Answers are grounded in this paper and cite their sources.
This analysis evaluates calibration methods for groundwater contributions to streamflow in a semi-arid catchment, suggesting effective solutions for data-scarce regions.
Hydrological models have become fundamental in groundwater recharge estimation, but they require large amounts of input data that are often lacking in the regions needing these studies the most. Innovative solutions are needed to constrain groundwater flow contribution to streamflow, despite the lack of fundamental hydroclimatic data. In this study, we assess four calibration procedures simulating groundwater contribution to streamflow in the semi‐arid, data‐deficient Goukou catchment of South Africa. The model realism is tested through a complementary evaluation with stable isotopes. Groundwater contributions were simulated after calibration with the non‐dominated sorting genetic algorithms‐2 (NSGA2), DiffeRential Evolution Adaptive Metropolis (DREAM), Monte Carlo analysis (MCA), and Latin Hypercube (LHC) calibration procedure in the fully distributed, conceptual rainfall‐runoff model J2000. DREAM performed best in finding a parameter set with a good balance between a realistic groundwater flow proportion as published for the Table Mountain Group aquifer (~31%), while maintaining fair Nash Sutcliffe efficiencies (~0.38) for streamflow. Isotopic characterisation showed that the catchment is recharged episodically during the most intense rainfall events, with precipitation to groundwater ratios all having negative values. The isotopes corroborated modelled groundwater contribution to streamflow in the upper part of the catchment, but results suggested over‐estimation of groundwater contribution to streamflow by J2000 in the middle section of the catchment. Isotopes proved effective in validating model results and detecting shortcomings in the J2000 model due to sparse hydroclimatic data and the catchment's position in a transition zone between winter and summer rainfall. We also provided evidence that the calibration procedure selection should be carefully considered in data‐scarce circumstances, as good model efficiency does not necessarily guarantee a realistic representation of hydrological processes.
No takes yet. Share an insight, caveat, or question.
Witt et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: