Mapping the changes in vegetation spatial extent within the Khayelitsha wetlands, Western Cape province, utilizing remotely sensed data
Abstract
Urban wetlands play an important role in providing ecosystem services and supporting biodiversity as a habitat. These ecosystem services include reducing severe impacts of floods by helping slow the overland flow amongst other services. However, despite the importance of wetland ecosystems and their services, their value and role across the board, is under threat from anthropogenic, and climate change-related events. Rapid urbanization and human encroachment are the major drivers of wetland vegetation fragmentation which leads to their degradation in urban areas. To prevent further destruction of urban wetland areas, it is essential to develop robust methods for inventorying their spatial distribution, and Land Use Land Cover (LULC) types. This information is important for inform decision- making and formulation of long-term strategies for wetland conservation. In this regard, this study sought to estimate changes in the spatial extent of the Khayelitsha wetland between the years 2000 - 2023 using freely available remotely sensed data obtained from Landsat 7 Enhanced Thematic Mappper Plus (ETM+) and Landsat 8 Operational Land Imager (OLI). By using satellite imagery and wetland fragmentation analysis techniques, this study sought to understand the patterns of wetland vegetation fragmentation during the years 2000, 2010, and 2023 as a proxy for assessing wetland degradation. To address the main objective this study (1) conducted a systematic review of the literature on the progress, gaps, and opportunities in the application of earth observation data in assessing and mapping changes in the spatial extent and productivity of wetland species, 2) assessed the performance of Support Vector Machines (SVM), Naïve bayes (NB) and Random Forest (RF) machine learning algorithms mapping wetland land use land cover types during the years 2000, 2010, and 2023, and 3) compared the performance of various vegetation Indices in classifying urban wetlands during the years 2000, 2010, and 2023 and assess the LULC fragmentation, thereof.