Assessing the sensitivity to climate change of the Berg River Voelvlei abstraction system, Western Cape Province, South Africa

Rabakali, Takalani

Assessing the sensitivity to climate change of the Berg River Voelvlei abstraction system, Western Cape Province, South Africa

Thesis / Dissertation

2025

Publisher

University of Cape Town

Department

Department of Environmental and Geographical Science

Faculty

Faculty of Science

Abstract

There is increasing evidence that climate change (CC) will negatively impact the water sector in many regions of the world, including South Africa, by reducing the availability of water resources in both dams and rivers, changing streamflow patterns, and disrupting aquatic biodiversity. The water resources of the southwest region of the Western Cape Province are particularly sensitive to anthropogenic CC, because the entire Western Cape Water Supply System (WCWSS) relies on rainfall-runoff to fill the main six renowned reservoirs, especially during the winter season. During the 2015-2018 drought event that occurred in the southwestern Cape (SWC) region around Cape Town, which ultimately lasted three consecutive years, these six reservors declined due to a shortage of rainfall that was exacerbated by anthropic factors associated with the growth of the urban population, as well as largely unabated demand amongst local water users. The records showed that rainfall in 2017 was the lowest since the 1880s, and subsequent analysis showed that human influence on climate had increased the likelihood of a drought of this magnitude in the SWC by a factor of three. Several options for augmenting supply in the WCWSS have been proposed, including the Berg River Vöelvlei Abstraction System (BRVAS), which will draw water from the Berg River (BR) in winter and pump it into the Vöelvlei dam (VD). However, there has been no analysis to date of the reliability of the BRVAS scheme under river flow conditions experienced in the recent 2015-2018 drought, or under projected climate change. This study aims to assess the reliability of the BRVAS to climate change. The two research questions for this study were as follows: (1) Given that the previous BRVAS reliability calculations made use of abstraction simulations using the earlier dry period of 1968-1972, would simulations using a period that spans the more severe 2015-2018 drought have the potential to alter the calculated reliability of the scheme? (2) How will the water available for abstraction change in the future under CC scenarios, and what are the implications of these changes for the reliability of the BRVAS? An abstraction model was set up in Excel that used historical streamflow (G1H020-Daljosafat streamflow gauge) and projected river flow data as input. The observed, extending from 1965 to 2021 were retrieved from the Department of Water and Sanitation (DWS) website. The CC model-driven runoff data came from an existing study on the future of water resources in South Africa for three different periods, the recent past (or baseline) from 1961-1990, the near future from 2015-2044, and the distant future from 2070-2099. For the observed period, analysis of the data using the G1H020 Daljosafat streamflow gauge historical data showed that the simulated abstraction amounts and estimates are reduced if one considers the recent drought of 2015-2018 that happened in the SWC region of South Africa. Similarly, for the near and far future, it was found that the water abstraction amounts available were reduced quite dramatically across a large majority of hydrological flows simulated using the six different climate change model outputs. The minimum annual abstraction amount in 2015-2018/19 was found to be 33% lower than the earlier period, while the median annual abstraction amount was 75% different. Similarly, for projections, the frequency of flows below the 25th percentile (Q1) in all models ranged from 1% to -13% in the near future of 2015-2044, and from -1% to -39% in the distant future of 2070-2099, whereas the median annual abstraction differences for the near future ranged from 0% to -6% and 0% to -29% for the far future. The abstraction of 95% of the time (FDCs) exceeded during the observational record and the range across the models in differences between the far future and the recent past was noted to be from -50% to -83 percent.