Journal of Hydrology: Regional Studies (Feb 2019)
Assessing the impact of a multimetric calibration procedure on modelling performance in a headwater catchment in Mau Forest, Kenya
Abstract
Study region: Nyangores catchment, Mara catchment, Kenya. Study focus: Hydrologic models are widely used tools in watershed management to assist in decision making by representing catchment functions under alternative scenarios. This study focused on the evaluation of the 5 segment Flow Duration Curve based calibration procedure (5FDC) for the period 1975–1978 using the Soil Water Assessment Tool (SWAT). The Normal Calibration Procedure (NCP) based on the hydrograph was compared to the 5FDC. Two separate ensembles each containing 10 Pareto calibration runs were obtained individually from each method and evaluated based on their goodness-of-fit. A final Combined Procedure (CP), which was an intersection of 5FDC and NCP was executed. With this supplementary analysis, the parameter distribution and water component balance were analyzed and compared. New hydrological insights for the region: The comparison of 5FDC and NCP shows that the 5FDC provides a better representation of the low and mid-level section of the flow duration curve as compared to the NCP, which greatly overestimates these flows. This method also closely matches the observed runoff ratios. This indicates that the 5FDC calibration may be well suited for water resource applications focused on low flows. The CP combines the advantages of both procedures by improving parameter identifiability, leading to better representation of high and low flows. Keywords: 5 segmented Flow Duration Curve (5FDC), Normal Calibration Procedure(NCP), Parameter identifiability, Pareto set, Pareto analysis, Water balance, Multimetric calibration, Latin hypercube sampling, Performance metrics, Uncertainty, Prediction, Hydrologic model, Simulation, Equifinality, Base flows, Variability, Optimum, Stepwise intersection, Ensemble, Water partitioning
