Survey of Groundwater Level Fluctuations by Soft Computing Techniques (Case Study: Urmia and Miandoab Plains)

The study of behavior of water level fluctuations is important for water balance due to the unbalanced distribution in the hydrological cycle. In present study, the groundwater level fluctuations of Urmia and Miandoab plains in two models SVM and Wavelet from the group of soft calculations have been investigated. For this purpose, information about the period years of 1381-1386 for both plains of Urmia and Miandoab by considering four scenarios of groundwater level -groundwater level with a delay of one month, groundwater level -groundwater level with a delay of two months, Groundwater level -precipitation and groundwater level -temperature. The simulation results were compared and the best model was introduced. The results of regression analysis showed that R2=0.9 and MSE=0.4 belong to the second piezometer well of Urmia plain and the third piezometer well of Miandoab plain in both SVM and Wavelet models for 3 scenarios of groundwater level -groundwater level with a delay of one month, Groundwater level was precipitation and groundwater level was temperature. Data on climatic variables of temperature, precipitation and groundwater level with a delay of one month as an independent variable and groundwater level data as a dependent variable formed the mentioned scenarios with 80% of educational data and 20% of test data. The results showed that the climatic variable of precipitation has a significant effect on groundwater level fluctuations. In general, the prediction results of Wavelet model were better than SVM model in investigating groundwater level fluctuations in Urmia and Miandoab plains. In the present study, an independent study of the mentioned models in order to simulate groundwater level fluctuations and compare each of the two models in their application.