The Influence of Anthropogenic Climate Change on the 2015-2017 Hydrological Drought in the South-Western Cape, South Africa

Hall, Andrew

The Influence of Anthropogenic Climate Change on the 2015-2017 Hydrological Drought in the South-Western Cape, South Africa

Master Thesis

2021

Abstract

The Western Cape Province in South Africa recently experienced below-average rainfall during the period 2015−2017, this resulted in a three-year compound hydro-meteorological drought event in the Province. The 2015−2017 Western Cape hydro-meteorological drought was the worst drought event since 1904 and caused severe unprecedented water shortages throughout the Western Cape region, with many municipal water supply systems close to failure by the first quarter of 2018; most especially the Western Cape Water Supply System that serves Cape Town. The drought gained a lot of interest from the public, media and climate scientists alike. The main aim of this study was to assess the extent to which human influence on climate from fossil fuel emissions has changed the likelihood of a hydrometeorological drought event with the magnitude of that experienced in 2015−2017 in the SouthWestern Cape. The Pitman hydrological model was set up for the Berg River catchment in a way that enabled multiple simulations with different rainfall inputs so that attribution experiments could be undertaken. The key differences to the standard Pitman model set up included: (i) constant abstractions, return flows, and land use conditions; (ii) reservoir and dam storages were set to reflect current storage volumes; and (iii) extending the observed rainfall inputs to include the drought period. A hydrological model evaluation was then undertaken, using updated streamflow gauging station data, to assess the ability of the Pitman model to realistically simulate runoff in the Berg River catchment. The model was deemed suitable for the purposes of this study in simulating runoff. To generate the climate attribution experiments, Coupled Model Intercomparison Project Phase 5 historical simulations (1861−2010) were merged with the Representative Concentration Pathway 8.5 greenhouse gas scenario simulations (2011−2100) of rainfall from 77 simulations From 42 models to create a long-term (150 years) time series. Attribution experiments were constructed by considering the average conditions in the 31 year period centred on the years of the event, i.e. 2002−2031 to represent current climate conditions and the period 1861−1890 to represent pre-industrial climate conditions. Five 150-year long stochastic time series of rainfall for each individual simulation were then generated conditioned on observed rainfall characteristics this was done to increase the sample size of the models available. These stochastic rainfall time series were then used as input to the Pitman model to generate outputs/realisations of runoff for a pre-industrial and current world; thus generating impact attribution experiments. To determine the role of anthropogenic climate change on the 2015−2017 hydro-meteorological drought in the South-Western Cape the risk-based approach was applied to the rainfall and runoff attribution outputs. The 2015−2017 meteorological/hydrological drought event was defined in terms of three-year mean annual rainfall/runoff received in the Berg River catchment and its individual 12 quaternary catchments. This event definition was used as a rainfall/runoff threshold in the attribution analysis for the 2015−2017 meteorological/hydrological drought in the South-Western Cape. The three-year minimum averages of rainfall/runoff were identified in each of the 150-yearstochastic time series generated from the 77 simulations; resulting in 385 values for both current and pre-industrial climates for rainfall and runoff. A normal distribution was then fitted to the 385 values of the current and pre-industrial rainfall/runoff. From this distribution, the probability of the current rainfall/runoff occurring, based on the defined threshold, was identified and compared to the pre-industrial time series to calculate the risk ratios of the Berg River catchment and its 12 quaternary catchments. Results show that the risk of the meteorological drought event occurring in the Berg River catchment was increased by a factor of 28.5, 95% confidence interval: 26.0−32.4, (but ranged from 11.5−41.0 in the individual quaternary catchments) due to anthropogenic climate change. The occurrence of the hydrological drought event in the Berg River catchment was found to be increased by a factor of 12.9, 95% confidence interval: 11.3−13.5 (2.7−61.0 in the quaternary catchments) due to anthropogenic climate change. The risk ratio for runoff was higher than for rainfall in the wetter southern quaternary catchments, while it was lower than for rainfall in the drier more northern quaternary catchments. Thus, the human influence on meteorological drought appears to have been amplified in those catchments most important to the Western Cape Water Supply System.

Keywords

Environmental and Geographical Science

Reference:

Collections

Masters

Full item page