Groundwater level forecasting using multivariate-multiple regression model

The significance of groundwater is no secret to anyone as an important source of water supply in arid and semi-arid areas. Precise understanding of the groundwater fluctuations is necessary for prediction of groundwater future behavior and eventually its management. Naishaboor plain is selected for this research because of presence of 45 observation wells that mostly have more than 18 years data. At the first of this paper, a report of preprocessing job done on the raw data using GIS for editing and generating of requirement of data in year scale is given. Then, a new Multivariate-Multiple regression model using Least Square method for estimation of groundwater fluctuations in many points simultaneously is presented. Aquifer extractions, precipitation, maximum and minimum temperatures and groundwater level the year prior to the predicted year were used as independent variables. The results showed that this method can be used as a suitable tool for prediction of groundwater table fluctuations in an extend area (many points simultaneously). The selected performance criteria indicators such as RMSE=3.23 and MAXE=1.33 reveals the relevance of this method.

Groundwater level forecasting using multivariate-multiple regression model Keywords:

Groundwater , Prediction , Multivariate-Multiple Regression , Least Square , Naishaboor Plain.