Sediment yield modelling in the upper Tsitsa Catchment, Eastern Cape, South Africa

Abstract

The Mzimvubu River is the largest river in South Africa without a dam. The Department of Water and Sanitation has identified the Tsitsa River Catchment on the Mzimvubu River as a potential site for a water resource development. The soils in the Tsitsa Catchment are prone to extreme soil erosion, in particular gullying. The sediment generated from these gullies and other forms of erosion will have a detrimental effect on any water resource development. Changing climate and land use will also affect soil erosion dynamics and thus need to be considered before any development is planned in the catchment. Previous studies have mapped the gully systems in the catchment as well as used hydrological models to determine erosion from sheet and rill processes. However, these studies did not account for the effects of change in land use or climate. The mapping of the gullies was also done manually, which is extremely time-consuming and is susceptible to human error. This study aims to determine the sediment yield in the catchment under current and future climate and land use scenarios as well as develop a methodology to identify and map the gullies automatically in order to determine the rate of gully growth from a time series of images. The study had two main components, the first was to study gully erosion in the catchment. In the second section the sheet and rill aspects of erosion under various climate and land use changes were modelled. Using object-based image analysis (OBIA) on SPOT 5 images a methodology was created to automate the task of gully mapping. This was applied to two SPOT 5 data sets one from 2007 and the other from 2012 in order to determine the rate of gully growth over the five-year period. Various accuracy assessments were also conducted to assess the accuracy of the methodology. It was determined that the methodology had an overall accuracy of 98% for the 2012 image and 99% for the 2007 image. There was an overall increase in gully erosion in the catchment by 28% in the five-year period. The estimated sediment yield generated from the gullies ranged between 7 and 14 t/ha/yr. It was concluded that OBIA resulted in faster processing times and more objective classification results. The second part of the study used the Soil and Water Assessment Tool (SWAT) to determine the sediment yield from sheet and rill erosion. SWAT only considers sheet and rill aspects of erosion and disregards gully erosion thus both methods needs to be incorporated into the study in order to understand the complete dynamics of soil erosion in the catchment. SWAT was used to model the current land use and climate scenario using land cover data and observed weather data for the 2007-2012 period. On average 0.18 t/ha/yr of sediment is generated in the catchment from sheet and rill erosion. Using climate data from 1969 and projected to 2100, future sediment yield from sheet and rill erosion was estimated. The effects of possible land use change on sheet and rill erosion was also estimated by changing the land use component in SWAT into various crops that may be cultivated in the catchment over the next century. The results of the land use change showed that the current land use is optimal for minimal sheet and rill erosion and converting to maize crops will have the greatest impact on sediment yield This study aimed to understand the dynamics of soil erosion under current and changing land use and climate scenarios. It was concluded that the majority of the sediment is derived from gully erosion, which accounts for up to 70 times more sediment yield annually than sheet and rill erosion. Gully formation and propagation in the catchment is of critical concern to any land or water developments proposed for the Tsitsa Catchment.