Groundwater Quality Analysis by Integrating Water Quality Index, GIS Techniques and Supervised Machine Learning: A Case Study in Duhok Province, Iraq

Abstract

This paper presents a case study focusing on the analysis of the Water Quality Index (WQI) using ArcGIS Pro and supervised machine learning (SML) techniques. The study aims to analyze the selection of physicochemical water quality indicators in water wells to determine the most effective physicochemical water quality parameters in water wells, in addition to finding the WQI of each well in Duhok province and its purpose of use. These parameters include Calcium, Magnesium, Chloride, Sodium, Potassium, Sulfate, pH, Electrical Conductivity (EC), Total Dissolved Solids (TDS), Nitrate, Total Alkalinity (TA), and Total Hardness (TH). The study generated a spatial distribution map of the WQI, revealing the highest values in wells located in the Sumil district, ranging between 18.47 and 57.9, while the lowest value of 18.47 was observed in the Akre district. Supervised machine learning algorithms were employed to identify the most influential physicochemical indicators of water quality. The results highlighted EC, TA, TH, and Ca+2 as the most crucial parameters affecting WQI. The mapping analysis further indicated that wells in the Sumil district exhibited the highest values of EC, TH, Mg+2, and TA. Conversely, the Duhok district demonstrated the highest calcium levels, while the lowest pH and nitrate levels were observed in the Duhok and Amedi districts, respectively. The Zakho district showcased the highest levels of sulfate and potassium, and the Bardarash district had the highest chloride and sodium values.