Search

Abstract: The aim of this study is to explore the capability of hyperspectral remote sensing to map areas affected by Red Dust (RD) pollution on soil and river water. The chosen study area contains residues from the extraction of aluminum from bauxite in a mud pond (i.e. the red mud), which contain high concentrations of contaminants such as heavy metals, so that special precaution must be taken when disposing to avoid RD pollution of land surface and river water resources. We perform laboratory analysis and in situ measurements of RD polluted samples to recognize the dominant minerals and identify the optical characteristics of the samples and their spectral features. Both an unsupervised methodology and a shape-based spectral analysis technique, using the significant RD reflectance spectral features, were applied to airborne hyperspectral remote sensing data to map the RD distribution on bare soils. Whereas, to assess a tool for the detection of RD in river waters a semi-analytical model for the radiative transfer in water bodies was used. The results of the spectral shape-based analysis and of the semi-analytical model used for this study were consistent with ground truth data and hence support the application of hyperspectral technologies for a rapid detection and mapping of industrial wastes, such as the RD. This information is suitable to support the development of effective intervention policies and monitoring programs. [Copyright &y& Elsevier]