Modeling Lake Urmia Water-Level Changes using Local Linear Neuro-Fuzzy Method

Abstract

According to the water resources and climate change and challenges of Urmia Lake basin, which is the discharge and final destination of North West Rivers, a model was presented. Due to climate change and water resources in river basin such as rainfall, climate change in basin that has direct impact on evaporation over water catchment areas and lake water, this model can be provided. In addition, the inflow to the lake and the lake water-level fluctuations with the high accuracy and acceptable to experts could be estimated by this modeling and the lake water-level is going to be predicted up to one month. In order to simulate monthly fluctuations of the Lake water-level, this paper dealt with modeling the lake water level using two methods, Water Balance Equation and Local Linear Neuro-Fuzzy Network. In this study, to evaluate models' accuracy, all of them were assessed by three most famous criteria including Root Mean Squares Error (RMSE), correlation coefficient (R), and similarity. The results obtained by Local Linear Neuro-Fuzzy Network modeling indicated that the concomitant use of cumulative flow (entering the lake), monthly precipitation and monthly evaporation on the lake surface provided the best performance with high accuracy regarding the simulated fluctuations of the monthly water level in Urmia Lake.