Hydrological modelling in east Gojjam Region, Ethiopia
MetadataVis full innførsel
The Blue Nile River is known for its huge hydropower potential. Future sustainable planning and management of the river system require advanced approaches in hydrological analysis. Rainfall-runoff models are one of the hydrological tools that can be used for runoff forecasting (real-timeforecasting), flood forecasting, generation of runoff time-series from meteorological data(precipitation and air temperature), filling-in of missing runoff observations, quality control of runoff data, determining the effects of land use changes in a catchment and studying the effects of climate change. In this study, two types of rainfall-runoff models: a lumped conceptual model( HBV) and a distributed physically based model (LANDPINE), both widely used at the Department of Hydraulic and Environmental Engineering at NTNU, were applied to compare their performance in the East Gojjam region, part of the Blue- Nile basin. Different evapotranspiration computation methods were also compared with measured evapotranspiration to determine the best method for assessment of climate impact. The achieved Nash-Sutcliffe efficiency coefficient (R2) values were 0.76 and 0.78 for LANDPINE and HBV models respectively which showed the acceptable performance of both models on the Chemoga catchment. The higher R2, less data requirement, less time (cost) of model set up, its flexibility to be integrated with daily evapotranspiration computation, and the purpose of the applications made the HBV model to be selected for further model applications over the LANDPINE model. Among the different methods of computing Potential evapotranspiration, the Thornthwaite method was found to be more representative and had the best fit to the observed records. However, to achieve this, an adjustment of the heat index on the original method and conversion into daily equivalent (by using temperature anomalies) were made. The potential of the HBV model to improve quality of flow data (i.e. filling missing data and estimating flow for nearby rivers) was found sufficient, even though, the monthly average of simulated flow on Jedeb and particularly Temcha rivers by HBV model showed lower values than the observed flows. This may have been due to the lack of meteorological data collected within these catchments. With respect to estimating the impact of projected climate change, the results from HADCM3A2 and B2 scenarios indicated that in the 2050s there will be small changes (increase or decrease) depending on the scenario. However, for CGCM3 both the A2 and A1B scenarios which were used showed a declining trend. The range of changes across the scenarios was larger for CGCM3 than HADCM3. The flows with 50% and the 90% exceedance across all the scenarios were in the range 0.87 to 1.23 and 0.7 to 0.96 m3/s respectively. The 50% value for the observed is 1.12m3/s. Infiltration excess runoff, variable soil moisture parameter and daily evapotranspiration computation method should be included with both models for these catchments so as to improve simulation of quick runoff.