Hydrogeochemical and hydrogeological study of Harzandat plain aquifer using multivariate statistical methods and graphical methods

This study investigated the hydrogeology and hydrogeochemical processes of groundwater in the Harzandat plain aquifer. The Harzandat basin shown in Figure 1 is located northwest of Tabriz, the city in the northwest region of Iran. It is a subbasin of the Khazar basin. The Harzandat basin is about 347 km2. This includes 146.7 km2 of the Harzandat plain and 200.3 km2 of Mountainous areas. The highest elevation of the Harzandat basin is 2623 m above mean sea level (amsl) at the old Harzandat and the lowest elevation is 1,274 m amsl at Zal railway station. Based on de Martonne (1925) and Emberger (1930), the prevailing climate in the Harzandat plain is arid-cold. Average annual precipitation is about 219 mm (Chercher climatological station, 2012-2013).
In the last decades, this aquifer suffered from severe groundwater level declination and caused degradation of groundwater quality. To better understand of hydrogeochemical processes in Harzandat plain, groundwater samples obtained 77 from 16 different water wells, springs and qanats between 2002 and 2013 and used them to characterize the hydrogeochemistry of the Harzandat plain aquifer. The location of the sample sites is shown in Figure 3. The samples were analyzed in the hydrogeological laboratory of University of Tabriz and in the laboratory of East Azerbaijan Regional Water Authority. The water quality parameters of interest were , , , , , , , and , which were determined by the standard methods (American Public Health Association 1998). Moreover, pH, electric conductivity (EC), the total dissolved solid (TDS) were measured from the samples. Total hardness (TH), sodium absorption ration (SAR), and sodium percent (Na%) were calculated from the measured chemical parameters. Samples collected from groundwater resources were analyzed through the graphical and multivariate statistical methods.
Graphical methods (e.g., a Piper diagram, a Durov diagram, a Stiff diagram, etc.) are commonly used to interpret hydrogeochemical processes (Stiff 1951; Piper 1944; Hem 1986; Durov 1948; Lloyd 1965). Piper diagrams were used to deduce groundwater types and an expanded Durov diagram improves the Piper diagram by providing a better display of different types of water as well as important hydrochemical processes, such as ion exchange, simple dissolution and mixing of waters with different qualities. A Stiff diagram compares analytical data in pairs and infers types of source rocks (Stiff 1951; Hounslow 1995). Although the graphical methods are commonly used to interpret hydrogeochemical processes, they have several limitations as they cannot analyze neutral chemical species (e.g., SiO2 and ) and nonchemical data (e.g., temperature) (Voudouris et al. 1997).
Multivariate statistical methods (e.g., factor analysis (FA) and hierarchical clustering analysis (HCA)) are able to complement the limitations of the graphical methods for hydrogeochemical process interpretation (Cloutier et al. 2008). Multivariate statistical analysis is able to explain the correlation among a large number of variables and reduce the number of variables into a small number of factors without loss of essential information. In multivariate statistical methods, factor analysis was frequently used to investigate hydrogeochemical origins (Voudouris et al. 2000). Factor analysis was also applied to investigating groundwater contamination. Clustering analysis was used as a classification technique for hydrogeochemical type investigation and the interpretation of their origin (Cloutier et al. 2008).
The results of the Piper diagram shows two groundwater types and the Stiff diagram showed six different sources of groundwater samples. The Durov diagram identified two major hydrogeochemical processes in the aquifer. However, Hierarchical Cluster Analysis (HCA) identified five water types in the groundwater samples, because HCA was able to use more chemical and physical data than graphical methods. The HCA result was checked by discriminant analysis method and the classification accuracy of groundwater resources samples and confirmed by this method. Based on the factor analysis method, four main effective factors were found on the hydrochemistry of the Harzandat plain aquifer. First, second and third factors show the effect of geological formation and general groundwater trend on hydrochemistry of the area. The fourth factor seems to be related the influence of human activities. In the first factor, chloride, sulfate, magnesium, calcium and electrical conductivity, and in the second factor percentage of sodium and sodium absorption ratio, and in the third factor potassium and bicarbonate is effective, while, in the fourth factor carbonate were related to nitrate and acidity. In fact, the first, second and third factors are geogenic, and the fourth factor is anthropogenic.