Simulating Fluctuations of Groundwater Level Using a Combination of Support Vector Machine and Wavelet Transform

Groundwater resources are known as one of the most important water resources. Proper knowledge and exploitation of groundwater, particularly in arid and semi-arid regions, plays an important role in sustainable development in different fields such as agricultural, social, and economic activities. Many models have been used in order to predict groundwater level including empirical time series models and physical models (Bierkens, 1998). It is true that these models have been widely used, but when the dynamic behavior of a hydrological system is changed during the time, the mentioned models may not be able to predict water resource parameters, therefore making them unsuitable (Nayak et al. 2006). On the other hand, physical models need a great deal data to simulate groundwater fluctuations and, since the relations among effective parameters may be nonlinear, the mentioned models are not practical enough to show the existing relations.