Digital satellite remote sensing for terrestrial coastal zone management

Pretorius, Cindy (2002-03)

Thesis (MA)--Stellenbosch University, 2002.

Thesis

ENGLISH ABSTRACT: The unique and often fragile environment of the coastal zone is placed under increasing pressure by human development. It is expected that three quarters of the world's population will be living within 60km of the coast by the year 2020. Thorough planning and management are required to prevent coastal degradation. In South Africa, coastal management efforts are being promoted through the implementation of a White Paper for sustainable coastal development. A Coastal Decision Support System (CDSS) was developed to empower local authorities to demarcate and manage sensitive coastal areas by giving them access to relevant botanical and physical information. Land cover/use information for the CDSS was previously mapped manually from hardcopy aerial photography. This method was found to be time-consuming and costly. This study investigated the potential for digital satellite imagery as primary source of data for populating the land cover/use information of the CDSS. A methodology was designed utilising semi-supervised isodata clustering for extracting relevant information for a study area covering 40x20km of coast along the southern coastal sector of South Africa. Digital mapping of SPOT 4 multispectral satellite data was used successfully to map land cover/use information such as wetlands, coastal lakes, dune vegetation, urban areas, forest plantations, natural forest and agricultural areas. A cost comparison was also made between the digital mapping method from satellite imagery used in this research project and the manual mapping from aerial photography. Digital mapping from satellite imagery was found to be more cost-effective in terms of both data and human resource costs. The method outlined and discussed in the research project should provide sufficient guidance for future application of the techniques in populating the CDSS with land cover/use information.

AFRIKAANSE OPSOMMING: Die unieke en dikwels sensitiewe landskap in die kussone is onder aansienlike druk weens ontwikkeling deur mense. Daar word verwag dat 'n derde van die wêreldbevolking teen die jaar 2020 binne 60km van die kus woonagtig sal wees. Dit sal deeglike beplanning en bestuur verg om die agteruigang van hierdie gebied te bekamp. Kussonebestuur word in Suid Afrika aangemoedig deur die implementering van 'n Witskrif vir volhoubare kussone-ontwikkeling. 'n Kusgebied-besluitnemingsondersteuningstelsel (KBOS) is ontwikkel in 'n poging om plaaslike owerhede te bemagtig om sensitiewe kusgebiede af te baken en te bestuur. Die KBOS verleen plaaslike owerhede toegang tot toepaslike inligting oor botaniese en fisiese o~standighede. Grondbedekkinginligting vir die KBOS is in die verlede vanaf hardekopie lugfoto's gekarteer. Hierdie metode is tydrowend en duur. Die potensiaal van digitale satellietbeelde as hoof databron om grondbedekkinginligting vir die KBOS te voorsien is in hierdie studie ondersoek. 'n Metode word in die tesis uiteengesit om 'semi-supervised isodata clustering' te gebruik om die nodige inligting uit die data te onttrek. Die studiegebied sluit 'n area van 40x20km langs die suid kus van Suid Afrika in. Digitale kartering vanaf SPOT 4 multispektrale satellietdata is suksesvol gebruik om grondbedekkingsinligting soos vleilande, kusmere, duin-plantegroei, stedelike gebiede, bosbou, natuurlike bos en landbougebiede te karteer. 'n Kostevergelyking is gedoen tussen die digitale karteringsmetode vanaf satellietbeelde in vergeleke met handkartering vanaf lugfotografie. Die digitale karteringsmetode blyk meer koste-effektief te wees beide in terme van die datakoste sowel as die koste verbonde aan mannekrag. Die omskrywing van die metode in die tesis behoort as goeie riglyn te dien vir die toepassing van die tegniek om grondbedekkinginligting voor te berei vir die KBOS.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/52804
This item appears in the following collections: