An assessment of the spatio-temporal urban dynamics in the city of Tshwane, South Africa

• Main Author: Magidi, James Takawira
• Format: Others
• Language: en
• Published: 2018
• Online Access: https://hdl.handle.net/10539/25838

A thesis submitted to the School of Geography, Archaeology and Environmental Studies, Faculty of Science in fulfilment of the requirements for the degree of Doctor of Philosophy in Geography. Johannesburg, May 2018. === Urbanisation, urban sprawl and loss of biodiversity in urban environments are major phenomena of the 21st Century cities and towns in both developing and developed countries. A study of the City of Tshwane (CoT), South Africa has shown that the city had been affected by unprecedented urbanisation, which led to encroachment of urban areas into non-urban environments. There is a need to monitor, quantify and predict urban dynamics for the sustainable management of urban environments. The advent of remote sensing and Geographical Information Systems (GIS) techniques have enabled researchers and decision-makers to have a historical perspective of the earth and detect change in urban areas. Remote sensing and GIS are powerful, cost-effective and efficient tools that are used in quantifying, monitoring and predicting land cover change using multi-temporal and multi-spectral spatial datasets. This helps decision-makers in designing decision support systems that are useful in evaluating alternative management scenarios and in the formulation of land use policies that are effective in the sustainable management of urban areas. Landsat TM (Thematic Mapper), ETM+ (Enhanced Thematic Mapper Plus) and OLI (Operational Land Imager) satellite imagery from 1984 to 2015 were used for the long-term change detection. These remotely sensed data were classified into two classes, which are built-up (urban) and non-built-up (non-urban) areas using the supervised maximum likelihood classifier (MLC) Post-classification change detection methods and landscape metrics were used to assess change and quantify the degree of urban sprawl. Short-term change detection was performed in the low, medium and high-density areas using classified SPOT (Satellite Pour l'Observation de la Terre) satellite imagery of 2008, 2012 and 2015. To predict future scenarios in urban dynamics the study made use of the classified land cover maps of 1986, 2005, 2009 and 2009 (Landsat TM and Landsat OLI) coupled with transitional areas, transitional probabilities and the Cell Automaton-Markov (CA-Markov) model. The prediction model was validated using the predicted maps and classified maps of 2009 and 2013. Change in vegetation was assessed using time series analysis, which was run on MODIS (MODerate-resolution Imaging Spectroradiometer) NDVI (Normalized Difference Vegetation Index) datasets with a 250m spatial resolution and a 16-day temporal resolution. Temporal (NDVI) profiles generated in different land cover classes coupled with the Mann-Kendall Statistic and Sen’s Estimator were used to assess the seasonal trends in vegetation from 2000 to 2016. Retrieval of change in land surface temperature (LST) was done using winter (August) and summer (December) Landsat imagery of 1997 and 2015. NDVI, emissivity and satellite temperature of the two different years and seasons were inputs in the retrieval of LST. There was a comparison of LST between the two years (1997 and 2015) and between seasons (winter and summer). Cross-sectional transects were run across different land cover types to show variations in LST. Results revealed an increase in urban areas in the CoT between 1984 and 2015. Urban predictions revealed an anticipated future increases in urban sprawl. Short-term land cover changes using SPOT imagery revealed an increase in urban areas in the high-density as compared to the low-density and the medium-density areas. Human settlements in the high-density areas especially the informal ones are also encroaching into areas earmarked for conservation. There were also remarkable seasonal variations in vegetation cover based on the MODIS NDVI temporal profiles. Mann Kendall trend analysis revealed a decreasing trend in vegetation cover in different land cover types. Temperature change in the CoT is evident as there was an increase in LST between 1997 and 2015 with high LST in summer and low in winter. The main aim of this study was to use remote sensing and GIS techniques to quantify, monitor and predict urban dynamics in the CoT. The objectives were to assess long-term and short-term land cover changes, to predict urban dynamics and to use available proxies such as vegetation cover, land surface temperature to assess urban growth. Keywords: Urban Sprawl, Urban growth, Predictive Modelling, GIS, Remote Sensing, Sustainable Development, Landscape Metrics, Land Surface Temperature, Time Series Analysis === LG2018