عنوان مقاله [English]
Increasing volatility and decreasing water depths in Lake Urmia along with the rapid progress of desertification during the last decade has become one of the most important and controversial environmental challenges in the country. In this study, from the viewpoint of the components that affect the elevation level of Lake Urmia water, the water level is analyzed and using the neural networks, its amount is foreseen for the coming years, so that an appropriate management plan for the program Design and implement necessary measures to prevent drought in Lake Urmia. Thus, in order to achieve the highest accuracy in predicting and simulating the water level, five scenarios were investigated, in which the input data were considered as evaporation, temperature, precipitation, temperature-precipitation and evapotranspiration respectively. The water level was selected as the output for network training. Finally, reaching the value of the value 0.026 for RMSE in the first scenario and 0.023 for RMSE in the scenario Second, as well as 0.093 for RMSE in the third scenario, we find that the effect of precipitation from temperature and evaporation and the effect of temperature on evaporation is greater, but given the results of 0.068 for RMSE in the fifth scenario. The most accurate scenario is the fifth scenario, which is the result of concurrent import of all three variables of temperature, evaporation and precipitation into the neural network.