Predicting Discharge at Ungauged Catchments Using Rainfall-Runoff Model (Case study: Omo-Gibe River Basin)
No Thumbnail Available
Date
2011-07
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Addis Ababa University
Abstract
This paper deals with predicting discharge at ungauged catchments, the case of Omo-Gibe river
basin, using the WATBAL conceptual lumped rainfall-runoff model. Parameters calibrated with
the model are extrapolated from gauged catchments to ungauged catchments of similar physical
characteristics with Regionalization technique. In the Omo-Gibe river basin, most of the
catchments are ungauged. Hydro-meteorological data from twenty four metrological and twenty
two hydrological gauging stations of the basin are used to calibrate and validate the model
parameters. Sizes of gauged catchments whose data are used vary from about 40 to 3242 km
.
Key model parameters considered consist of sub-surface runoff coefficient (α), surface runoff
coefficient (ε), and maximum water holding capacity of catchments (Smax). These parameters are
calibrated using an automatic optimizing routine of the WATBAL model and through routine
iteration; Smax is found to vary from 3403mm to 466mm, α from 64mm/day to 0.26mm/day and ε
from 9.6 to 2.6. Some more model parameters, namely, direct runoff coefficient (β), subsurface
runoff coefficient (γ), and base flow (Rb) are calibrated manually. The hydrograph characteristics
of observed and simulated events are compared using various evaluation criteria consisting of
Nash-Sutcliffe coefficient (NS) greater than 0.6, relative volume error (RVE) between -10 and
10%, and coefficient of determination (r
2
) greater than 0.7. Among the catchments of the basin
that are used in the modeling work, 73% fulfilled the criteria. Based on availability of required
data, six physical catchment characteristics (PCCs) are selected among those commonly used in
many regionalization studies. These are catchment area, mean annual precipitation, mean annual
evapotranspiration, average catchment slope, longest flow path and mean catchment elevation. In
addition to the PCCs, input data including precipitation, observed flow, temperature, relative
humidity and sunshine are used in the WATBAL rainfall-runoff model. Four parameter
2
regionalization methods (multiple regression, spatial proximity, area ratio and sub-basin mean)
were applied to transfer model parameter values from the gauged to the ungauged catchments.
Description
Keywords
Omo-Gibe River Basin, Ungauged Catchments, Rainfall-Runoff Model, Discharge