and Guspini–Arbus in southwestern Sardinia have historically served as the epicenter of the mining industry. From 1850 to 1995, these areas witnessed substantial extraction activities focused on zinc and lead ores. Our comprehensive study aimed at mine mapping in a vast area of 80 mines demonstrated the effectiveness of a systematic approach involving multiple images processing techniques, notably utilizing Sentinel-2 imagery. The application of vegetation removal through NDVI, band ratio analysis, and Principal Component Analysis (PCA) proved instrumental in refining mine detection accuracy. Real mine data, incorporating information on materials, activity status, and years of operation, played a crucial role in calibrating and validating the results. The correlation between GIS-visualized activity status and detectability highlighted the method's dependency on this feature. While acknowledging the complexity and uniqueness of mine mapping initiatives, our study underscores the significance of tailoring methods to specific geological and spatial characteristics for optimal results in mineral detection and mapping.