Application of SWAT hydrological model in a mine catchment in South Africa

Abstract

A comprehensive rainfall−runoff model was developed for the B11C−B11E study area, encompassing three quaternary catchments (± 1 330 km2). The aim was to effectively oversee a catchment area, ensuring the fulfilment of fundamental human requirements and promoting the sustainable use of water resources amidst the impact of mining activities. Additionally, the aim was to minimise adverse environmental effects, such as large-scale dewatering. The study area, located in the Mpumalanga province, falls within the Vryheid Formation, renowned for significant coal reserves. Over the past century, extensive mining activities have taken place in this catchment, directly impacting water resources through pollution and local resource depletion. The choice of the Soil and Water Assessment Tool (SWAT) rainfall−runoff model was based on its effectiveness in simulating hydrological processes, especially in watersheds with limited data availability. The study area was divided into 15 sub-basins, and a Digital Elevation Model was employed to establish watershed boundaries, delineate sub-basins, define land use and soil data, and configure model parameters to ensure accurate representation of reality in the model. Observed data, including rainfall, flow, and water quality data, were collected for 18 years (1991-2008), covering the period with available data for most flow and rainfall gauges. The catchment, hosting various land uses (urban, agricultural, and mining), underwent changes over this period. To model the impact of mining effectively, a literature-derived approach segmented the time series data into three periods: a reference period, a middle period, and a recent period. In the calibration of the SWAT rainfall−runoff model, key parameters included runoff curve number, soil hydraulic conductivity, and deep aquifer percolation recharge. Calibration and validation were performed using the B1H021 flow gauge, located at the catchment outflow. The Nash−Sutcliffe Coefficient values of 0.73 and 0.77, along with coefficient of determination (R2) values of 0.82 and 0.77, indicated a strong correlation between simulated and observed values. The comparison of the reference and recent period models revealed shifts in the water balance influenced by weather fluctuations, conservation efforts, increased water usage, and mining impacts. Coal mining and other factors collectively contributed an average of 14.79 × 106 m3 annually, indicating a notable increase in water consumption. The study demonstrates that the SWAT model effectively simulates monthly runoff within the appropriate parameter range. While the model holds potential for future scenario analysis and water resource planning, caution is warranted.The employed model in this study area provides a range of essential benefits for achieving the primary aim of effective management. It enables the prediction and simulation of various scenarios related to water resources, aiding decision-makers in anticipating changes, and to plan accordingly. The model facilitates the assessment of mining impacts on the hydrological system, supporting informed decision-making and targeted mitigation efforts. Decision-makers can optimise resource allocation, identify and manage risks, and conduct scenario analyses to understand the potential effects of different variables. Additionally, the model aids in long-term planning, community engagement, regulatory compliance, and fosters data-driven decisionmaking by relying on accurate hydrological representations. The continuous feedback loop ensures that real-time data would improve the accuracy of the model over time, enhancing the overall effectiveness of water resource management strategies. The research successfully achieved its main aim of utilising hydrological modelling as a potential tool to explore the hydrology of a mining catchment in South Africa. The primary achievement was to provide valuable insights for water resource management in mining-affected areas.