Simulation of Urmia Lake Level Changes and Its Uncertainty and Sensitivity to Water Budget Components

This research work is an attempt to simulate and analyze monthly water level changes in Urmia Lake. To this end, water budget, multiple regression and artificial neural networks (ANNs) approaches have been investigated, using monthly data of effective components of water budget equation such as input discharge, average rainfall and average evaporation. Furthermore, uncertainty and sensitivity analysis were employed to compare the simulation methods capabilities. The results suggested that ANNs model using monthly discharge, rainfall and evaporation as inputs gave best results with less sensitivity, but greater uncertainty.