Applicability of the daily hydrological models (GR4J, GR5J, and GR6J) in the South Korean basins

Despite the increasing frequency of extreme droughts caused by climate change, which prompts a reevaluation of the need for low-flow simulation, there is a lack of evaluation of low-flow estimation using GR5J and GR6J models in Korean basins compared to studies on GR4J. This study assesses the applicability of the GR5J and GR6J models in South Korean basins. It aims to determine how to calibrate these models effectively in South Korea. The models were calibrated for 10 basins using seven PET equations and 100 samples. Results show that GR5J and GR6J performed better than GR4J for both all samples and the best cases. Romanenko, FAO-24, and FAO-56 delivered prior performances rather than other PET equations. The basin area showed a significant relationship with performance, while drainage density and Q90/Q50 ratio exhibited weak relationships. When calibration periods are similar to validation periods, performance notably improves. Reservoir capacity parameters (X1, X3, and X6) demonstrated greater sensitivity than other parameters, whereas the time parameter (X4) was below 1.0. Future research involving various basins and additional sampling trials is necessary to validate these findings and gain more insights into how basin properties and calibration period characteristics impact model performance.