Impacts of Surface Water Quality on Groundwater Quality (Case study: Tabriz plain)

In order to determine effective parameters on Aji Chay river pollution and the aquifer of Tabriz plain located in eastern part of Urmia Lake and better management and controlling of water resource quality, principal component analysis, artificial intelligence techniques (neural networks and decision tree M5) and geostatistics (Kriging and Cokriging) through water quality data analysis (Cations, Anions) in the period of 2001 to 2013 have been used. Statistical multivariate analysis has an ability to process a big range of data, and is known as a quantitative method in order to manage water quality of the East Azarbaijan province’s rivers. Determination of the water quality parameters would result in costs reduction and aiming the monitoring, controlling, and protection of water quality. The results show that M5 method presents simple linear relations in predicting chloride and salinity, also the decisions for categorizing decision tree algorithms are clear and extractable. Also the results of geostatistics showed that Tabriz plain's aquifer is in a bad condition from qualitative viewpoint compared to the last years, and in western part of the plain due to extreme well extraction, salinity in outlet from 9000 in 2001 to over 11000 (µmhos cm-1) in 2012 has been increased.