Abstract:
Expansive soils generally pose a significant geological hazard, which causes damage to civil infrastructure that could amount to billions of rand every year. Traditionally, expansive soils have been mapped and characterized using methods such as X-ray Diffraction (XRD) and standard engineering tests such as Atterberg limits. These traditional methods have proven to be time consuming, labour intense and expensive. Remote sensing has shown potential as a faster, cheaper non-laborious complementary tool that will support traditional methods of mapping expansive soils and geotechnical investigation. Notwithstanding the significance of remote sensing techniques and their applications in mapping expansive soils, research in this field is at its infancy in South Africa. The purpose of this study is to map expansive soils by use of multi-disciplinary techniques i.e., identify expansive soils using field observations, laboratory techniques and multispectral remote sensing. Soil samples (53) were collected from Brits in the North West Province. The soil samples were classified into swelling and non-swelling soils using laboratory-based techniques. In this regard, a) X-ray Diffraction (XRD) was used for mineralogical identification and classification, b) Atterberg limits tests were used to estimate swell potential of soils and support soil classification, and c) spectroscopic measurements of soil were acquired under laboratory conditions using a hand-held Analytical Spectral Device (ASD). Results show that expansive soils are dominant in the study area, with smectite being the most dominant clay mineral. Analysis of malleability demonstrated that soil plasticity varies from non-plastic to very high plasticity. Soils with medium to very high plasticity are in majority. Using an ASD that collects data in the 350-2500nm wavelength region, swelling soils were identifiable in the SWIR region of the electromagnetic spectrum. Mean reflectance spectra of swelling and non-swelling soils were used to map these soils on ASTER imagery, using SAM classifier. In spite of the difficulties of acquiring enough ASTER images, mapping of expansive soils using remotely sensed imagery demonstrated potential of studying the distribution of expansive soils as well aid in geotechnical investigations with inherent spatial-temporal resolution merits.
