Evaluation of groundwater quality for drinking and irrigation purposes using proxy indices in the Gunabay watershed, Upper Blue Nile Basin, Ethiopia

Evaluation of groundwater potential and its quality assessment for drinking and irrigation has recently become a major concern, especially in developing countries due to various constraints. The primary aim of this study is to evaluate the quality of groundwater and establish whether they are safe for domestic and agricultural usage. 78 samples were collected during dry and wet seasons from 39 locations in the Gunabay district of the upper Blue Nile, Ethiopia. The following physicochemical parameters were evaluated successfully (T, pH, EC, TDS, Na+, K+, Ca2+, Mg2+, Fe, Cl−, F−, SO42−, PO43−, CO32−, HCO3−, and NO3−-N). Then, Entropy Weight Water Quality Index (EWQI) and irrigation water quality indices (SAR, %Na, MAR, RSC, PS, KI, PI, and IWQI) were used to assess the distribution of groundwater quality in the study area. The Piper diagram used to characterize the groundwater types revealed that Ca–HCO3 is dominant in the area and rock-water interaction regulates the chemical characteristics of groundwater. Wilcox diagram was used to analyze the salinity level in the groundwater. The findings showed that the groundwater had higher nitrate levels relative to the permissible level of WHO standards due to excessive use of fertilizers in rural areas. Depending on the EWQI approach, the study area was categorized as excellent, good, and medium zones, covering 84.6%, 12.8%, and 2.6%, respectively. The results depict that high-quality drinking water was available in rural areas, n high to medium in the urban regions. The comparative irrigation water indices record 85% of water wells are suitable for irrigation, but some well sites are unsuitable due to higher salinity hazards and deep rock interaction. These integrated water quality indices were effective in validating drinking and irrigation water quality in the study area.


Geological setup
This study is occupied by Quaternary and tertiary rocks with thicknesses of up to 3000 m covering more than half of the country. The early quaternary and tertiary volcanoes are concentrated in the study area [44]. As shown in Table 1, slightly fractured and weathered basalt has a 33.3 m thickness and the top layer is covered with black cotton soil with a small 3 m thickness.

Land cover/uses
Land use (land cover) is one most influencing factor for the spatiotemporal variability of groundwater quality [45]. The dominant land cover in the study area was crops land (human-planted cereals) and shrub land (shrubs, bushes, and tufts of grass, savannas with sparse grasses or plants), which covers 57.04% and 36.16% respectively. Open water, forest, grassland, bare land, and settlement cover 6.8% of the total area. The land cover dataset is extracted from land cover databases of High-Resolution Maps (10:10) (https:// livingatlas.arcgis.com/landcover).

Materials and methods
Constructing physicochemical parameters, developing drinking and irrigation water indices, and mapping groundwater quality potentiality zones of the study were the three research domains that were considered primarily. Matlab, Arc GIS tools, and AquaChem software were used for processing, and EWQI, SAR, RSC, ESP, MAR, KR, PS, PI, and IWQI were the indices calculated in detail (Fig. 3).

Data collection and analysis
In this study, different materials have been used to achieve the defined objectives. The following materials were used in the field: a Garmin GPS H72, a Micro Multi 800, and a water sample container (an icebox). The field samples were taken for two seasons, April 2022 (dry season) and August 2022 (wet season) and the laboratory tests were evaluated using the Palintest photometer 800 and flame atomic absorption spectroscopy.

Sampling of groundwater
A systematic sampling technique was implemented to determine sampling sizes in the study area. 39 sampling sites were selected after considering the topographic setup (high land, escarpment, and lowland), geological setup, road accessibility, water well availability, literature reviews, and financial capacity. Therefore, 78 samples were collected with standard instruments from 8 springs (GW3, GW4, GW8, GW12, GW14, GW36, GW37, and GW38), 13 open wells (GW5, GW6, GW7, GW9, GW11, GW13, GW15, GW16,  GW18, GW22, GW24, GW27, and GW31) and 12 hand-dug wells (GW10, GW17, GW19, GW20, GW21, GW23, GW25, GW26, GW32, GW33, GW34, and GW39), and 6 deep wells (GW1, GW2, GW28, GW29, GW30, and GW35). All shallow and deep wells have a depth of 6m-210 m. The basic parameters such as temperature (T⁰C), conductivity (EC), total dissolved solids (TDS), and pH were evaluated in the field using a micro multimeter. Using a pump to sample deep groundwater well is much more effective and the obtained sample is representative of deep aquifer water when using a pump [46]. This helps in removing the stagnant/polluted water from the wells. Polyethylene bottles of half-liter capacity were used to store sampled water. All sample bottles were stored in ice-packed coolers immediately after collection. The temperature of all stored samples was maintained at 0-4 • C until analyses were conducted.

Laboratory methods
A Palintest photometer was used for chemical parameters and flame atomic absorption spectroscopy (AAS) to measure sodium concentration. The laboratory analysis of chemical parameters has been analyzed at Abay Basin Authority, Amhara Design and Supervision Works Enterprise Laboratory Service, and ORDA, Ethiopia.

Data accuracy assessment
The groundwater sampling data reliability has been checked using ionic balance error (IBE) or reaction error (RE), for good data representation. The results revealed that RE values are within ±5%. The accuracy of the chemical results was determined using equation (1) [42,24].

Water quality standards
To check the quality of groundwater for drinking, physicochemical parameters were evaluated and compared with the maximum permissible limit set by WHO [21].

Analysis of index for drinking purpose
Estimation of groundwater quality techniques has been developed and improved by different methods to get high performance and effective results. Recently, Entropy weighted water quality index (EWQI) has become the most acceptable approach for drinking water [34]. This is a reliable tool to evaluate overall quality for drinking purposes from the individual groundwater quality parameters [15,47]. In this study, the entropy-based groundwater quality index has been applied to evaluate the suitability of groundwater intentionally for drinking.

Entropy weighted water quality index (EWQI)
The main aim of the water quality index is to characterize the overall quality of groundwater. According to the reports of numerous researchers, different water quality indices have been used for domestic groundwater usage [48]. Yet, the available approaches are makeshift by newly accurate techniques in apprehending the goal of representing the water quality. The entropy-weighted water quality Index (EWQI) is one of the most competent methods, which uses the entropy model to assign weights [49,39]. The approach of evaluating entropy weighted was analyzed in five steps, when m water samples (i = 1, 2 …, m) and each sample is analyzed for ''n'' quality parameters (j = 1, 2 …, n), according to observed data [50]. The first step is constructed through matrix X (equation (2)).
Where, X represents the initial matrix value of all parameters, m represents the total number of water samples and n represents the number of hydrochemical parameters.
In the second step, the initial process has to be standardized to remove the influences of dimensions and magnitude. Therefore, the standardized process was evaluated using equation (3), while the normalized process was carried out using equation (4).
Where, xij is the initial matrix; (xij) min and (xij) max are the minimum and maximum values of the hadrochemical parameters of the samples, respectively.
In the third step, the entropy "ej" and entropy weight "wj" have been computed by equations 5-7.
This is entropy information, which can be calculated as in equation (6).
The entropy weight (wj) can be calculated in equation (7).
In the fourth step, the quality rating scale (qi) has been calculated in equation (8).
Cj is the concentration of water quality parameter j (mg/L), and Sj is the maximum permissible limits of drinking water by modified guideline of WHO [51] of parameter j (mg/L).
In the last step, the EWQI has been calculated using equation (9).

Analysis of induces for irrigation purposes
The quality of groundwater may vary in space and time due to the extraction of groundwater, aquifer recharge, and the intensity of rainfall. Therefore, it is important to validate the quality of irrigation water [52], since, toxicity affects sensitive crops, salinity affects crop water availability, permeability affects soil infiltration rate, and miscellaneous effects susceptible crops [53,54].
The suitability of groundwater for irrigation purposes is mainly governed by SAR (Sodium adsorption ratio), %Na (Sodium percentage) or ESP (Exchangeable sodium percentage), PS (Potential salinity), MAR (magnesium adsorption ratio), RSC (residual sodium carbonate), Kelley index (KI), PI (permeability index) [55] and IWQI (Irrigation water quality index) [56,57]. However, as suggested by many agencies and organizations, suitability for this purpose relies on numerous indexes and parameters [58].

Sodium adsorption ratio (SAR)
The initial concept of SAR was proposed by Richards [59], which is used for the detection of sodium hazards in the soil. It is vital for agricultural water classification, since the reaction of sodium and soil, forms sodium hazards. SAR can be calculated using equation (10) [60].

Exchangeable sodium percentage (ESP)
The Na concentration in irrigation water is the result of the exchange of Ca and Mg concentrations, and the process reduces soil permeability. Sodium concentration (% Na) is an important approach for determining the quality of groundwater for agricultural activities. %Na (sodium percentage) or % ESP (exchangeable sodium percentage) can be computed by equation (11) [54].
Na percentage can be classified as excellent quality for values (<20), good quality for values , medium quality for values , doubtful/poor quality for values , and unsuitable quality for values (80-100) [15,54].

Potential salinity (PS)
The potential of salinity in irrigation water is measured by electrical conductivity (EC), which is presented in microsemes per centimeter. Salinity hazards can be classified based on the EC (μS/cm) results: (very high, no detrimental effects for all crops) for the values (<750), (high, may have some detrimental effects for sensitive crops) for the values (750-1500), (moderate, may have adverse effects for many crops) for the values (1500-3000), (severe, not suitable for all crops) for the values (3000-7500) [55,52].

Residual sodium carbonate (RSC)
It is a vital indicator of deep water in the confined aquifer with a higher amount of bicarbonate concentration (equation (12)) [61].

Kelley index (KI)
Kelley's ratio (KR) is known as Kelley index, which was suggested by Kelley [62] for irrigation water quality analysis (equation (14)).
Water is classified into three types based on the Kelly index. If the Kelly index value is less than one, the water is suitable for irrigation. Water is marginally suitable for irrigation if the Kelly index is between 1 and 2, and water is unsuitable if the Kelly index is greater than 2 [30].

Permeability index (PI)
The permeability index (PI) is one of the most important indicators of groundwater suitability for agricultural activities. It is a measure of the ability of soil to move water (permeability). It correlates sodium, calcium, magnesium, and bicarbonate concentrations in soil, which are influenced by long-term irrigation practices (equation (15)).
Where qmax is the upper value of the corresponding class of qi X ij denotes the data points of the parameters (observed value of each parameter) X inf refers to the lower limit of the class to which the observed parameter belongs. q imap refers the class amplitude for qi classes X imap corresponds to class amplitude to which the parameter belongs.
Where, m is the number of parameters considered (qi) is a water quality measurement parameters values and (wi) is the weight of each parameter (equation (17), Table 2).

Graphical representation of water irrigation uses
The much-accepted Wilcox diagram was used to evaluate groundwater quality for irrigation purposes. The vertical and horizontal axes of the diagram are SAR and electric conductivity (EC, μS/cm) which ranges from 0 to 40 and 100-12000, respectively [54,65].

Physicochemical parameters and its seasonal variation
Evaluation of groundwater quality was conducted by considering proxy indices in this research. An entropy-weighted water quality index and graphical representation have been used to validate the water quality for drinking use in both seasons. In the same way, optimal irrigation water quality indices such as SAR, % Na (ESP), PS, RSC, MAR, KR, PI, IWQI, and graphical representation were also applied. The descriptive statistics of physicochemical values at 39 locations of Gunabay watershed during dry and wet seasons are given in Table 3. They revealed that, the variability of chemistry in the groundwater datasets in the area.
The seasonal variation of each physicochemical parameter is not significant in confined aquifers and deep groundwater wells. A slight variation had been observed in the study area. However, the seasonal variability in the two seasons is high for T(⁰C), EC (μS/cm), and TDS (Table 3). Unprotected springs and shallow wells have high NO 3 − -N levels. Due to extensive use of fertilizers, the concentration of nitrate in some shallow wells was above the maximum permissible standard limit [51]. The source of high nitrate concentrations in the groundwater is frequently attributed to anthropogenic activity (domestic waste and agricultural) due to lack of spring protection or wellhead [66]. The result shows that 5% of the samples were acidic in the dry season, whereas 13% of the total samples were acidic in the wet season. Only 2% of the total sample was basic, and more than 74% of the pH values decreased in the wet season (Fig. 4 a and b). This indicates that the locations of groundwater samples may become acidic in the future. Moreover, 83% of the TDS values were higher in the dry season than the wet season. Arbgebya, Este, and Wanka well sites are the most exception sampling aquifers, where saline water intrusion makes groundwater unsuitable for irrigation purposes. Especially, Wanka well has medium sodium hazard (class: S2) and high salinity hazard (class: C3), which indicates the unsuitability of internal rock-water interaction for plant growth (Fig. 5 a and b). Therefore, for better irrigation purposes, such soil primarily needs efficient drainage patterns and good permeability.

Hydrogeochemical analysis
Classification of water type and its characterization of hydrogeochemical processes were represented using a piper diagram that identifies the controlling factors of water chemistry [67,49,68]. Three water types accounted for 69% of the samples: Ca-HCO 3 (left quadrant), Ca-Cl (18% of total samples), and Na-HCO 3 (13% of total samples). As shown in Figure (6 a and b), most results revealed that groundwater in the study area is fresh and recharged. Around 18% of the samples indicated cement pollution or reverse ion

Gibbs plot
Groundwater chemistry can be controlled by physical properties of aquifer lithology (Table 1), weather condition, and bedrock mineralogy, which represents geochemical process of the rainfall, rock weathering, evaporation and precipitation-evaporation dominance [69,14].
As indicated by Gibbs's diagram, the water quality in the study area is mainly influenced by rock weathering process (Fig. 7 a and   Fig. 6. Type of groundwater based on the graphical representation of piper diagram (a) dry and (b) wet seasons. ). This type of groundwater is showing the presence of water rock interaction due to carbonate and silicate rock weathering [14].

Water for drinking purpose
The water quality assurance, control, and water safety plans are responsible for drinking water supply agencies [51]. Therefore, this research provides a detailed investigation of water quality evidence for the communities and policymakers.

Entropy-weighted water quality index
EWQI is an advanced technique to evaluate the status of groundwater quality for domestic uses. The results revealed that 84.6% of the samples had excellent water quality, 12.8% (GW1, GW2, GW3, GW24, and GW28) had good water quality, and 2.6% (GW31) had medium water quality in both seasons (dry and wet) (Fig. 8 a and b). The GW31 has a medium water quality status, which is the lowest in comparison to the others. This is due to that the well is located in Este Town near the Wanka River, and the aquifer is shallow in nature. Due to poor water management practices, untreated water may interact with the river and enter directly. Good quality water is also found nearby Bahir Dar (GW1 and GW2), Arbgbya (GW24), Tisabay (GW3), and Mekane Eyesus (GW28 and GW31). This shows that the overall water quality of urban areas was lower than that of rural areas (urban areas are highly contaminated) [11]. This indicates that except for nitrate concentration, aquifers located near urban areas are more prone to groundwater pollution than rural aquifers. Therefore, urban areas and agricultural lands are the major sources of groundwater pollution [8].

Groundwater for irrigation purposes
Evaluation of water quality and application of proxy indices are essential in the study area because irrigation water quality is susceptible to chemical parameters [63]. Therefore, a graphical representation and eight indices were applied for a better understanding of the current groundwater situation as well as its suitability for agricultural activities (Fig. 4 a and b).
SAR is a common approach to evaluate irrigation water which relies on sodium concentration [70]. Thus, very high sodium concentration in the groundwater harms the soil and is unsuitable for plant productivity. During the analysis, it was found that 3% of samples have very high sodium water resulting in unsuitability for any crops. 10% of the total samples have medium sodium water (good), and 85% of samples have low sodium water (excellent). %Na is a measure of soil permeability in agricultural lands. Soil drainage permeability can reduce the higher sodium availability through irrigation water [71]. In the study, groundwater samples are categorized as excellent (18%), good (46%), medium (28%), and poor (8%). PS indicates the level of soluble salts in the water. PS values were classified as very high (69%), high (28%), and moderate (3%). The samples have moderate water and may have adverse effects on many crops. RSC results showed that groundwater samples (18%) were unsuitable with RSC values ≥ 2.5 meq/L and poor water quality was recorded in about 28% of samples. However, in 54% of the studied region, the groundwater was suitable regarding The KI values varied from 0.08 to 11.34, and thirty (77%) samples are evaluated as suitable for irrigation (KI < 1). Three samples (8%) are considered unsuitable for irrigation (KI > 2). The remaining samples are marginal for irrigation uses. PI is one of the most important irrigation water quality indices to evaluate the long-term effects of soil permeability. Twelve locations of groundwater samples are suitable for irrigation and twenty-seven samples are considered marginal water for irrigation.
According to the IWQI values, 28 water wells have no restriction on irrigating for any type of farming. 8 wells also represent low restriction, and the remaining 3 wells represent moderate quality. Generally, the seasonal spatial and temporal groundwater quality status for drinking and irrigation purposes has been figured out in both seasons (Fig. 8 a and b and Fig. 9 a -p).
Generally, sodium and salinity hazards have a potential to reduce the soil permeability (which restricts plant growth through reduction of water circulation in the soil) and soil salinity (which restricts root to intake salt), respectively [53].

Limitation of the study
Due to limitation of funding and laboratory access, this research work did not focus on heavy metals (trace elements) which linked with public health. Because heavy metals have a potentially harm for human and aquatic environment [72]. Boron concentration was not evaluated in the study area for irrigation purposes. However, the main physicochemical parameters were considered under the guideline of World Health Organization [51]. The next scholars should focus on trace elements. Moreover, suitable and simple groundwater management measure should be required to improve the crop yields and public health in study area.

Conclusions
In the study area, groundwater is the only source to meet the demand for drinking and irrigation purposes. Thus, it is mandatory to validate their water quality status before it enters the distribution system for drinking and irrigation purposes. The physicochemical parameters results were compared with modified WHO guidelines. The laboratory results showed that some nitrate concentration,  TDS, EC, and pH values are not within the standard limits. The seasonal variation of physicochemical parameters between dry and wet was insignificant. However, shallow unconfined aquifer wells have some variation in the case of nitrate, pH, and EC. Based on proxy indices such as EWQI and SAR, %Na (ESP), MAR, KR, PI, RSC, PS, and IWQI), about 85% of the total samples are categorized as suitable water for drinking and irrigation purposes, but a few samples are affected by various geological and anthropogenic activities located near urban areas. The relative water quality status in a rural area is better than in the urban area, except for the contaminated level of nitrate. Excessive use of fertilizers by farmers in rural areas somewhat affects the quality. Hydrogeochemical results revealed that Ca-HCO3> Ca-Cl > Na-HCO3 are the most dominant groundwater types in the study area. Calcium carbonate water is found in shallow aquifers increasing the hardness. Sodium carbonate water is located in deep groundwater which is due to higher exchangeable rock-water interactions. These comprehensive evaluation techniques will provide a map-based groundwater quality status for planners, decision-makers, and water resources managers. More than 85% of samples are almost suitable for drinking and irrigation uses. Excessive application of fertilizers and improper wastewater disposal needs to be monitored on regular basis. Once the groundwater becomes contaminated, it is too difficult to treat. Therefore, monitoring of a large number of physicochemical parameters will be recommended to control the degree of groundwater contamination. Integrally, the model is a useful tool for water resources management for both environmentalists and public health decision-makers.

Author contribution statement
Asnakew Mulualem Tegegne: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data; Wrote the paper.
Tarun Kumar Lohani: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Wrote the paper.
Abunu Atlabachew Eshete: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Wrote the paper.

Data availability statement
Data will be made available on request.

Declaration of interest's statement
The authors declare no conflict of interest.