Die lokalisering van bodemdegradasie in Noord-Transvaal met behulp van satelietafstandwaarnemingstegnologie

Abstract

Satellite data are used to gather regional scale soil degradation data. These data are necessary for the planning and management of the environment. The spatial occurence of soil degradation was needed by the South African Development Trust Corporation in order to determine priorities for water control planning in the various districts of the former Lebowa. The PU for CHE helped completing the survey by gathering the soil degradation locations through image interpretation of satellite data. Soil degradation data were gathered by means of remotely sensed satellite technology (RSST). The use of RSST concerns two types of data needed viz. satellite data and ground truth data. The satellite data used were Landsat Thematic Mapper (TM) data. Landsat TM was chosen due to the fact that the spatial resolution of these data could easily be applied to the need for regional data. The satellite data were classified on an IBM High Level Image Processing System using the Bayes Maximum likelihood classification algorithm. This image processing system was made available by the PU for CHE. Ground truth data were gathered during various field trips and mapped on 1 :50 000 maps. These data are used for training and testing during and after image processing. The data derived from the satellite were then compared with the ground truth data to assess the accuracy of location of soil degradation. Accuracy assessment was undertaken by the Chi-squared test and the Kappa test, among others. The overall accuracy of location was assessed to be 69%. This number exceeded the initial goal of 51%. According to the results from the satellite data, approximately 33% of the total area of Lebowa proved to be degraded. This percentage could even be higher if it is taken into account that this number includes the mainly uninhabited and non-degraded mountainous areas of Lebowa. The results , however, present environmentalists with useful regional scale information to be used during planning on a local scale.