Determination and evaluation of surface and groundwater quality parameters of Amasya province with artificial neural networks approach

Correct determination of the quality of water resources used as drinking and potable water is extremely important for human health. In this study, it was aimed to reveal the quality of the surface and groundwater of Amasya Province and to model the relationship between water quality parameters and nitrate concentration by using data form previous years (2003-2018) with the help of Artificial Neural Networks (ANN) and Multiple Linear Regression (MLR) methods. Adding to this purpose, in the study area, by using the analysis results of the surface and groundwater, hydrochemical facies of groundwater and quality classes of surface waters were determined. The quality indices of surface and groundwater based on the values recommended by TS 266 were calculated by using the Canadian Water Quality Index (CWQI) method. Following the evaluations carried out, it can be seen that the groundwater is mainly in the Ca-HCO3 water type; According to the parameters SAR, MH, RSC and %Na, it was determined that some of the groundwater in Merzifon and Suluova are not suitable for irrigation water. Based on the parameters given by RWQC, Yeşilırmak Stream and its tributaries are in different quality classes. Seasonal differences in water quality are observed, but no upward or downward trend has been observed for any parameter from the past to the present. Based on CWQI values of surface and groundwater, it was determined that in general groundwater is in good quality and surface water is in lower quality. To determine the superiority between the ANN and MLR methods used to model the relationship between the water quality parameters and nitrate concentration determination coefficient (R2), root mean squared error (RMSE) and mean absolute error (MAE) criteria were used. It emerges from this comparison that the ANN method gives better results than the MLR method.