Modelling Impacts of Land Use and Climate Change on Hydrology: A case study of Mokolo River Catchment in Limpopo Province, South Africa

Abstract

The Mokolo River Catchment (MRC) of the Waterberg District Municipality has been experiencing water security challenges which have attributed to water supply crises over the years. Due to its importance in water supply and development, quantifying water resources under changing climatic conditions is essential for effective water resources planning and adaptation. This study aimed to model the impacts of projected climate change on the hydrology of the MRC. Land use and Land Cover (LULC) change quantification was computed using the Semi-Automatic Classification plugin (SCP) in Quantum Geographic Information Systems (QGIS). The Inverse Distance Weighting (IDW) method was used for rainfall interpolation, while the Web-based Hydrograph Analysis Tool was employed for baseflow separation analysis. Mann-Kendall trend test was used to detect trends in temperature, precipitation, and streamflow. The QGIS interface Soil and Water Assessment Tool (version 2012) hydrological model was calibrated and validated for the MRC. Climate change projections were generated using five CMIP6 GCM-driven ensemble datasets framework in near future (2020-2049) and far future (2055-2084) were compared to the historical period (1993-2022) for the SSP2-4.5 (medium emission scenario). Delta Statistical Downscaling (DSD) and bias-correct adjustments were conducted using CMhyd at a spatial resolution of 0.22° (25 km). The LULC classification demonstrated 100% overall accuracy, with Kappa statistics of 0.99, 0.99, and 0.98 for 2000, 2010, and 2020, respectively, using the Landsat dataset. For 2017 and 2022, Kappa statistics of 1 and 0.99 were obtained using Sentinel-2 imagery, indicating a close match with reference data. The IDW method showed effectiveness as a spatial interpolation technique, with correlation coefficients ranging from 0.75 to 0.97. Observed and simulated Baseflow Index values were consistent and ranged from 0.22 to 0.31. Following calibration and validation results obtained using historical data set-up, SWAT demonstrated sufficient reliability for assessing the impacts of climate change on hydrology. The performance evaluation over the historical period showed R² values ranging between 0.52 to 0.88, NSE ranged between 0.05 to 0.81, PBIAS ranged between 2.07 to 68.44 and RSR ranged between 0 to 0.32. The MK results indicate a statistically significant increase in temperature and a significant decrease in rainfall and streamflow in the catchment over both near future (2020-2049) and far future (2055-2084) study periods. The streamflow magnitude decreased over time with annual average totals of 100 m³/s, 45 m³/s, and 38-40 m³/s for historical, near future, and far future, respectively. Projected annual average streamflow is expected to decrease by 21% in the near future and by 49% in the far future relative to the historical baseline period. Projected near-future to far-future seasonal reductions are expected to reach 39.71% in summer, 49.51% in autumn, 34.18% in winter, and 48.97% in spring. Results on projected temperature increases and reduction in streamflow should trigger the initiation of support for the development of effective adaptation measures to address the ongoing impacts of climate change and facilitate long-term water resource management planning in the Mokolo catchment. Keyw