Assessment of groundwater quality using water quality index (WQI) and multivariate statistical analysis in Amman-Zarqa area/Jordan

Groundwater is considered the basic element in the arid and semi-arid regions where water resources are limited. Natural and anthropogenic activities have a considerable impact on groundwater quality and quantity.

Due to population rate increase, excessive pumping, long periods of drought, and improper management, groundwater qualities became highly affected in many areas of the world (Causapé et al. 2004). The groundwater deterioration issue reflects the significance of the control and sustainability of the quality of water. The changes in the concentrations of different hydrochemical constituents of the groundwater due to natural or anthropological activities will change the suitability of the aquifer system. Accordingly, the periodic assessment of water quality becomes a necessity. This can be done by determining the concentrations of ion species in the water and comparing the findings with the already set standards (Hasan et al. 2020). Moreover, it is very important to understand the processes that control groundwater quality.

Different known methods are used to assess water quality by determining the water quality index and statistical treatment of the measured variables. In addition, hydrochemical studies involve an evaluation of the chemical composition of groundwater, therefore offering a better understanding of possible changes in water quality. The application of these techniques, over long periods, allows the accumulation of huge data sets, promotes sustainable development and effective management of groundwater and it has become very popular in determining the suitability of groundwater, mainly because of the rising need for clean water resources (Shrivastava et al. 2018).

There are many types of water quality indices (WQIs) that are frequently used in assessing water quality. For low parameter values the most efficiently used one is that of the Canadian Council of Ministers of the Environment (CCME) and that of the British Columbia Ministry of Environment Water Quality Index (BCWQI). The National Science Foundation (NSF) uses a WQI that converts concentration data into one of five water quality classes ranging from ‘ very bad’ to ‘excellent’. However, the most efficient WQI to be used is one that uses parameters that are carefully selected depending on the source and time. The application of these indices provides a general glimpse of water quality status and assesses the suitability of water for some specific utilization such as irrigation and suitability for drinking water supply. A WQI is an indispensable tool that can summarize huge information on many parameters, and allow the interpretation of water quality data into a single value. The obtained values can then be then classified into different groups describing water quality. Semiromi et al. 2011 state that a WQI is used in most countries to study water quality only in terms of their physicochemical parameters and that there is a need to involve the biological parameters in calculating WQI. This can be done using fuzzy logic. Doing so would produce a new index called the fuzzy water quality index (FWQI). The obtained indices can be then grouped into different classes. The validity of these classes can be checked by using both one-way ANOVA and Tukey-HSD test. This gives a reliable picture of water quality in a simplified form, which is easy to understand by citizens and policymakers.

The developed indices have a common structure predominantly based on physical, chemical, and biological parameters such as weighted arithmetic water quality index (WAWQI), National Sanitation Foundation Water Quality Index (NSFWQI), Canadian Council of Ministers of the Environment Water Quality Index (CCMEWQI), Oregon Water Quality Index (OWQI), etc. (Abbasi & Abbasi 2012). These indices have been modified by applying different mathematical methods or by choosing suitable water quality parameters for a particular purpose for a particular region.

In this study, the WQI was applied. The parameters used in indexing are often weighted according to their importance to water quality. However, a small change in weighting will affect the overall interpretation of water quality (Mukate et al. 2019). Jehan et al. (2020) conducted a study about the evaluation of the Swat River, northern Pakistan, by using a water quality index and multivariate statistical techniques. The authors conclude that the mean concentrations of physicochemical parameters and the heavy metals concentrations are within permissible limits of the WHO (2017) except 34, 60, and 56% of copper (Cu), nickel (Ni), and lead (Pb), respectively. Solangi et al. (2019) evaluated the groundwater quality in Sujawal district, Pakistan by using the WQI, the synthetic pollution index (SPI), and geospatial tools. They revealed that groundwater in most of the investigated areas does not meet WHO guidelines and the used water is highly contaminated and of high risk for human health. Varol (2020) finished a study on the use of water quality index and multivariate statistical methods for the evaluation of the water quality of a stream affected by multiple stressors in Sürgü in Turkey. The study concludes that the water quality of the stream is affected by multiple stressors such as untreated domestic sewage, effluents from fish farms, agricultural runoff, and stream bank erosion.

El-Naqa & Al Raei (2021) conducted a study for the assessment of drinking water by using the WQI in the Greater Amman area, Jordan. This study assessed the drinking water from 26 samples and the calculated WQI ranged from 29.17 to 62.32. The WQI analysis reveals that the water quality varies from excellent to good water quality. In addition, the spatial distribution mapping of the WQI compared with the WQI for the five years from 2012 to 2016 indicates that the water quality of potable drinking water has deteriorated in 2016 due to the high population growth of Greater Amman in comparison to the precedent years.

Obeidat & Awawdeh (2021) finished a study about the Assessment of groundwater quality in the area surrounding Al-Zaatari Camp, Jordan, using cluster analysis (CA) and WQI depending on samples collected from the Basalt and B2/A7 aquifers (26 and 4 samples respectively). This study concluded that the groundwater showed two main hydrochemical facies: mixed Ca-Mg-Cl and Na-Cl and the groundwater chemistry in the study area is influenced by the processes of ion exchange of both types and rock weathering as was deduced from the Gibbs diagram. Finally, the WQI calculations revealed three categories of groundwater: (1) excellent, which involved 46% of the sampled wells, (2) good, which involved 50% of the sampled wells, and (3) poor, which involved only two sampled wells.

Similarly, this study aims at evaluating and classifying the groundwater quality in the Amman Zarqa area by constructing spatial distribution maps for the calculated WQIs, based on the different measured parameters.

The study area is a part of the AZB (Figure 1) and occupies an area of about 866 km². This area is occupied by more than 6.5 million inhabitants distributed in the three main cities, Amman, Zarqa, and Ruseifa (DOS 2020). Since the early 1980s, tremendous industrial development has taken place in this area. Most industrial enterprises were constructed without provisions for adequate treatment of wastewater or safe disposal of waste. These industries represent a major threat and potentially hazardous pollution points to the groundwater. These pollution point sources originate from the pharmaceutical, dairy, textile, detergents, soft drink, stone cutting, oil refinery industries; thermal power plant; Ruseifa solid waste disposal site; tannery; iron steel company; industrial paper company; scrap yards for cars; and numerous military facilities (Al Kuisi & Abdel-Fattah 2010; Al Kuisi et al. 2014).

The exposed rock units in the study area are of sedimentary and volcanic origin. The stratified sedimentary sequence consists of carbonates of various facies associated with chert and phosphate. Volcanic rocks in the study area are part of the Arabian Volcanic Province (Harrat Ash-Shaam), which extends from Syria via Jordan to Saudi Arabia (Abu Qudaira 2001; Diabat & Abedelghafoor 2004).

The annual precipitation decreases towards the east in the area, where the annual rainfall is around 400 mm in the west and 150 mm in the eastern part of the study area (Al Kuisi et al. 2014). The winter season prevails mostly from October to May. The summer season, on the other hand, is extensively dry. Differences in soil types of the area are related to variation in the topography and rainfall.

Al Saodi (2022) assessed the hydrologic vulnerability of the AZB to climate change impacts using the MOLUSCE tool. She concludes that the climatic changes reflected negatively on the groundwater quality and consequently on the land-use classes in the study area between 2001 and 2021. The study area has witnessed an increase in urban extent by 14.76% during the last twenty years. This increase in urban fabric was accompanied by a decrease in the bare soil and cultivated soil extents by 18.64% and 6.02%, respectively, which indicates that the urban growth was at the expense of the bare soil areas mostly.

Groundwater resources in the study area are used for different purposes other than the drinking ones. More than 800 abstraction wells in the basin are utilized for domestic, agricultural, and industrial purposes in both Zarqa and Amman cities as well as in parts of Mafraq, Jarash, and Balqa governorates (Al-Salihi 2006).

Due to immigration from unstable countries around Jordan, the water demand has increased in this area. This will affect the groundwater quality by local, natural, and anthropogenic activities leading to its contamination and salinization (see Shaqour et al. 2016). In addition, overexploitation of groundwater has become one of the serious problems in the study area. The annual extraction of the groundwater in Amman Zarqa Basin is exceeding 150 MCM when the safe yield is about 70 MCM (MWI 2017). Such circumstances prompted this study.

Major cations (Ca²⁺, Mg²⁺, Na^+, and K⁺) and anions (Cl⁻, HCO₃⁻, SO₄²⁻, and NO₃⁻) were analyzed in the Laboratory of the Department of Geology at the University of Jordan according to standard methods for the examination of water and wastewater, (APHA 2017). Twelve physicochemical parameters, namely: electrical conductivity (EC), potential for hydrogen (pH), total dissolved solids (TDS), hardness (TH), calcium (Ca²⁺), magnesium (Mg²⁺), sodium (Na⁺), potassium (K⁺), bicarbonate (HCO₃⁻), chloride (Cl⁻), sulfate (SO₄²⁻), and nitrate (NO₃⁻), were used to calculate WQI and describe the qualitative conditions of groundwater in the study area. Laboratory quality assurance and quality control methods were used to ensure the quality of the analytical data. Quality control samples included replicates and field blanks. Replicate samples were collected after the routine sampling in the field, and all differences measured in concentrations between replicate pairs were within the precision of the method. Analysis of blank samples did not show any inherent bias in the method of analysis. The accuracy of the analysis was calculated by the charge balance error equation, which resulted in ±5% concentration of the major cations and anions.

The results of the groundwater analysis were obtained using Statistica 13 and Minitab 20 interactive statistics software. The analyses included descriptive, linear modeling, multivariate, and mean categorical comparisons tools.

Univariate analysis for numerical variables included measures of central tendency (e.g. mean, mode, and median), measures of dispersion (e.g. range, standard deviation, variance, and coefficient of variation), and shape (skewness and kurtosis). Data were tested for normality using the Shapiro–Wilk distribution test (Royston 1982).

The bivariate platform has been used to explore the relationship between two numerical variables, while multivariate analyses were used to derive a Pearson correlation coefficient matrix. Pearson correlation matrix was used to determine the relationships among the 12 water quality variables to indicate the strength of the relationships between anthropogenic activities and water properties.

Factor analysis (FA) was used to interpret the hydrochemical facies and origin of the groundwater (Varol 2020; Venkatesan et al. 2020). The main purpose of factor analysis is to reduce the variance of a multivariable data set and preserve the existing data while generating new variables based on linear combinations of the original variables (Mohamed et al. 2015). The calculated factors are rotated by the Varimax rotation method, which results in stronger loadings of closely related variables in each factor. The resultant component will describe a certain amount of the statistical variance of the analyzed data and is interpreted according to the inter-correlated variables. Variable loadings are defined by the orthogonal projection of the variables on each of the factors. The selection of the factors is based on both the significance (eigenvalues >1) of the factor and the cumulative percentage of data variance explained. The final step is to interpret each loading factor in association with its sources and reasons. The factor scores showed the relation between samples and the components quantitatively, which indicated the extent of the impact of each factor on samples (Jiang et al. 2015).

CA was used to identify groups of similar sampling sites based on correlation coefficients as measures of similarity to explore spatial heterogeneity of water quality. There are various cluster techniques such as hierarchical clustering, k-means cluster analysis (KCA), and two-mode clustering that could be used. Among these techniques, hierarchical clustering was applied in this study (Zolekar et al. 2020). Hierarchical clustering is an algorithm that groups similar objects into groups called clusters, resulting in different clusters (Fouedjio 2016). The resultant dendrogram provides a visual summary of the clustering processes that presents a diagram of the groups and their proximity, with a dramatic reduction in the dimensionality of the original data (Fouedjio 2016).

Geostatistical analyses were performed to determine the spatial extent of the groundwater quality in the study area using Kriging interpolation techniques by using ArcGIS (10.8).

Geostatistical analysts use sample points taken from different locations to create and interpolate a continuous parameter through the use of the value from the measured parameters to predict values for each location in the landscape. In this study, the sample points are taken from groundwater wells penetrating the usage aquifer. Geostatistical analyst provides two groups of interpolation techniques: deterministic and geostatistical. All methods rely on the similarity of nearby sample points to create the surface and parameter predictions. In this study, the geostatistical analyses were performed using ArcGIS Geostatistical Analyst (ARCGIS10.7) software. The ordinary Kriging method was used because the value at the unsampled point can be predicted by a linear weighting of the variation between the surrounding points derived from variogram analyses.

A WQI is a tool to assess the status of water quality at certain times and locations (Singh 1992). The process that is generally followed to compute a WQI consists of three different steps:

• (1)

  The selection of the water quality parameters that are considered in index calculation;

• (2)

  Each of the parameters is given a weight (W_i) according to its significance in the overall quality for drinking purposes. Parameters like pH, Ca²⁺, Mg²⁺, and SO₄²⁻, were assigned weights between 2 and 5 depending on their significance in water quality determination. The maximum weight of 5 has been assigned to the parameters such as TDS, Na⁺, Cl⁻, NO₃⁻, HCO₃, and TH due to their major importance in water quality assessment (Tiwari et al. 2014). K⁺ is given the minimum weight of 2 as it plays an insignificant role in the water quality assessment.

• (3)
  The relative weight (W_i) for each parameter is computed according to Tiwari & Mishra (1985). The calculated relative weight value of each parameter is presented in Table 1.

  

where, SI_i is the sub-index of the i^th parameter; q_i is the rating based on the concentration of the i^th parameter, and n is the number of parameters.

The statistical summaries of the physicochemical parameters of the groundwater samples from the 246 wells in the study area are presented in Table 2.

TDS in the groundwater ranged from 229 to 4,211 mg/L with an average value of about 1,054 mg/L. 40% of groundwater samples were found to exceed 1,000 mg/L, the permissible limit of JISM (2015). The EC of water ranged from 358 to 6,580 (μS/cm) with an average value of 1,647 (μS/cm). The EC, as well as the TDS values, appear to be positively skewed (Figure 6), indicating that it is distributed towards the higher values. 40% of the Values of EC are exceeding the permissible limit (1,500 μS/cm) according to the JISM (2015) for drinking water.

The pH value has been known to influence the dissolution of minerals in a groundwater system as well as affect the quality of water for various purposes. In this study, it ranged between 6.5 and 8.5 with an average of 7.39 not exceeding thus the permissible limit range of JISM (2015) as shown in Table 2.

Most of the samples appear to be in the neutral range and will make HCO₃⁻ the dominant carbonate species, and the carbonate rock-forming minerals dolomite [CaMg(CO₃)₂] and calcite (CaCO₃) are most likely dominant and are involved in the buffering mechanisms in the study area. TH recorded is in the range between 60 and 1,677.56 mg/L with an average value of about 529.71 mg/L, which shows soft to very hard types of the samples (Todd & Mays 1980).

On the other hand, the Durov diagram is based on the percentage of the major ions in meq/L. Both the positive and the negative ion percentages total 100%. The values of the cations and the anions are plotted in the appropriate triangular and projected into the square of the main field to display some possible geochemical processes that could affect the water genesis. Durov diagram for the major cations and anions plotted by Rockworks 22.0 software (Colorado, USA) is illustrated in Figure 7(b). The fields and lines on the diagram show the classifications of Lloyd & Heathcoat (1985). The result of the Durov plotting indicates most of the samples located on the simple dissolution or mixing line trend, where the Ca²⁺ and HCO₃⁻ ions are dominant indicating the recharge water. Along the groundwater flow the samples show no dominant anion or cation, which indicates water exhibiting simple dissolution from the aquifer material and mixing water. However, some samples show that Na and Cl are dominant, which indicates end-point water.

Ca²⁺ concentrations ranged from 15.16 to 372 mg/L with an average value of 138.63 mg/L. 14% of calcium values exceeded the allowable limit (200 mg/L) of JISM (2015). Magnesium concentrations (Mg²⁺) ranged from 2.46 to 310 mg/L with an average of 60.86 mg/L. Mg²⁺ concentrations exceeding the maximum permissible limit (150 mg/L) of JISM (2015) represent 3.5% of the studied samples. The dissolution of carbonate minerals (calcite and dolomite) in the B2/A7 aquifer contributes mainly to the enrichment of these elements in AZB waters.

Sodium (Na⁺) and potassium (K⁺) concentrations ranged from 2.78 to 661.71 and 0.0 to 46.80 mg/L, respectively, in the collected samples. 6% and 2% of samples have Na⁺ and K⁺values exceeding the permissible level (200 mg /L) for drinking water (JISM 2015), respectively. Sodium input into the groundwater is attributed to evaporites and the effect of cation exchange between Ca and Na in the aquifer rocks. Potassium in the groundwater is largely controlled by fertilizers and clay minerals in the rock beds. Elevated K⁺ concentrations may cause kidney problems, hypertension, and nervous disorder (Singh et al. 2008).

Bicarbonate (HCO₃⁻) concentration in the collected samples ranged from 65.05 to 1,146.80 mg/L. Among these 86% of values exceeded the permissible limit (250 mg/L) of JISM (2015). High concentrations of HCO₃⁻ may come, most likely, from the dissolution of limestone into the water, and least likely, from the dissolution of atmospheric CO₂ during recharge (Sheriff & Hussain 2012). On the other hand, the sewage systems may reach the groundwater and contaminate it, leading to this increase in the concentration of HCO₃⁻.

Chloride ion appears to be the most abundant anion in the study area. The Cl⁻ concentration ranged from 26.98 to 1,493.13 mg/L with an average value of about 374.92 mg/L. 29% of chloride values exceeded the permissible limit (500 mg/L) of JISM (2015). The high presence of Cl⁻ in water is attributed to agricultural activities and overexploitation of the aquifer. Increasing Cl⁻ concentrations may cause serious issues to human health (indigestion, kidney, palatability) (Singh et al. 2008; WHO 2017). It is clear from the spatial distribution maps (Figure 9(c) and 9(d)) that sodium and chloride are following the same trend in their distribution. Both increase along with the groundwater flow towards the north and northeast of the study area.

Sulfate (SO₄²⁻) concentrations range from 2.90 to 711.50 mg/L with an average value of 119.32 mg/L Figure 9(e). SO₄²⁻ values exceeded the 500 mg/L allowable limits (JISM 2015) in 2% of samples. The high sulfate content in some wells is attributed to the dissolution of gypsum and anthropogenic activities like irrigation return flow and wastewater from septic tanks.

Nitrate (NO₃⁻) concentration ranged from 0.06 to 237.42 mg/L with an average value of around 45.59 mg/L. 46% of these values exceeded the 50 mg/L permissible limits of JISM (2015) and WHO (2017) for drinking water. This increase in nitrate concentrations in the groundwater is attributable to using fertilizers (chemically and naturally) for agriculture activities. In addition, the influence of wastewater effluents on the groundwater in the area affected the level of nitrate concentration (Al Kuisi et al. 2009, 2014). The spatial map for NO₃⁻ (Figure 9(f)) indicates elevated concentrations of nitrate due to anthropogenic activities and their increases along with the groundwater flow. However, increasing NO₃⁻ concentration in drinking water may cause hypertension, goiter, and gastric cancer (Manivasakam 2005).

The obtained chemical data is plotted on a Piper diagram (Figure 8(a)). This figure shows four water types as follows:

• (1)

  Type 1: Earth alkaline water with increased portions of alkalis with prevailing sulfate and chloride;

• (2)

  Type 2: Earth alkaline water with increased portions of alkalis with prevailing bicarbonate;

• (3)

  Type 3: Alkaline water with prevailing sulfate-chloride; and

• (4)

  Type 4: Normal earth alkaline water with prevailing bicarbonate.

As precipitation water, which contains CO₂ gas, percolates through the soil horizons and travels through rock media down to the aquifer, it dissolves minerals and carries the dissolved particles along its path. As such, TDS values are usually lowest at points of infiltration, which are considered recharge zones, usually with TDS values similar to those of the precipitation in the study area. However, due to many factors that influence the chemistry of the groundwater like anthropogenic activities (over pumping, irrigation return flow, and wastewater) (Al Kuisi et al. 2009, 2014) the water will contain the highest TDS at the discharging points.

For a better understanding of the hydrochemistry of the studied water, the Durov diagram was used (Figure 8(b)). As shown in this diagram the mixed water type prevails in the study area where 70% of the samples plot in field 5 of Durov plot along with the dissolution or mixing line with no dominant anion or cation. In addition, 21% of the remaining samples showing Ca²⁺ and HCO₃⁻ as dominant cation/anion, indicated that this dominance is related to ion exchange. The rest of the samples (9%) exhibit Cl⁻ and Na⁺ which can be accounted for due to the reverse ion exchange of Na⁺- Cl⁻ waters.

The spatial distribution map for the study area shows areas of high salinity are located mostly in the northern part of the study area. The groundwater in this area is highly depleted and the pumping rate exceeds the safe yield of the aquifer leading to deeper salty water abstraction. Moreover, some localities show signs of water salinization due to the infiltration of irrigation return flow and wastewater contaminants.

In this study, the analytical data was assessed by a three-step approach for the all sample populations: (1) univariate analysis (descriptive analysis of individual chemical parameters); (2) characterization of major ion concentrations using traditional geochemical plotting tools; and (3) multivariate analysis (factor and cluster analysis). Cluster analysis and factor analysis have been used to identify inter and intra-correlations of measured parameters and their probable sources of elements in groundwater.

In addition, a strong correlation appears between Ca²⁺, Mg²⁺, Na⁺, Cl⁻, and SO₄²⁻, indicating that common sources constitute the main pollution for groundwater, particularly in areas with excessive use of fertilizers and dominated by industrial activities. Moreover, overexploitation is the principal cause of water quality degradation in irrigation and highly urbanized areas.

To add, there is a high correlation between Na⁺, Cl⁻, SO₄²⁻ and NO₃⁻ that can be accounted for as a result of the same anthropogenic activities (Figure 11).

The dendrogram shows close associations between TH, chloride, and carbonate, which suggest the impact of carbonate dissolution of infiltration water. The second group, which forms cluster 2, includes the parameters pH, Ca²⁺, Mg²⁺, Na⁺, K⁺, NO₃⁻, and SO₄²⁻. This group represents 20% of the total samples and suggests possible anthropogenic activities which led to the increase in these parameters in the groundwater. In the northeastern and central parts of the study area where the farmers are using fertilizers, primarily NKP and the natural manure from chickens, sheep, and goats. Moreover, the infiltrated water in these areas has a great influence on groundwater parameters. The third cluster shows the similarity between TDS and EC. This association suggests the domination of groundwater by infiltration water and its associated interaction with atmospheric CO₂ and CO₂ released from organic matter decomposition from manure, with all parameters contributing significantly to the EC. Besides, over-pumping from the aquifer system will lead to this increase. This cluster represents 18% of the total samples and it's dominated by sodium and chloride concentrations.

Factor analysis is a statistical technique that attempts to reduce a large number of variables into a smaller number of underlying factors. In this study, factor analysis is performed for all samples using the Statistica 13 software package. The data were subjected to varimax factor rotation. The extracted factors have a variance percentage of 78.5, which means 78.5% of the measured parameter variability is attributed to these factors (Table 4).

Table 4

Varimax rotation factor analysis for water samples

Red highlighted correlations represent high loading value on each factor.

View Large

Factor 1. Salinity Factor: This factor has high loading on electrical conductivity, chlorides, and hardness, and the correlated variables such as Ca²⁺, Mg²⁺, Na⁺, K⁺, SO₄²⁻_, and NO₃⁻. It is assigned to geogenic influence in the groundwater because of the rock-water interaction. The highest amount of electrical conductivity and total dissolved solids indicate the mineral dissolution in the water especially contributed by calcium and magnesium-bearing minerals (Al Kuisi et al. 2014).

Factor 2. Carbonate Factor: This factor shows positive loading values of HCO₃⁻, K⁺, and negative values of SO₄²⁻. This factor is highly influenced by the HCO₃⁻ formation, which is derived from surface water leaching to groundwater in addition to anthropogenic mixed sources such as human waste, manure waste, and agricultural runoff.

Factor 3. pH Factor. This factor shows high negative loading values of pH and could be attributed to the effects of anthropogenic activities like using manure and fertilizers. Moreover, the northeastern part of the study area is cultivated by using treated wastewater from Kheribit As-Samara Waste Water Treatment Plant.

A low value of WQI signifies excellent quality while a high value signifies poor quality. Parameters exceeding the permissible are considered responsible for high WQI value. 12% of water samples lie in the ‘excellent’ category, 53% in the ‘good’ one, 31% in the ‘fair’ one, and the remaining 4% in the ‘poor’ category. The obtained WQIs were then contoured using geostatistical techniques to pinpoint areas of excellent and poor water qualities (Figure 14).

Generally speaking, excellent waters are present in the southern part of the study area, whereas poor waters prevail in the northern part of the basin.

The measured values of 246 samples from the Middle Aquifer Complex (B2/A7) in the study area were evaluated and assessed for their water quality for drinking purposes. The average chemical composition for the analyzed samples exhibits cation dominance in the following decreasing order Na⁺ > Ca²⁺ > Mg²⁺ > K⁺ and anion content in the following decreasing order Cl⁻ > HCO₃⁻ > SO₄²⁻ > NO₃⁻.

The bivariate statistics of the physiochemical parameters of the groundwater indicated a strong correlation between EC and Ca²⁺, Mg²⁺, Na⁺, SO₄²⁻_, Cl−, and NO₃⁻, indicating that dissolution of these ions is the main source of the groundwater salinity. In addition, a good correlation exists between EC and NO₃, indicating that salinization of groundwater is taking place, namely, in the northern part of the study area. Nitrate is the most important water pollutant globally as it enters the groundwater from anthropogenic inputs and can reflect the degree of impact of human activities on the water environment. The spatial distribution of nitrate concentration reaches the highest value in the northern part, where this area is highly cultivated, and is exceeding the maximum permissible limits of JISM and WHO. The central part of the study area also presents a high concentration of nitrate due to the direct seepage of the polluted surface water from Zarqa river which mainly includes sewage water flows from the urban areas surrounding the river and from the treated wastewater from Ain Ghazal treatment plant.

The present study demonstrated the importance of multivariate statistical analysis in groundwater studies for drawing a meaningful conclusion. Factor analysis extracted three factors that explain 78.5% of the total variance of the original data matrix. These three factors explain clearly the hydrochemistry of the study area groundwater. The most important factor with approximately 54.6% variation could be due to the source of the ‘salinity’. The second factor with approximately 14.4% could be due to the ‘anthropogenic influences’. The third factor with approximately 9.5% variation is due to the ‘pH’ interaction effect. Both the hydrochemical and statistical classification techniques have indicated the presence of a dominant type of water. CA showed that the groundwater quality gradually improves from the northern to the southern parts of the study area.

Finally, the calculated water quality indices ranged from 30 to 335 in the study area, with the lowest values located in the northern part of the study area, and the highest values in the southern part. The chemistry of groundwater is dominated by cations and anions pairs such as (Na⁺ and Ca²⁺) and (Cl⁻, HCO₃⁻), respectively. 12% of the samples were plotted in the ‘excellent’ category, 53% of the groundwater samples were plotted within the ‘good’, 31% in the ‘fair’, and 4% were plotted in the ‘poor’ categories.

The combined impact of many different factors that characterize the water quality and the challenges of classifying the significant parameters used to measure the status of water resources quantitatively are very complex to understand. However, the WQI is considered a mathematical tool that significantly minimizes the complex water quality data sets and provides a single classifying value that describes the water quality status of water bodies or the degree of pollution. Therefore, the produced WQI final map can be a guide for the decision-makers in the water sector in Jordan. This map will help them in planning for any development projects or urban development in the study area because water quality is a significant criterion in matching water demand and supply and its an assessment for pollution control.

The authors would like to extend their thanks to the Ministry of Water and Irrigation of Jordan for providing the active wells coordinates and sampling permission. The authors are highly indebted to the anonymous reviewers of the journal for their valuable comments, and suggestions that improved the manuscript. This research has been accomplished during the sabbatical leave offered for Prof. Dr Mustafa Al Kuisi from the University of Jordan starting October 2021 to September 2022.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.