Analysing multitemporal vegetation density in the upper Molopo river catchment using remote sensing techniques

Abstract

In arid and semi-arid environments like most of South Africa, the state of vegetation density in catchments is an important indicator of the state of the environment, particularly because vegetation influences water availability by encouraging groundwater recharge. Because of the scarcity of water and consequent limits in abundance of vegetation in healthy green condition, in addition to climatic pressure, vegetation density in semi-arid environments are under human use pressure. The upper Molopo river catchment area (UMRCA) in the North West Province of South Africa is under this combined human use and rainfall pressure. This study aimed at assessing long-term changes in vegetation density in the upper Molopo river catchment area, resulting from anthropogenic and rainfall pressures. Four Landsat images were utilized in analyzing the vegetation density change. For purposes of interpreting the changes identified, ancillary long-term data on anthropogenic factors (human population, number of houses, household use of wood as energy source, livestock populations) and rainfall were obtained from state sources. Vegetation density on the images was enhanced using the Normalized Difference Vegetation Index (NDVI). The training data was then used for supervised maximum likelihood classification of the images into separate LULC as well as vegetation density (low, medium, high) maps for each elate. Assessment of accuracy indicated high classification accuracy of over 80%. The errors in classification were mainly clue to spectral signature confusion for the LULC classes. Change detection was then performed using the post-classification comparisons technique. Results indicated a growth in built up area from 3% in 1989 to 16% in 2013. The main indication of disturbance to .the vegetation was a sustained decline in medium vegetation density and its replacement by low vegetation density, particularly within 5km of human settlements. Hence, the study concluded that anthropogenic factors were the main cause of the decline in vegetation density in the UMRCA. Rainfall showed a cyclical pattern, with low seasonal rains (and high, negative Standardized Precipitation Index (SPI) values in the mid-1980s and positive SPI values thereafter, indicating that in the image period of analysis had wet conditions in general. Differences in rainfall prior to image date accounted for most of the inter-date variation in the vegetated LULC classes as well as water bodies. Long-term rainfall pattern did not have direct impact in the decline of medium vegetation density class in the UMRCA. However, there was a statistically significant negative correlation between human population and area of cover by medium vegetation density (r =-0.960, P < 0.01), implying that as human population increases, the medium vegetation density declines in area of cover. The decline in medium vegetation density in the Upper Molopo River Catchment is of ecological concern. There is a need for short term and long term strategies to ensure sustainable land management in the catchment area, and in order to preserve vegetation density and biodiversity in their natural state.