Remote sensing survey of archaeological sites in the Shashi- Limpopo Region
Date
2020
Authors
Thabeng, Olaotse Lokwalo
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The African continent is rich with archaeological heritage, which needs to be preserved for the current and future generations. The majority of archaeological heritage sites in Africa are facing disappearance due to a number of challenges including looting, destruction from developments, expansion of agricultural land and natural hazards. Documentation and monitoring of archaeological heritage sites, therefore, is of paramount importance for effective site management and preservation. However, archaeological heritage sites in the continent are poorly documented and monitored due to a number of factors including lack of funds by heritage management institutions, lack of trained personnel and inaccessibility of some areas due to conflicts or land ownership rights. Traditionally, the documentation and monitoring of archaeological heritage sites in Africa have been done through fieldwork, which is costly, time-consuming and difficult to carry out over large areas. Remote sensing offers a relatively fast, cheap, systematic and reproducible method of surveying and monitoring archaeological sites over large and/or restricted areas. Remote sensing techniques are used to identify earth surface features based on their spectral signature, which is the variation of reflection or emittance of materials’ electromagnetic energy. Spectral signatures for identifying archaeological sites are not universal, and an assessment of the applicability of remote sensing techniques in different archaeological landscapes is needed. The aim of this study, therefore, was to investigate the potential of using remote sensing techniques to document archaeological sites previously occupied by farming communities, which are traditionally associated with the Iron Age period in Southern Africa, using the Shashi-Limpopo case study.
The first part of this study gives a review of the use of remote sensing in the African archaeological context. Despite it being a fast, cost-effective and systematic method of survey, the results of this study have demonstrated that remote sensing is not widely used in archaeological applications in Africa. The aforementioned situation calls for studies investigating the potential of using remote sensing techniques to fast track archaeological site survey, documentation and monitoring in the continent.
The chemical composition of materials characterising different features have more or less subtle variations that, in turn influence the spectral behaviour of soil. This is an important principle that can be used for distinguishing archaeological soils from non-archaeological soils and can potentially help in discriminating different archaeological signatures. As such, the second part of this study investigated the possibility of using field spectrometer measurements to discriminate middens, non-vitrified dung, vitrified dung and non-sites (natural soils) characterising archaeological landscapes previously occupied by farming communities. It then investigated the presence of differences in the chemical composition of elements between middens, non-sites, vitrified dung and non-vitrified dung. The findings indicated that there is a statistically significant difference in the concentration of soil elements between non-sites, middens, vitrified dung and non-vitrified dung byres. They also indicated that some bands in the visible and shortwave infrared regions of the electromagnetic spectrum important bands for predicting the aforementioned archaeological sites and non-archaeological sites.
In the third part of this study, the ability of multispectral sensors to discriminate archaeological and non-archaeological features in Shashi-Limpopo confluence area was investigated using field spectral data resampled to the spectral resolutions of common multispectral satellites namely GeoEye, Landsat 8 OLI, RapidEye, Sentinel-2, SPOT 5 and WorldView-2. This is because the spectral and spatial resolutions of various multispectral sensors determine the size and the type of archaeological data a sensor can detect. As such, another goal of this study was to identify multispectral sensors with the optimum spectral resolutions for detecting middens, non-vitrified dung, vitrified dung and non-sites. Additionally, the performance of advanced classification algorithms (random forest and support vector machines) in discriminating middens, non-vitrified dung, vitrified dung and non-sites was also investigated. The results proved the possibility of using multispectral satellites in mapping middens, non-sites, vitrified dung and non-vitrified dung sites. These results initiated the need to upscale the test to actual satellite images.
The fourth part of this study assessed the possibility of prospecting for archaeological sites previously occupied by farming communities in the Shashi-Limpopo Confluence Area, using a very high-resolution satellite WorldView-2 image. The findings have shown that WorldView-2 satellite images and advanced classification algorithms can be used in prospecting for archaeological sites previously occupied by farming communities in Shashi-Limpopo Confluence Area.
Finally, the ability of geographic object-based image analysis (GEOBIA) based on random forest and support vector machines, to discriminate archaeological and non-archaeological features on a very high-resolution satellite WorldView-2 image was investigated. The results of this study demonstrated the robust ability of the GEOBIA to integrate spatial attributes into the classification model improves the chances of separating materials with limited spectral contrast.
Generally, this study has shown that remote sensing techniques can be used to map archaeological landscapes characterised by middens, non-vitrified dung, vitrified dung and non-sites. This will help archaeological heritage managers and researchers to document and monitor sites in archaeological landscapes characterised by the aforementioned features in a fast, systematic, reproducible and cost-effective manner
Description
A thesis submitted to the Faculty of Science, University of the Witwatersrand, Johannesburg, in partial fulfilment of the requirements for the degree of Doctor of Philosophy (Geography and Environmental Science), 2020