Multi-objective regionalisation for lake level simulation, the case of Lake Tana in the Upper Blue Nile, Ethiopia

Abstract

The aim in this study is to simulate lake levels of Lake Tana by solving the water balance at daily time step. Since 42% of the basin is ungauged regionalisation procedures are applied. We examine the predictive capability of a regionalisation approach that combines multi-objective calibration of a simple conceptual model and multi regression analyses to establish relations between model parameters and catchment characteristics. Recently few studies are presented on lake level simulation of Lake Tana. In these studies the water balance of the lake is closed by estimation of runoff contributions from ungauged catchments. Studies partly relied on simple ad-hoc procedures of area comparison to estimate runoff from ungauged catchments. In this study a regional model is developed that relies on principles of similarity of catchments. For runoff modelling the HVB-96 model is selected while multi-objective model calibration is by a Monte Carlo procedure. <br><br> Assessment of the lake water balance was established by comparing measured to estimated lake levels. Results of daily lake level simulation show a water balance closure term of 85 mm and a relative volume error of 2.17%. Results show runoff from ungauged catchments of 527 mm per year for the simulation period 1994 to 2003 that is approximately 30% of Lake Tana stream flow inflow. Compared to previous works this closure term is smallest