Sensitivity analysis of a physically based distributed hydrological model

Sensitivity analysis using a distributed physically based hydrological model was carried out and the results obtained are presented. The model used in this research is the one known as NASIM which is a German abbreviation meaning simply Rainfall-Runoff Simulation Model. It was originally developed by Ostrowski (1982). Later on it was modified to represent various sub-processes as accurately as possible. In this dissertation, this model is applied to two catchments; the Umrobo catchment for the daily simulation and the Agucho catchment for event simulation. The model was calibrated and verified using the technique of split sampling. The results of the calibration phase are quite good. In the verification period however, the model results are fairly good. The inconsistency of results could be due to inconsistency of the model and reliability of the data amongst others. The sensitivity analysis performed on the selected model parameters showed that the most important parameters that control the rainfall-runoff process in event simulation are the overland flow retention constant, initial soil moisture, surface roughness, and time step of simulation. In the continuous simulation mode the important parameters considered are the overland flow retention constant, the base flow retention constant, saturated hydraulic conductivity, maximum infiltration rate, initial soil moisture, and initial groundwater storage.