LAND COVER MAPPING USING MULTI-TEMPORAL SATELLITE LANDSAT TM AND LANDSAT OLI/TIRS IMAGERY IN THE ARTISANAL ALLUVIAL TIN MINING AREAS OF CENTER AND SOUTH BANGKA, BANGKA BELITUNG ISLAND

• Main Author: El Hakim, Muhammad
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/43743
• Institution: Institut Teknologi Bandung

Bangka Island is known as the largest tin-producing area in Indonesia that lies within the metallogenic province of the Southeast Asian tin belt. With the decresing alluvial tin reserves for large-scale mining both on land and off shore of Bangka Island, mapping the artisanal mining to check the remaining reserves of alluvial tin becomes crucial. Land cover mapping throughout this area was conducted using multi-temporal Landsat satellite imageries to analyze land cover changes that suggest the artisanal mining activities. The target areas are in the center and in the south of Bangka Island from 2000 to 2017 and from 2004 to 2017, respectively, using Landsat TM and Landsat OLI/TIRS data. Several image processing methods were used to identify land cover changes over time, including vegetation indices, the maximum likelihood classification, and spectral angle mapper mapping (SAM). Multi-temporal mapping was able to identify several land cover types, such as vegetation/forest, barren land, built-up areas, agricultural land, and water bodies. The result for barren land was validated using X-ray diffraction (XRD), grain counting, and spectral analysis of field samples to distinguish overburden (original land) and tailing. SAM method successfully identifies the distribution of overburden and tailing areas in barren land which are the prospect for small scale mining. The multi-temporal results show an increase in barren land from 2000 to 2017 in some study locations as well as overburden and tailing areas which indicate of artisanal mining activity changes in the center and southern Bangka over time.