Characterising vegetation cover in relation to land use in the Inkomati catchment, South Africa, using Landsat imagery

Abstract

Vegetation cover in catchments in semi-arid environments with distinct rain seasons is important in influencing water quality and availability through its effect on runoff and groundwater recharge. This study assessed catchment-scale vegetation cover and condition, indicated by land degradation and vegetation vigour, using recent rain season Landsat TM/ETM+ images supplemented by dry season images for comparison. Image classification into vegetation density categories was used in mapping land degradation, the bare land category used as indicator. Vegetation vigour indicates vegetation productivity levels, and has been shown to correlate with the NDVI in similar previous studies. The 30 savanna biome vegetation types in the catchment were used as strata for assessing relative vigour using NDVI values. The NDVI frequency distributions per vegetation stratum were normal (Gaussian), the location of whose peaks enabled comparisons. Land degradation could be mapped with certainty mainly in the vicinity of settlements. Rivers draining through degraded areas had elevated colour and turbidity values. As the NDVI profiles per vegetation type could be explained by the land use and landscape position context of the vegetation types, NDVI profiling proved useful for assessing vegetation condition. NDVI profiling of vegetation types is potentially useful in monitoring savanna vegetation condition using remotely sensed images.