Sediment Yield Potential in South Africa's Only Large River Network without a Dam: Implications for Water Resource Management

Abstract

Soil erosion not only involves the loss of fertile topsoil but is also coupled with sedimentation of dams, a double barrel problem in semi‐arid regions where water scarcity is frequent. Due to increasing water requirements in South Africa, the government is planning water resource development in the Mzimvubu River Catchment, which is the only large river network in the country without a dam. However, previous soil erosion studies indicate that large parts of the catchment are severely eroded. These studies, nonetheless, focussed only on rill and interill erosion processes and provided insufficient information about sediment yield. The present study maps and models the sediment yield potential by considering important connectivity factors (flow path length, slope steepness and surface roughness) over a 5‐year timeframe (from 2007 to 2012). Sediment yield contribution from rill–interrill erosion was modelled with ArcSWAT‐2012, whereas gully erosion contributions were estimated using sequential gully mapping with SPOT 5 imagery followed by gully‐derived sediment modelling in a GIS. Integration of the rill–interrill and gully results provided the total sediment yield potential, with an average of 5100 Mg km−2 y−1. The average sediment yield where the irrigation and hydropower dams are planned is around 20 000 Mg km−2 y−1. This study provides relative differences of the potential sediment yield and a framework to predict sediment yield to assist area‐specific rehabilitation in large, ungauged catchments. Without catchment rehabilitation, the life expectancies of the future irrigation and hydropower dams could be 55 and 43 years, respectively. Copyright © 2017 John Wiley & Sons, Ltd.

The impact of human behaviour and restoration on the economic lifespan of the proposed Ntabelanga and Laleni dams, South Africa: A system dynamics approach

“…This difference can be explained as the trapping efficiencies used were not assumed to be 100% throughout for both dams with monthly variation in trap efficiency values randomly sampled from a uniform distribution for some range of trap efficiency values. The estimates of dam lifespan for the Laleni dam differ considerably from the 43 year reported earlier [45].…”

“…A baseline reading from the model, that is the reading without allowing for the Monte Carlo analysis, indicates a lifespan of 57 years when no attempt is made to correct for the impact of increasing populations and incomes on sediment yields over time. The outcome differs slightly from the 55 year calculation mentioned earlier [45]. This difference can be explained as the trapping efficiencies used were not assumed to be 100% throughout for both dams with monthly variation in trap efficiency values randomly sampled from a uniform distribution for some range of trap efficiency values.…”

“…An explanation for this is that the sediment yield rate used to calibrate the model was derived from observations from the Ntabelanga dam catchment and applied for the entire Mzimvubu river catchment [51]. The earlier study focussed on differences between the Ntabelanga and Laleni dam catchment, the latter of which was found to have a sediment yield much lower than the first [45]. This highlights the dependency of these lifespan estimates on sediment yield figures and the pitfalls of estimating lifespans where data is scarce.…”

“…It has a total surface area of approximately 150ha [43] and contributes roughly R6 783 215 (~US$522 000) per annum to marine fisheries [44]. This return is under threat since the catchment produces high sediment yield rates, is considered to be in a poor condition and is subject to probable future deterioration [45,46].…”

“…None of the official documents related to the MWP report on the projected lifespans of the dams [39,1,47,49,43]. An external study focussing only on possible soil movements, however, suggests that the lifespan of the Ntabelanga dam is 55 years while the Laleni dam is 43 years [45]. Land use data is mentioned as an important factor in most of these studies and is used in the calculation of dam lifespan and sedimentation.…”