Baseflow estimation using streamflow hydrograph data in South Africa

Abstract

The reason we have perennial rivers running all seasons of the year is mainly due to baseflow. Baseflow can be defined as a groundwater contributing factor that sustains perennial rivers. Baseflow can be estimated through various methods such as hydrograph separation (HYSEP), recursive digital filter (RDF), Herold, recession curve displacement (RCD), Smakhtin low flow, United Kingdom Institute of Hydrology (UKIH), and tracer tests. Baseflow is a very important factor in relation to environmental impact assessment (EIA) studies, hydrological sustainability, and ecological reserve, however its estimation is limited. Tracer tests are considered the most accurate estimation of baseflow, but are not used frequently due to the time, effort, and money involved. Due to the limited research on the applicability of the above methods in a South African context, this research study's objective was to establish an estimation method built from the advantages of existing methods with tracer test results as a benchmark. Learning from these existing methods, a newly proposed baseflow estimation method, namely, the South African Baseflow Estimation Method (SABEM) is introduced in this dissertation. This method makes use of readily available hydrograph data from the Department of Water and Sanitation through a Python scripted data scraper. By leveraging the advantages of established methods and incorporating a foundation established through tracer studies, the parameters of the SABEM were calibrated with parameters identified by existing methods proven to have direct impact on baseflow. The designated study areas are adequately broad and tailored to cover various South African geohydrological regions. They are distinct enough to include diverse parameters, ensuring calibration extends across the spectrum of South Africa's geological and hydrological features. Therefore, specific parameters were identified for upper and lower spectrums emphasising the quality of parameters rather than quantity of case studies. Lastly, the areas also required existing studies of tracer tests on baseflow to be sufficient. Study areas across South Africa combining various and unique characteristics were selected at the following flow-gauging stations: G1H043, T3H009, V1H010, X2H008, X2H010, X2H022, and X2H031. For SABEMto accurately and successfully estimate baseflow, it was crucial for the characteristics of all study areas to exhibit a directional linearity with SABEM parameters. This necessitated running all study areas in parallel through the SABEM calibration process to obtain equal baseflow estimations to that of the case study tracer results for baseflow, creating a directly proportional system to encompass all catchment characteristics across South Africa. Thus, the characteristics of South African catchments can serve as a basis for baseflow estimation, given that the parameters demonstrate directional proportionality with tracer results for baseflow. The tool underwent development and calibration through multiple iterations, ensuring the parameters became directly proportional to the catchment characteristics within these areas and aligned with the tracer study findings. Later, the tool was effectively applied in a test case within the verification catchment area, where tracer study results were absent, and the area wasn't among the initial study areas. The outcomes indicated an accurately estimated baseflow, asserting the applicability of SABEM to diverse catchments with varying characteristics across South Africa (SA), yielding reliable results without the need for tracers. Catchments characteristics included catchment size, groundwater occurrence, and groundwater gradient towards rivers. This means that the SABEM is applicable to other catchment areas besides those used in the study. One of the four catchment parameters, namely, hydraulic conductivity, when calibrated to SABEM parameters was found be negligibly small (0.014 K). For this reason, hydraulic conductivity could not be calibrated completely linear to the parameters and therefore can be used in addition to SABEM according to user preference as it is found to be a very sensitive parameter for catchments. The SABEM estimation method can be utilised in a variety of South African contexts and can be incorporated into EIAs, hydrological and ecological reserve studies with confidence without requiring the use of tracer tests.