| Summary: | The prediction of dominant hydrological processes is imperative with the available information in data scarce regions by means of the lumped hydrological models for the purpose of water resource management. This study is aims at an intercomparison of the performances of the conceptual hydrological models in predicting streamflow. The Veralgemeend Conceptueel Hydrologisch (VHM) and NedborAfstromnings Model (NAM) lumped rainfall–runoff models were manually calibrated and validated for periods of 1 January 1990–31 December 2000 and 1 January 2001–31 December 2005, respectively. Some of the parameters of the models (i.e., recession constants of subflow components) were estimated from the preprocessing of the streamflow data using the Water Engineering Time Series PROcessing tool (WETSPRO). These parameters were used for the initial model setup and subjected to slight adjustments during calibration. The performances of the models were evaluated by graphical and statistical means. The results depicted that the models reproduced the streamflow in a good way and that the overall shape of the hydrograph was properly captured. A Nash Sutcliffe efficiency (NSE) of 0.71 and 0.67 were obtained during calibration, whereas, for the validation period, NSE of 0.6 and 0.58 were obtained for VHM and NAM, respectively. The water balance discrepancy (WBD) of −0.1% and −13.7% were achieved for calibration, while −17% and −9% were acquired during validation for VHM and NAM, respectively. Though the models underestimated the high flows, the low flows were relatively well simulated. From the overall evaluation of the models, it is noted that the NAM model performed better than the VHM model in predicting the flow. In conclusion, the models can be used for water resource management and planning with precautions for extreme flow.
|
|---|
