Stream flow prediction of ungauged catchment using Hydrological model the case of Geba Sub-Basin: Tekeze River Basin, Ethiopia

Abstract

Prediction of stream flow of ungauged catchment is important for water resources design, planning and management system. Runoff estimation in interest catchment is probably one of the most basic and oldest tasks of hydrologists. This long-standing issue has received increased attention recently due to the prediction in ungauged basin initiative by the International Association Hydrological Science. In developing countries like Ethiopia most of the rivers are ungauged. Therefore, applying regionalization techniques on an ungauged or poorly gauged river basin is crucial. This thesis deals with stream flow prediction in an ungauged catchments using hydrological model in Geba Sub-Basin. The HBV-96 model is selected to simulate discharge for four gauged catchments on a daily basis in the period of 1998-2008 and with data input such as precipitation, air temperature, potential evapotranspiration, and Geographical zones. Four regionalization methods were applied to transfer model parameter values from the gauged to the interest catchments. Those methods are regional model, sub basin mean, area ratio and proximity methods. In regional model, gauged catchments model parameters and physical catchment characteristics of ungauged catchments were used to develop the equations in order to estimate stream flow from ungauged catchments. To have better understanding of model parameter performance, the sensitivity analysis of eight model parameters were performed manually by trial and error. The evaluation shows that the model parameters, runoff coefficient (Beta), recession coefficient of upper reservoir zone (Khq), limit for evapotranspiration (LP), and Field capacity (FC) are more sensitive than the others. The model performance was evaluated using Nash Sutcliff efficiency and Relative volume error. The result shows that the four gauging river have good agreement and distribution since Nash Sutcliff efficiency greater than 0.67 and relative volume error lies between +10% and -10%. In predicting model parameters from ungauged catchments in regional model method p-value ≤ 0.05 for 95% confidence interval and determination coefficient (R2) ≥0.98 were obtained. Stream flow from ungauged catchments simulated by regional model, spatial proximity, area ratio and sub-basin mean contribute high and less runoff volume respectively.