A comprehensive scheme for lithological mapping using Sentinel-2A and ASTER GDEM in weathered and vegetated coastal zone, Southern China

Abstract

Abstract Lithological mapping in weathered and vegetated coastal zone is always a crucial but challenging issue for geological remote sensing. This study tested the applicability of Sentinel-2A and ASTER GDEM data for lithological mapping in the coastal zone of Shanwei district, southern China. Here, a comprehensive scheme for tuff and granite outcrops extraction and lithological classification was employed to process the remote sensing data: 1) Using the matched filtering with image-derived spectra as references, the weathered but representative outcrop-related patches were preliminarily extracted based on the fractal “DN-A” (digital number-area) algorithm; 2) The image masks of NDXI (NDVI, NDSI and MNDWI) in conjunction with spatial intersection, were conducted to minimize the influence of vegetation coverage; 3) Considering terrain features caused by differential weathering, slope analysis have allowed us to discriminate from tuff to granite outcrops; 4) A set of empirical verification criteria were established to assess the classification accuracy, which are exactly 82.4~83.2% of tuff and 91.3~92.3% of granite. The peak values of tuff and granite patch areas are 0–0.1 km2 (51%) and 0–0.5 km2 (60%), and 90% of patches are < 0.6 km2 of tuff and < 3.5 km2 of granite, respectively. The protolith of extracted weathered outcrops were further identified by several excavated trenches. This research has contributed a useful case study for remote-sensing lithological mapping in weathered and vegetated coastal zone, and the proposed approach should have a great possibility to be extrapolated for other similar cases.