Assessment of Major Sources Controlling Groundwater Chemistry in Kombolcha Plain, Eastern Amhara Region, Ethiopia

The study area, Kombolcha town, forms an important industrial town situated in the Eastern Amhara region, Ethiopia. The geology of the area is mainly composed of basalts, rhyolitic ignimbrites, and Quaternary sediments. Hydrogeochemistry and the source of ions in the groundwater of the study area are poorly understood. Therefore, the current study aims to assess the factors and the different hydrochemical processes significantly controlling groundwater quality, source, and chemistry. For this purpose, a total of eighteen groundwater samples were collected using 250 ml sampling bottles at selected points in the dry season (May 2017) and wet season (November 2017). Gibbs diagram, correlation analysis, scatter plots of ionic molar ratio relations, saturation index values (estimated using PHREEQC Interactive 2.8) were used to decipher the hydrogeochemical process. Gibbs diagram shows that the rock-water interaction process is the predominant, Na/Cl and Ca/Mg molar ratio value of all groundwater samples in both seasons reveals that the groundwater chemistry of the area is controlled by silicate minerals weathering. The strong correlation of Ca with Mg in the dry season, and Ca with HCO3 and Na with HCO3 in the wet season could also be an indication of silicate weathering and ion exchange processes. The impact of anthropogenic practices on groundwater chemistry is also seen from the strong correlation of Ca with Cl, NO3 , PO4 , and F, NO2 with K, Mg, and PO4 3, PO4 3with F , and NO3 with Na, Cl, HCO3 . The negative values of chloro-alkaline indices in both seasons indicate base-exchange reaction where an indirect exchange of Ca and Mg of the water with Na and K of the host rock occurs. Saturation indices results for the wet season show that the groundwater is undersaturated with respect to calcite, aragonite, dolomite, gypsum, and anhydrite. In the dry season, however, some of the waters are oversaturated with respect to calcite and aragonite. To sum up, the groundwater quality of the study area is controlled by geological processes and anthropogenic effects.

According to Mekonnen and Yihenew (2013), due to the release of wastewater from the town to the river, boron concentration was found to be higher than the allowable irrigation limits at Borkena River and analysis results of vegetables also showed that cabbage accumulated more Fe at Kombolcha agricultural areas. Furthermore, in the leafy vegetables, high concentrations of Zn, Fe, Mn, Ni, and Pb were detected. It also was found that the chemical parameters in irrigation water had accumulated and modified the soil characteristics of the farmlands, and higher soil pH values were found in irrigated farmland soils (Eskinder et al., 2011).
Mostly surface water bodies are vulnerable to contamination and affecting the groundwater resource quality when there is an interconnection between surface water and groundwater. Hence, discharging of untreated industrial and municipal wastes to surface water increase with increasing urbanization (Hamere et al., 2017) and deteriorates and degrades the water quality. The geochemical characteristics of groundwater, hydrochemical evaluation, and the factors governing the quality of groundwater are poorly understood in the Kombolcha town, where the groundwater is the only source for drinking, industrial, and agricultural purposes. Therefore, the current study is carried out to assess the hydrogeochemical characteristics of the groundwater and to investigate the main source that controlling the groundwater chemistry around Komblcha town area using water samples collected during May 2017 (dry season ) and November 2017 (wet season ).

STUDY AREA
The study area lies between 571500 and 583800 mE longitude and 1219000 and 1234600 mN latitude in the Eastern part of Amhara region, Ethiopia (Fig 1). It is located about 375 Km north of the state capital, Addis Ababa, along the road to Mekelle and 23 Km southeast of Dessie city. It is situated on the western margin of the Main Ethiopian Rift (MER) valley and covers about 120 Km 2 area. The MER is located at the northern termination of the East African Rift System (EARS) and extends from the Afar triple junction in the north to the Turkana Rift in the south (Marco et al., 2005) (Fig 1).
The general physiographic map of the study area was prepared from DEM (Fig 1)   The flat land is covered by human settlements and agricultural areas. Small trees mostly acacia and thorny bushes are common plants in the area and eucalyptus trees are also seen in the ridges in the western part area. The dominant soils in the study area are loam, silty, silty loam, and clay loam and the pH of the soils ranges between 7.6 and 8 (Mekonnen and Yihenew, 2013

GEOLOGY AND HYDROGEOLOGY
The geology of the study area is constituted by the rocks ranging in age from Eocene-Oligocene to Recent or Quaternary deposits. Stratigraphically, from oldest to young, they are Ashangie basalts (Eocene-Oligocene), Dessie Basalt Formation and Ancharo Rhyolitic ignimbrite (Oligo-Miocene) and Recent sediments (Quaternary) (Fig 2) (GSE, 2010)).
The Ashangie basalts are also exposed as a faulted block along NW-SE trending escarpment. Its thickness decreases towards the south (< 200 m) along Dessie Kombolcha road (Mengesha et al., 1996;Tesfaye et al., 2010) and it covers about 35.5% of the study area ( Fig   2). The basalts are characterized by strong weathering, different directional tilting, columnar jointing, intense fracturing and crushing. In many of its exposure, it is dominated by inclined columnar jointed aphanitic basalts. It is also characterized by continuous and patchy outcrops with a sheet and blocky forms. The Ashangie basalt is exposed along road cuts, stream beds, gentle and steep slopes of undulating mountain chains and low lying flat plains. The unit is unconformably overlain by Dessie basalt formation. Ashangie basalt is found in N and NW, and SW of the highlands of the study area. The contact with the underlying Ashangie basalt is marked by 50 cm thick paleosoil from top and bottom as seen along Dessie-Kombolcha road.
Its thickness is about 20 m thick plagioclase phyric basalt (Mengesha et al., 1996;Tesfaye et al., 2010) and it covers about 4 % of the study area. The physical characteristic of this unit is represented by an association of different types of aphanitic and porphyritic, massive and vesicular basalts, with subordinates of pyroclasts and ash layers. Ancharo Rhyolite forms a N-S trending hill in the east part of the study area, which is named after the locality called Ancharo. It is well exposed in quarry sites and its thickness is about 50m (GSE, 2010) and covers about 15.5 % of the study area (Fig 2). The rock is white, pink, gray and medium to coarse grained and on the quarry site shows columnar joints. It contains plagioclase, alkali feldspar phenocrysts, glass shards and pumice rock fragments.
The alluvial-colluvial sediments are exposed in the low land plain and cover about 45 % of the mapped area. They are represented by black cotton and reddish brown silty to sandy soil with few outcrops of diatomite. The black cotton soil is commonly seen along the northern part of the study area. The thickness of the soil is more than 3 m as observed along the Borkena River cut and stream cuts. The sediments are characterized by clay, sand, gravel, pebbles and boulders derived mainly from the escarpments (Fig 2) and its thickness reaches up to 240 m as observed from the drilling logs of the boreholes.
Tectonic events that led to the growth of the Rift System control the geological structure of the area. The marginal grabens are slightly elongated depressions bounded on both sides by normal faults facing each other, followed in N-S direction by most of the faults in the study region (Fig 2). The eastern and western ridges bounding the valley area are characterized by a system of opposite dipping faults oriented parallel to the plateau escarpments. The N-S oriented faults show that normal dip-slip movements and form a graben like structure where the town of Kombolcha is situated including the flat low lying area around (Figs 1 and 2).
The main aquifer in the study area is the alluvial sediment. The alluvial deposits show alternating layers of sands, gravel and clay implying that there are multiple layer aquifers. The alluvial sediment is heterogeneous and their porosity and permeability making them a very good aquifer in the study area. The alluvial aquifer, as it was observed from lithological logs, is found at different depths forming a layer in most cases below 30m from the ground surface.
According to pumping test data available the existing wells in the area showed average yield 10 l/s and it has good groundwater potential where transmissivity ranges from 88m 2 /day to 335m 2 /day with average transmissivity 210 m 2 /day. The volcanic basalts and rhyolites are underlying the alluvial sediments. So far no drilling has been conducted in the volcanic section.
All the wells in and around Kombolcha tape from the unconsolidated sediment aquifers.
However, the fracture permeability is the most dominant permeability group in the highland areas. The scoriaceous basalt is highly porous due to its abundant vesicles and secondary structures such as joints and fractures interconnection of vesicles (Abraham and Assay, 2011).

MATERIALS AND METHODS
Groundwater wells drilled by government, and by private companies and hotels were used to collect groundwater samples. Systematic sampling was carried out in the dry and wet seasons, i.e., during May 2017 and December 2017, respectively. Each season, a total of eighteen (18) groundwater samples were collected.
Sample containers were washed two to three times with the sample solution and rinsed three times with distilled water. Samples were taken after the wells were pumped out for about 5 minutes to remove the stagnant water and collected in polyethylene bottles (250 ml plastic bottles). The containers were sealed and taken to the laboratory and were stored in a refrigerator without freezing to minimize volatilization and biodegradation until analysis at Mekelle University geochemical laboratory (Clesceri et al., 1998).
Water samples were analyzed in Mekelle University, School of Earth Science, geochemical laboratory for major ions and some minor ion elements. Atomic Absorption Spectrometer 5Ob Variant was used to analyze the major cations like Na + , K + , Ca 2+ and Mg 2+ . Reliability of the chemical analyses was calculated through a calculation of electrical neutrality between cations and anions, whereby the ionic balances were within ± 5% (Hem, 1989;Appelo and Postma, 1993;Li et al., 2016) and was calculated using equation 1 (Freeze and Cherry, 1979).
Correlation between different parameters of water chemistry was done by SPSS software package version 20. The saturation indices (SI) of the major mineral phases in the investigated groundwater samples were calculated using the software package PHREEQC Interactive 2.8. Diagrams like Gibbs diagram, scatter diagrams of different between major ions were used to identify the source water chemistries.

RESULTS AND DISCUSSION
The summarized chemical and physical parameters of the laboratory result of the surface water and groundwater for the dry and wet seasons of the study area are presented in appendix 1 and 2, respectively.

Hydrochemistry of Groundwater
The result and statistical description of the physical (temperature, pH, TDS and EC) and chemical parameters of groundwater are presented in appendix 1 and 2. The pH values of all waters of the study area elaborate a tendency of basic reaction among the groundwater system.
The mean concentration of calcium in groundwater samples is 75.22 and 70.67 mg/l in the dry and wet seasons, respectively and for magnesium, the average concentration in groundwater is 6.39 and 9.79 mg/l in the dry and wet seasons, respectively. Analysis of groundwater samples indicates that the mean values of potassium are 0.81 mg/l and 0.95 mg/l in the dry and wet season, respectively. According to Kolahchi and Jalali (2006), the lower potassium concentration in groundwater is owing to its greater resistance to weathering and fixation in the form of clay minerals leading to nutrient loss. The mean sodium concentration in groundwater is 31.9 mg/l in the dry season and 13.9 mg/l in the wet season. The high concentration of Na + is recorded in the dry season than in the wet season. This is because Na + will easily continue to dissolve as long as the groundwater level rises in the wet season.
The average concentration of major anions HCO3 -, SO4 2and Clin groundwater is 175.32 mg/l, 92.14 mg/l and 23.98 mg/l in dry season, respectively, while in wet season the mean values are 193.17 mg/l, 71.61 mg/l and 10.64 mg/l in wet season, respectively.

Hydrogeochemical Processes
The relationship of water composition and aquifer lithological characteristics can be established using Gibbs diagram. Three separate fields such as evaporation process, precipitation process and rock-water interaction dominance areas are shown in the Gibbs diagram (Gibbs, 1970).
Gibbs diagram (Gibbs, 1970) represents the ratio of (Na + + K + ) / (Na + + K + +Ca 2+ ) and Cl -/(Cl -+ HCO3 -) as a function of logarithm of the total dissolved solids (TDS) separately is widely used to assess functional sources of dissolved chemical constituents such as rockweathering, precipitation and evaporation. The major solute acquisition mechanisms controlling the concentration of chemical constituents in the groundwater are natural processes such as weathering, ion-exchange, and inputs from atmospheric and anthropogenic sources (Xiao et al., 2012).  The rock-water interaction dominance field shows the interaction between the rock chemistry and the chemistry of the percolated waters below the subsurface. In the study area.
Gibb's diagram (Figs 3 and 4) for groundwater of Kombolcha area displays that the hydrogeochemical processes in all seasons were significantly controlled by weathering (rock dominance) (Kumar et al., 2009). Nevertheless, evaporation and precipitations do not affect water quality.

Ionic Relations
Various hydrogeochemical mechanisms involved in the evolution of water chemistry are explained by the relationships between ionic elements using different dispersion diagrams. in groundwater (Maya and Loucks 1995;Elango et al., 2003). In the current study, all the samples of the two seasons had a ratio >2, indicates the effect of silicate minerals (Fig 5).

Na + /Clmolar ratio
The calculated value of Na + /Clratio helps to understand the source of Na + and Clin groundwater samples. At natural rock-water interaction process, the halite dissolution is accountable for sodium and the Na + /Clmolar ratio is closely one, however, if the ratio is greater 1, it is typically assumed that the sodium ion concentration in groundwater samples is released from a silicate weathering reaction (Meybeck, 1987;Prasanna et al., 2019) (Fig 6) The excess Na + in groundwater samples resulted either from silicate weathering reaction or from anthropogenic activities.

Correlation Analysis
The computed correlation coefficients are used to prove the degree of correspondence among the different hydrogeochemical parameters of groundwater in the study area. Statistical analysis was executed on the major ion concentration and physio-chemical parameters to detect the relationship and differences between the groundwater samples (Tables 1 and 2).  The strong correlation Ca 2+ and Mg 2+ in the dry season and Ca 2+ and Na + in the wet season with HCO3could be an indication of silicate weathering and ion exchange processes.
A strong correlation disclosed between Ca 2+ -SO4 2and Na + -Cland (Tables 1 and 2) designates that these ions evolve simultaneously and having probably the same origins and indicating anthropogenic input in groundwater (Srivastava and Ramanathan, 2008).
The correlation of Ca 2+ with Cl -, NO3 -, PO4 3and Fis very strong and ranges from 0.8 to 0.93, NO2with K + , Mg 2+ and PO4 3ranges from 0.74 to 0.84, PO4 3with F -(r=0.99), and NO3with Na + , Cl -, HCO3ranges from 0.89 to 0.97 and is very high (Tables 3 and 4). This could result from the poor sanitation municipal and industrial liquid wastes, and agricultural activities like fertilizer, pesticides and herbicides.

Ion Exchange
Ion exchange is accountable for the concentration of ions in groundwater. Schoeller (1967) recommended two chloro -alkaline indices (CAI) to indicate the exchange of ions between groundwater and its host rocks during traveling in the aquifer materials. These indices are: and Mg 2+ of the water with Na + and K + of the rocks and is positive CAI indicates the exchange of Na + and K + from the water with Ca 2+ and Mg 2+ of the rocks.
In both seasons, 100 % of the groundwater samples from the study area disclosed negative values representing an indirect exchange of Ca 2+ and Mg 2+ of the water with Na + and K + of the host rock (Table 3).

Saturation Indices (SI) and Chemical Equilibrium
The saturation indices describe quantitatively the deviation of water from equilibrium with respect to dissolved minerals (Jalali, 2010). The degree of water saturation with respect to a mineral is given by: Where, SIis the saturation index KIAP -is the ionic activity product, KSP -is the solubility product of the concerned mineral.
The water is at equilibrium or saturated with the mineral phase when SI is equal to zero, SI value < zero (negative value) indicates the water is under-saturated with respect to the given mineral and the mineral phase tends to dissolve, whereas over/supper-saturated and the mineral phase tends to precipitate when the SI value > zero (positive value). The calculated saturation indices for different carbonate and sulphate minerals and presented in table 3.
In the study area, all groundwater samples were under-saturated with respect to carbonate minerals (calcite, aragonite and dolomite) and sulphate minerals (anhydrite and gypsum) in both dry and wet seasons. However, in dry season, 78 % and 55 % of the samples were super-saturated with respect to calcite and aragonite, respectively. In general, the saturated index values of groundwater chemistry of the study area are greater in dry season as compared with the wet season (Table 3). This may be resulted from the less concentration major ions in the wet season due to the greater water levels in the wells during the wet season than in the dry season.

Hardness
The concentrations of certain metallic ions, predominantly magnesium and calcium in the water controls the hardness of water. Water hardness usually expressed as total hardness (TH) is given by equation 5. TH = 2.5 Ca 2+ + 4.1 Mg 2+ …………………….5 Where, TH (Total Hardness), Ca 2+ and Mg 2+ concentrations are all in mg/l (Todd, 1980 The total hardness value ranges from 81 mg/l to 479 mg/l and from 116 mg/l to 521 mg/l in May 2017 and November 2017, respectively (Appendix 1 and 2). During the dry season, classification of water based on TH by Freeze and Cherry (1979) (  (Table 4).

CONCLUSIONS
The chemical composition of groundwater of the study area is strongly influenced by rock water interaction and weathering of silicates minerals as well as ion exchange processes.
Gibbs diagram disclosed that the rock-water process is the dominant parameter in the area. Scatter plots of Ca 2+/ Mg 2+ , molar ratios of Na + /Cland (Ca 2+ + Mg 2+ )/(SO4 2-+ HCO3 -) reveals that weathering of silicates minerals as well as ion exchange processes are the controlling processes of the groundwater chemistry in the study area.

AKNOWLEDGEMENTS
This work was supported by Wollo University research project (WU/125/2017) and the author is grateful to Wollo University for its valuable financial support. The author is grateful for the critical comments and suggestions of two anonymous reviewers for improving the manuscript.