Potential impacts of climate change on hydrological extremes in the Incomati River Basin

Master Thesis


Permanent link to this Item
Journal Title
Link to Journal
Journal ISSN
Volume Title
Climate change has been shown to influence extreme rainfall and flooding events over many river basins, yet there is a dearth of information on how to mitigate future risks and vulnerabilities in the Incomati River Basin (IRB), a basin known for extreme devastating flood events. This thesis investigates the potential impacts of climate change on extreme hydrological events that induce flood in the Incomati River Basin (IRB). A series of climate and hydrological simulation datasets were analysed for the study. The climate simulation datasets were acquired from the Global Meteorological Forcing Dataset (GMFD) and the CO-ordinated Regional Downscaling Experiment (CORDEX), but the hydrological simulation datasets were generated with the latest version of the Soil and Water Assessment Tool (called SWAT+), using GMFD and CORDEX as the climate forcing data. The CORDEX dataset was biased-corrected with GMFD, using the Quantile Delta Mapping (QDM) method. The SWAT+ was calibrated and evaluated over the basin to investigate the role of objective functions in SWAT+ calibration, four sensitivity experiments were performed using four objective functions (hereafter, 1-NSE or RMSE, 1-R 2 and PBIAS). To study the influence of the bias correction of CORDEX on hydrological simulations, the SWAT+ simulations were performed using the original and biased-corrected CORDEX datasets as the climate forcing. The impacts of climate change on the mean hydroclimate variables and on characteristics of extreme hydrological events (i.e. the intensity and frequency of extreme precipitation and streamflow events) were examined at four global warming levels (i.e. GWL1.5, GWL2.0, GWL2.5, GWL3.0) under the RCP8.5 future climate scenario. The results of the study show that SWAT+ gives realistic simulations of hydrological processes in the basin, although with notable biases in the simulated streamflow. The SWAT+ calibration over the basin is sensitive to the choice of objective function for the calibration. The calibration converges faster with 1-NSE or RMSE than with R2 or PBIAS. The performance of SWAT+ in simulating the streamflow over the basin depends on the statistical metrics used in the evaluation, while the NSE of the model SWAT+ simulation is poor (i.e. NSE ≈ -0.08) over all the stations, the PBIAS is very good (i.e. PBIAS ≈ 13.7%) at some stations. The bias correction of CORDEX datasets substantially reduces errors in the climate datasets and improves the quality of SWAT+ simulations over the basin. Moreover, it also reduces the level of uncertainty in the simulations. With global warming, a future increase in temperature is projected over the basin, but a decrease in annual precipitation is indicated over most part of the basin except at the south-west tip of the basin (i.e. around Nooitgedacht Dam), where precipitation is projected to increase. The changes in hydrological extreme events generally follows the precipitation pattern, in that, while less intense and less frequent extreme precipitation and streamflow events are projected over most parts of the basin, more intense and more frequent precipitation and streamflow are indicated in the vicinity of the dam. However, the projection also suggests that an increase in extreme precipitation and streamflow activities surrounding this water body could induce extreme streamflow events downstream of the basin. The results of this thesis have applications in mitigating the impacts of climate change on extreme hydrological events in the basin.