Groundwater quality assessment using the water quality index: case study in the north-western part of Drini i Bardhë River basin, Kosovo Open Access

K

constant of proportionality given by the relation (2) (Kalavathy et al., 2011)

Q_n

sub-index

S_n

standard acceptable value of the nth parameter

T

temperature

V₀

actual value of the parameter in pure water

V_n

estimated value

W _n

weight of the nth parameter

The continuous increase in the needs for drinking water, technological or agrarian, is faced in many cases with limited opportunities to meet these public and private needs. Expansion of residential areas and an increase in industrial activities constitute an increasing stress for the quality of water bodies (Çadraku, 2014). This promotes the necessity of undertaking scientific studies of quantitative and qualitative of water resources nationwide and beyond. Therefore, through this study, it is intended to determine water quality using the water quality index (WQI) method (Brown et al., 1970, 1972). Groundwater is an important natural resource of water for catering to the domestic, agricultural and industrial needs of many countries in the world. Groundwater resources are one of the safest sources of drinking water supply. Water quality depends on water composition, which is influenced by natural processes and human activities. The quality of drinking water indicates water acceptability for human consumption. Demographic growth and overall economic development have increased industrial and agro-productive capacities and producers at the country level, posing an increased pressure on the use of groundwater. Water quality is characterised on the basis of water parameters (physical, chemical and microbiological), and human health is at risk if values exceed acceptable limits (WHO, 2012). To determine the drinking water quality for end users, the WQI is a significant parameter and one of the best active tools for sharing the details of each water body with policymakers. Political decision makers, non-technical managers of water resources and the public in general usually have neither the time nor the knowledge required for interpretation of the traditional, technical reports on the state of water quality (Mladenović-Ranisavljević and Žerajić, 2018). Consequently, the WQI becomes an essential parameter for the management and assessment of groundwater. WQIs vary according to the water quality parameters used in the index, as well as the algorithm of calculation and the scale of the water quality rating (Feng et al., 2016). The WQI method numerically sums up a set of different water quality parameters into one value (SWA, 2007). It is primarily developed as an auxiliary tool for summarising data on water monitoring and reporting these to the wider public, which is why it is rather general in nature (Brown et al., 1970). Water quality, in fact, expresses the convenience of water to support a variety of uses or processing. The WQI is used to assess the suitability of surface water and groundwater for drinking and agricultural purposes (Alastal et al., 2016). What is common to all of the aforementioned WQI models is that each individual, specific application has certain requirements concerning the physical, chemical and biological characteristics of water (Alam et al., 2007). Horton (1965) proposed a WQI to describe the suitability of water for human consumption, distinguishing five water classes according to water quality – excellent, good, weak, very weak and unusable – which are easy to understand for decision makers and consumers. The same WQI method was used (Çadraku et al., 2013) for the assessment of the surface water quality of the Drini i Bardhë basin (Kosovo), which found that the surface water of the Drini i Bardhë River basin mostly falls in medium (WQI = 50–70) and good (70–90) water quality categories. Despite the studies conducted so far on the quantitative and qualitative assessment of water in Kosovo (Çadraku, 2018, 2021; Çadraku et al., 2013, 2016; Rizani et al., 2016), other studies should be carried out on the physico-chemical and biological parameters of water. This study aimed to assess the quality of the groundwater (WQI) in the north-western part of the Drini i Bardhë River basin according to the method of Brown et al. (1972). Today, a set of methods is used for water quality assessment, such as the WQI method according to Brown et al. (1970), National Sanitation Foundation (NSF) WQI, Canadian WQI and Serbian WQI. Considering that the WQI model (Brown et al., 1970) is a popular tool for water quality assessment and provides a standardised method for comparing the water quality of different water bodies, therefore, in this study, this method was used. On the other hand, all methods related to water quality assessment derive a WQI that provides a suitable tool for summarising complex water quality data and facilitating the communication of these data to a general audience. This study highlighted how the variation of different parameters can be reduced to a single number when reported with the help of a WQI, thus making it quite suitable for interpretation regarding water quality. The WQI also provides a more appropriate way to express the quality of water resources for consumption by using water quality data, making it very useful for modifying water-management and resource-management policies.

The study area, the Dukagjini basin, is the richest region of water, surface and groundwater, in the Republic of Kosovo (Çadraku, 2014). The Drini i Bardhë River basin represents one of the most important water resources in this basin. The Drini i Bardhë River basin has an area of 4340.14 km² (Kepa, 2022) and is in the western part of Kosovo. The study area is located in the north-western part of this basin between the coordinates 42°44′00″–42°55′00″ north and 20°12′00″–20°45′00″ east (Figure 1) and has an area of 874.85 km². The hydrographic network is relatively developed with the main river Drini i Bardhë, which has a water flow of 61.7 m³/s, and the Istog River, with a water flow of 6.98 m³/s (Kepa, 2020). The water sources are formed in the karst aquifer, are of lithological contact source type and drain the water gravitationally. The most popular sources in the study area are the Radavc source, with an average annual flow of water of 3.5–8 m³/s; the Vrelle–Istog source, with an average annual flow of 2.5–4 m³/s (IHJC, 1983); and the Vrelle source in the Vrelle locality, with an average annual flow of 0.5–1.5 m³/s (Peric, 1978). The surface and underground waters are fed by atmospheric precipitation (rain, snow etc.). This is shown by the fluctuation of water flows both in the surface hydrographic network and in springs and underground water wells. A total of 39 289 inhabitants live in the study area. The agricultural sector is actually developed with an area of 20 653.91 ha, of which 19 940.37 ha is made up of areas used by the agricultural sector (Municipality of Istog, 2022). The impact of pollution is evident in surface waters due to the fact that over 95% of polluted urban and industrial water is released into the receiving environment (rivers, streams etc.) without being treated beforehand. Likewise, the impact of diffuse pollutants is evident, taking into account that about 97% of the earth’s surface has agricultural activity. This statement is valid considering the data presented in Table 1, which shows the water quality of the 50 samples analysed based on the WQI.

The area consists of the several small mountains streams; water flows into tributaries and the Drini i Bardhë River. In geotectonic terms, Kosovo is located within the Dinaric Mountains, and all three divisions of rocks – namely, magmatic, sedimentary and metamorphic – of Precambrian to Quaternary ages are present in the territory of Kosovo (ICMM, 2006a; Mekshiqi et al., 2009; Sikosek, 1971). The sediments of the rivers in Kosovo are composed of alluvial deposits containing largely varying proportions of unconsolidated to semi-consolidated sand and gravel materials (ICMM, 2006a; Sajn et al., 1998). As regards hydrogeological aspects and geological, lithological, structural and tectonic characteristics, the following areas of the aquifers are distinguishable in the study area: intergranular porous aquifers, porous cracks, porous karst and waterless terrain (Çadraku, 2014; ICMM, 2006b). The geological construction and the hydrogeological characteristics of the study area affect whether surface and underground waters have good hydraulic communication links, particularly in the lowland part of the river valleys where there are alluvial sediments (sand and gravel). The character of mutual hydraulic communication is variable. In dry periods, when groundwater levels drop, the river feeds the alluvial aquifers in the study area. Favouring this statement are data on some chemical parameters of underground water that are indicators of pollution (ammonium (NH₄⁺), nitrogen dioxide (NO₂), phosphate (PO₄³⁻), chlorine (Cl), biochemical oxygen demand, chemical oxygen demand, viruses etc.), which, in the dry periods of the year, are the same as those of river water (Malcom, 2008–2010).

The WQI is the most effective tool for monitoring surface water as well as groundwater pollution, and it can be used efficiently in water quality improvement programmes (Saleem et al., 2016). According to Diersing (2009), water quality comprises the overall characteristics (physico-chemical and biological) of water. The weighted arithmetic index method was utilised depending on the measured values of the physico-chemical parameters for estimating the WQI of the sampled well water (Brown et al., 1972). In this study, nine physiochemical parameters were chosen for the calculation of the WQI based on the drinking water quality standards of the World Health Organization (WHO). The WQI is based on the three equations, which play an important role in the calculation the index, which is discussed in the following.

A location of the study area and the sampling sites are shown in Figure 1 and Table 2, respectively. Water samples were collected from 50 different wells (Figure 1 and Table 2) in November 2018–January 2019. Different physico-chemical water quality parameters of the samples were examined by using standard protocols of the American Public Health Association (APHA, 2005). The standard sampling method was used for water collection; water samples were taken from the bottom of aquatic sources (at different depths). Geographic information systems are a new technology used for analysing and interpreting the distribution of pollutants in environmental studies (Khosravi et al., 2021; Merchant; 1994). The inverse distance weighted technique is one of the techniques applied by the ArcGIS software to obtain the spatial distribution of pollutants, which, based on the distance between points and the concentration of pollutants at each point, simulates pollutant concentrations in other parts of the studied area (Fallahzadeh and Ghadirian, 2018). Figure 2 shows the work methodology.

A set of 13 water quality parameters, shown in Table 3, was taken to generate the WQI. The WQI is calculated with the help of the index method (Table 4) (Brown et al., 1972).

Step 1. Calculate the unit weight (W _n) factors for each parameter by using the formula

where

S _n is the standard desirable value of the nth parameter.

On summation of the unit weight factors of all selected parameters, W _n = 1 (unity).

Step 2. Calculate the sub-index (Q _n) value using the formula

where Q _n is the sub-index, V _n is the estimated value, V ₀ is the actual value of the parameter in pure water (generally V ₀ = 0 for most parameters except for pH) and S _n is the standard value.

Step 3. Combining step 1 and step 2 the WQI is calculated by using the following expression:

Water is the most important human need, and its conservation is of great importance. The assessment of surface and underground water quality (partially) was carried out by several authors (Bytyçi et al., 2018; Shala-Abazi et al., 2020) by application of the WQI method. Shala-Abazi et al. (2020) evaluated the WQI of Sitnica River water (central part of Kosovo) and showed that the WQI varies from 46 to 95, classifying these waters in the marginal category. Bytyçi et al. (2018) established that the WQIs of the waters of the Lepenc and Nerodime Rivers (southern part of Kosovo) ranged from 36 to 76, classifying their water quality as varying from moderate to poor. The same WQI method was used (Çadraku et al., 2013) for the assessment of the surface water quality of the Drini i Bardhë basin (Kosovo), which found that surface water of Drini i Bardhë basin mostly falls in the medium (WQI = 50–70) and good (70–90) water quality categories. The study area of the study drains water into the Drini i Bardhë River. For evaluation of water quality in the north-western part of the Drini i Bardhë River basin, the method by Brown et al. (1972) was applied. The values of water quality parameters in the present study were taken at 50 sampling stations, of which 48 are wells and two are springs; the results are presented in Table 2. The quality of groundwater has changed dramatically due to the release of pollutants in the environment and the excess consumption of resources (Eslami et al., 2015). A major influence on the chemical composition of groundwater in this study area is the activities of humans. Land use practices can directly influence groundwater quality, particularly in shallow and unconfined aquifers where there is a greater hydraulic connectivity between surface and unconfined aquifers.

• Temperature. Temperature is one of the main properties of water. It affects the development of chemical processes in water and the amount of gas and salts dissolved in it. Temperature affects the chemical composition of water. Increasing the temperature also increases the solubility of minerals and reduces the percentage of dissolved gases. An increase in temperature accelerates the chemical reactions that take place in water. The water temperature in the study area ranges from 5.2 to 15°C with an average value of 11°C (Table 3). This shows that shallow groundwater temperature is affected by air temperature fluctuations and average annual air temperatures. Based on the results obtained, the groundwater in this aquifer is cold to warm and within the limit values of drinking water according to Administrative Instruction (AI) no. 10/2021 (The Goverment of Republic of Kosovo, 2021) and the WHO standard (Table 5). Temperature fluctuation from this aspect comes as an effect of the impact of monthly and annual air temperatures, as well as the impact of the coverage of the basin space. The temperature analysis (variation) of groundwater in the study area is shown in Figure 3(a).

• pH. The pH of a solution is the negative common logarithm of the hydrogen (H⁺) ion activity: pH = −log(H⁺). The pH of water is a measure of acid–base equilibrium and, in most natural waters, is controlled by the carbon dioxide (CO₂)–bicarbonate (HCO₃⁻)–carbonate equilibrium system (WHO, 2007). The effects of acids and alkalis depend on the strength of the acid or alkali and the concentration. Strong concentrated acids or alkalis are corrosive, whereas dilute and weak acids and alkalis are not corrosive. In the study area, the pH of the groundwater samples varied from 6.39 to 8.65 with an average value of 7.74 mg/l (Table 3). The spatial and temporal variations of pH are controlled by the quality and the infiltration rate of recharge water, the replenishing water rate and water–rock interactions in the aquifer. The pH values in the study area are all within the desirable limits set by the WHO and AI no. 10/2021), except in SP32, where the pH value is 6.39. Based on the measured pH values, it is shown that the groundwater in the study area is slightly alkaline water. pH is an important parameter in water body since most of the aquatic organisms are adapted to an average pH and do not withstand abrupt changes. The pH analysis (variation) of groundwater in study area is shown in Figure 3(b).

• Electrical conductivity (EC). EC is the ability of water to conduct electricity. The EC of water is a direct function of its total dissolved salts (Harilal et al., 2004). It depends mainly on the geological structure – that is, the medium through which water flows – than on inorganic solution suspended substances in water such as chlorides (Cl⁻), nitrates (NO₃⁻), sulfates (SO₄²⁻) and phosphates (ions that carry a negative charge) or sodium (Na⁺), magnesium (Mg²⁺), calcium (Ca²⁺), iron and aluminium ions (ions that carry a positive charge). In the groundwater of the study area, the EC ranges from 267 to 1745.84 μS/cm, with an average value of 651.47 μS/cm (Table 3). The EC values were compared with AI 16/2012 values (Table 5), and it turns out that all are within the limit values of this instruction. The EC analysis (variation) of groundwater in study area is shown in Figure 3(c).

• Total dissolved solids (TDS). High values of TDS in groundwater are generally not harmful to human beings, but high concentrations of these may affect people, who may suffer from kidney and heart diseases (Gupta et al., 2004). The range of TDS levels in the study area is 137–877 mg/l, with an average value of 327.28 mg/l (Table 3). Based on comparisons of TDS values in the study area with the standards by WHO (2012), it was found that for 50 water samples, four or 8% of the samples have a total value above the limit of 500 mg/l, while 46 or 92% are below the value of 500 mg/l of the WHO standard. Therefore, water containing higher than 500 mg/l TDS is not considered desirable for drinking water. The TDS analysis (variation) of groundwater in the study area is shown in Figure 3(d).

• Total hardness (TH). The hardness of water mainly depends on the amount of calcium or magnesium salts or both. In natural water, TH is contributed mainly by dissolved calcium and magnesium ions (Ikomi and Emuh, 2000), with all other divalent cations contributing to its value. Water hardness is the traditional measure of the capacity of water to react with soap, with hard water requiring considerably more soap to produce lather (WHO, 2011a). The TH in the study area varies from 120.89 to 744.19 mg/l, with an average value of 324.47 mg/l (Table 3). Based on the comparisons of chemical constituents with WHO (2012) standards, it is found that for 50 water samples, three or 6% of the samples have a TH value above the limit of 500 mg/l. The highest values were identified in wells SP3 and SP6, while the lowest in that in well SP37. The TH analysis (variation) of groundwater in the study area is shown in Figure 3(e).

• Calcium. Over 99% of the total body calcium is found in bones and teeth, where it functions as a key structural element (Cotruvo and Bartram, 2009). Food is the principal source of both calcium and magnesium. Both calcium and magnesium are essential to human health. The calcium ion is considered the most abundant ion in freshwater; moreover, it is significant in the precipitation of lime in plants, bone building and shell construction (Qureshimatva et al., 2015). Values of calcium ions range from 35.4 up to 230.2 mg/l, while the average is 98.77 mg/l. Calcium values in the study area were compared with WHO standard values (Table 3), and it turned out that two or 4% of the samples are below the 200 mg/l limit, while only 48 samples or 96% are above the WHO standard value of 200 mg/l. The calcium ion analysis (variation) of groundwater in the study area is shown in Figure 3(f).

• Magnesium. Magnesium is considered an alkali earth metal and is a cause of water hardness. This ion comes from groundwater by dolomite and alienation of crystalline mineral rocks (olivine, pyroxene, amphibole, mica, serpentinite, magnesium) (Çadraku et al., 2016). The concentration of magnesium ions in the study area groundwater ranges from 2.4 to 95.8 mg/l, with an average value of 18.91 mg/l (Table 3). Magnesium ion values for 50 samples were compared with WHO standard values, and it turned out that all are below 100 mg/l. The magnesium ion analysis (variation) of groundwater in the study area is shown in Figure 3(g).

• Sodium. The sodium ion is ubiquitous in water (WHO, 1996). According to standards (WHO etc.) for drinking water, sodium may affect the taste of drinking water at levels above about 200 mg/l. The sodium content in the study area ranges from 0.53 mg/l (SP10, spring) to 61.61 mg/l (SP40, well), with an average value of 7.74 mg/l (Table 3). According to WHO data, most water supplies contain less than 20 mg of sodium per litre, but in some places, levels may exceed 250 mg/l. Therefore, referring to these values, it turns out that the sodium content in the groundwater of the study area is within the limit values to be used for drinking water. Also, the sodium values in the study area are within the AI no. 10/2021 limit values. The sodium ion analysis (variation) of groundwater in the study area is shown in Figure 3(h).

• Potassium. Potassium is an element with chemical properties similar to those of sodium, but the potassium (K⁺) ion content in water is many times lower than the sodium ion content. Clay minerals, feldspar and some micas are considered major natural sources of potassium in groundwater (Alikhan et al., 2020). In the groundwater of the study area, potassium comes mainly from the decomposition of organic matter. The values of potassium ion in the groundwater in this area range from 0.43 to 78 mg/l, with an average value of 4.61 mg/l (Table 3). Potassium ion values are compared with WHO standard values (Table 5), and it turns out that two samples or 4% are above the limit values, while 48 samples or 98% are below the WHO standard values. The potassium ion analysis (variation) of groundwater in the study area is shown in Figure 3(i).

• Bicarbonates. The bicarbonate ion usually correlates positively with calcium and magnesium ions. The amount of bicarbonates in groundwater is also affected by the pH of the water. Bicarbonates in the groundwater of the study area vary from 63 to 488 mg/l, with an average value of 303.59 mg/l (Table 3). Higher bicarbonate values are found at measuring station SP30 (well). According to the WHO standard, it turned out that in 50 samples, the values of bicarbonates are below the value of 125 mg/l. This ion probably comes from dissolved carbon dioxide in water and dissolution of limestone and dolomite, hydrolysis of silicate minerals and bacterial reduction of sulfates and anthropogenic pollution. The bicarbonate analysis (variation) of groundwater in the study area is shown in Figure 3(j).

• Sulfates. Sulfate is an important chemical factor for water quality and has an effect on the odour and taste of water (Bouslah et al., 2017). It is a characteristic of shallow groundwater. In groundwater, it comes from dissolution of sulfate rocks and oxidation of sulfide minerals. Also, it can enter shallow groundwater from decomposition of plant and animal substances that have sulfur in their composition (Çadraku et al., 2016). The sulfate values in the study area ranged from <0.1 mg/l (in SP8, SP9, SP10 and SP11, two springs and two wells) to 288 mg/l (PS40, well), with an average value of 39.66 mg/l (Table 3). All measured values and the average value of sulfates (39.66 mg/l) in the study area are below the value of 250 mg/l of the WHO standard, except that of SP40, which is above the WHO standard. The sulfate analysis (variation) of groundwater in the study area is shown in Figure 3(k).

• Chlorides. Chlorides occur naturally in all types of water; however, their main contributing sources are run-off of inorganic fertilisers from agricultural fields, sewage discharge and so on. Chlorides are important in detecting the contamination of ground water by waste water. According to Agarwala et al. (2012), in people who are not accustomed to high chloride contents, it may cause a laxative effect (Agarwala et al., 2012). The chlorides of 48 well and two spring water samples were found to be between the 9.23 and 192 mg/l, with an average value of 29.62 mg/l (Table 3). The chloride content of the samples was found to be well within the permissible levels of 250 mg/l of WHO standard. The chloride analysis (variation) of groundwater in the study area is shown in Figure 3(l).

• Nitrates. Nitrates are ions that occur naturally and are part of the nitrogen cycle. Nitrate is used mainly in inorganic fertilisers. Nitrate can reach both surface water and groundwater as a consequence of agricultural activity (including excess application of inorganic nitrogenous fertilisers and manures), from waste water treatment and from oxidation of nitrogenous waste products in human and animal excreta, including septic tanks (WHO, 2011b), which is also characteristic of the study area given that most of the inhabitants in this area engage in agriculture. The natural nitrate concentration in groundwater under aerobic conditions is a few milligrams per litre and depends strongly on soil type and on the geological situation. Concentrations of nitrate in rainwater of up to 5 mg/l have been observed in industrial areas. Nitrate values in the waters of this basin range from 0.1 to 182.5 mg/l with an average value of 18.55 mg/l. The maximum (182.5 mg/l) value was found at well measuring station SP16, while the minimum (0.1 mg/l) value was found at spring measuring station SP9 (Table 3). Nitrate values are compared with WHO standard values (Table 5), and it turns out that two samples or 4% are above the limit values, while 48 samples or 98% are below the WHO standard values.

A correlation shows the relationship between the two variables. The study of correlation (correlation) often aims to show the statistical independence of two variables – that is, to prove that they are not related to each other (Selenica, 2009). Correlation coefficients with values of 1 and −1 correspond to functional dependence, while a value of 0 corresponds to statistical independence. Depending on the correlation coefficient, there are different types of relationships between variables: r = 1, positive functional linear relationship (between two variables); r ≈ 1, strongly positive linear bond; r > 0, positive linear relationship; r = 0, no linear relationships, or independent variables; r < 0, negative linear relationship; and r = −1, negative functional relationship. Table 6 and Figure 4 show the results regarding the correlation for physico-chemical parameters in the study area. Within the study period, pH showed a highly significant positive relationship with nitrate. TDS showed a highly significant positive relationship with EC. TH showed highly significant positive relationship with EC and TDS. Calcium ions showed a highly significant positive relationship with EC, TDS and TH.

Hierarchical clustering (or hierarchical cluster analysis) presents an alternative approach to grouping objects based on their similarities. In Figure 5, the hierarchical clustering of 13 parameters in the study area is shown. Figure 6 shows the percentile plot. Principal component analysis was used to calculate the basis of the multivariate data analysis (Tables 7 and 8). Cluster analysis of the R mode (Figure 5) showed mutual links between the studied variables, and it could be observed that the variable EC has the closest association with calcium and bicarbonate ions. After that, the variables calcium ions and bicarbonates form two branches of the dendrogram. The first branch, calcium ions, is linked with the other one, in which sulfates, nitrates, chlorides, potassium ions, T, pH, sodium ions and magnesium ions were linked. In the second branch, bicarbonate is linked with TDS and TH.

Table 1 and Figure 7 show that 27 or 54% of water samples in the study area are at the limit (WQI = 25–50), ranking them in the group of good-quality water. Fifteen water samples or 30% of the total number of samples are on the border (WQI = 0–25), ranking these waters as excellent.

Five or 10% of water samples have a poor water quality (WQI = 51–75), two samples or 4% have a very poor water quality (WQI = 76–100) and one sample or 2% have a water quality unfit for consumption (WQI > 100). In general, the groundwater in the study area evaluated in accordance with this method turns out to be good with a slight tendency towards deterioration of its quality.

Groundwater was collected from 50 sampling points from November 2018 to January 2019. In the study area, the groundwater is odourless, colourless and tasteless. Most of values of the physico-chemical parameters are within the WHO standards, while a few exceed the limits. The physico-chemical parameters show different variations throughout the basin, which are indicated by their respective values of variance and/or standard deviation. According to the method by Brown et al. (1972) the groundwater quality index in the north-western part or River Drini i Bardhë basin ranged from 10.69 to 116.23, with an average value of 36.50. WQI methods showed that these waters are of the better quality, which should be expected due to two factors: firstly, groundwater is to some extent protected by surface contaminations and, secondly, the study area is a peripheral area of the Drini i Bardhë River basin and is not highly exposed to agricultural, industrial and civil pollutants. The spatial distribution of the WQI showed that most of the groundwater in the study area fall within the ‘good’ to ‘excellent’ categories. This study could be a baseline for the authorities to establish a groundwater-management plan in the study area in the future. Cluster analysis of R mode showed mutual links between the studied variables, and it could be observed that the variable EC has the closest association with calcium and bicarbonate ions. After that, the variables calcium ions and bicarbonate form two branches of the dendrogram. The first branch, calcium ions, is linked with the other one, in which sulfates, nitrates, chlorides, potassium ions, T, pH, sodium ions and magnesium ions are linked. The second branch, bicarbonate, is linked with TDS and TH.