Abstract:
Hydro-climatic projections in West Africa are attributed with high uncertainties that
are difficult to quantify. This study assesses the influence of the parameter sensitivities and
uncertainties of three rainfall runoff models on simulated discharge in current and future times
using meteorological data from eight Global Climate Models (GCM). The IHACRES Catchment
Moisture Deficit (IHACRES-CMD) model, the GR4J, and the Sacramento model were chosen for this
study. During the model evaluation, 10,000 parameter sets were generated for each model and used
in a sensitivity and uncertainty analysis using the Generalized Likelihood Uncertainty Estimation
(GLUE) method. Out of the three models, IHACRES-CMD recorded the highest Nash-Sutcliffe
Efficiency (NSE) of 0.92 and 0.86 for the calibration (1997–2003) and the validation (2004–2010)
period, respectively. The Sacramento model was able to adequately predict low flow patterns on the
catchment, while the GR4J and IHACRES-CMD over and under estimated low flow, respectively.
The use of multiple hydrological models to reduce uncertainties caused by model approaches is
recommended, along with other methods for sustainable river basin management.
