Application of “panel-data” modeling to predict groundwater levels in the Neishaboor Plain, Iran

The aim of this research was to predict groundwater levels in the Neishaboor plain, Iran, using a "panel-data" model. Panel-data analysis endows regression analysis with both spatial and temporal dimensions. The spatial dimension pertains to a set of cross-sectional units of observation. The temporal dimension pertains to periodic observations of a set of variables characterizing these cross-sectional units over a particular time span. Firstly, the available observation wells in the Neishaboor plain were clustered according to their fluctuation behavior using the "Ward" method, which resulted in six areal zones. Then, for each cluster, an observation well was selected as its representative, and for each zone, values of monthly precipitation and temperature, as independent variables, were estimated by the inverse-distance method. Finally, the performance of different panel-data regression models such as fixed-effects and random-effects models were investigated. The results showed that the two-way fixed-effects model was superior. The performance indicators for this model (R ² = 0. 97, RMSE = 0. 05 m and ME = 0. 81 m) reveal the effectiveness of the method. In addition, the results were compared with the results of an artificial-neural-network (ANN) model, which demonstrated the superiority of the panel-data model over the ANN model.