Parameter estimation using regionalization for ungaged basins: An example from the Upper Blue Nile River Basin

Document Type


Journal/Book Title/Conference

Presented at the Spring Runoff Conference

Publication Date



Hydrologic models, regardless of their complexity, need to be calibrated to reproduce the observed basin responses using historical data such that a model can be used reliably for prediction purposes. Many optimization algorithms have been developed and tested to determine the best model parameters to reproduce the observed basin responses especially for gauged basins. Frequently hydrographs need to be simulated for ungaged basins using available regional information, i.e., basin characteristics. Unfortunately, no best approach of regionalization is available due to the uncertainties of models and observations. This study investigates several regionalization methods to estimate plausible model parameters of the upper Blue Nile River Basin of Ethiopia where its major tributaries are ungaged and its flow variability affects water availability to the riparian countries. Six model parameters of a two-layer water balance model based on monthly time scale are being regionalized using (a) averaging, (b) aggregated calibration, (c) volume fraction, (d) multiple regression, and (e) regional calibration. The performance of each method in representing the observed hydrographs of small and large basins is evaluated using model efficiency and volume error. Since most gauged sites of the upper Blue Nile River Basin have much smaller drainage areas than the major tributaries, the applicability of the regionalization methods to accommodate small and large basins is also evaluated. The results and benefits of this study will play an important role in estimating the water yield and its variability of rural ungaged drainages of the upper Blue Nile River Basin and other similar basins.

This document is currently not available here.