Integrated Artificial Neural Network Modeling and GIS for Identification of Important Factor on Groundwater Hydrochemistry (Fe-,Ca2+ and PO4-3)

Background & Aims of the Study: Groundwater resources are a crucial component of the ecosystem. Management and cleanup of contamination from groundwater resources requires a long term strategy and a huge amount of investments. Artificial neural networks (ANN) and Geographic Information System (GIS) can be useful in determining management strategies. To protect these valuable resources, groundwater hydrochemistry (Fe-, Ca2 and PO4-3) spatial distribution is evaluated; also, the important parameters that affect their rate and spatial distribution are identified.
This study employed GIS technique and Modeling technique based on artificial neural network for identification and investigation of important factor on groundwater hydrochemistry such as Fe-, Ca2 and PO4-3. The case study is Ghareh-su basin of Golestan province of Iran. The maps of land use, soil, geology, population density, digital elevation model, distance from built-up areas, roads and rivers, cultivated land density and water table are the parameters that used for running ANN model. Sensitivity analyses were also performed to identify the effective parameters of ground water hydrochemistry
The results show that the concentration of the parameters around Gorgan and Kordkuy cities, and areas where the cultivated land is denser, is high.
Results indicated that the highest concentrations of these parameters were located around Gorgan and Kordkuy cities and where the cultivated lands have a high density. The present contribution confirms that a significant relation between the concentration of pollutants in groundwater resources and different land uses/land covers is found. Soil type, geological structure and high groundwater level in the north of Ghareh-su basin have a great impact on groundwater quality.
These techniques have successfully implemented in groundwater hydrochemistry mapping of Ghareh-su basin.