The study area is a part of Rupsha-Kazibacha River system of the Shailmari River under Khulna districtis confined between the longitude 89˚31'18.6" and 89˚28'31.9" East and the latitude 22˚44'45.1" and 22˚46'23.8" North (Figure 1) of

which 607.80 km² is riverine and 2028.22 km² is under forest (Roy et al., 2005) . The present study was conducted particularly on three riverine unions i.e. Jalma on the left bank, Batiaghata and Dumuria on the right bank of the river in the context of its hydro-geochemical characterization and suitability evaluation. The hydro chemical characteristics of this river system are mostly controlled by unique climatic variation, local geologic and anthropogenic structure. It is highly attributed by widespread shrimp farming, agro farming, livestock, poultry industries, brick kiln and other industries (Islam et al., 2016) .

The study area is characterized by the Quaternary alluvial sequence marking the Halocene to Recent Alluvium deposits (Alam, 1990) . Hydrogeological properties of this area are therefore, largely governed by the litho-stratigraphy and prevailing tectonic features of this region (ADB, 2011) . The present study involves the deep hydrogeology of Batiaghata, which consists of a confined and deep Pleistocene aquifer where deposition of sediments helped the formation of aquifers with high transmissivity and storage in this region (Rahman et al., 2011) . The hydrogeology of the Khulna region has generally been marked as a three aquifer systems separated by aquitards composed of clay minerals (Figure 2) (ADB, 2011; DPHE-JICA, 2015) . The upper or first aquifer is semi-confined, shallow Holocene aquifer type and occurs at a depth of 50 to 100 m from the surface which consists of gray fine to very fine sand and overlain by 5 - 45 m thick silty clay aquitard (ADB, 2011) . The second aquifer encounters at depth around 150 to 270 m below the surface and extends up to 350 m in places which is confined to leaky confined in nature and serves as the main aquifer in most parts of the

study area. The aquifer is dominated by gray very fine to medium sand, underlain by gray clay aquitard and is separated from the upper shallow aquifer by a thick clay aquitard. The third aquifer is the deepest aquifer and occurs below 350 m from the ground surface (ADB, 2011; DPHE-JICA, 2015) . The study area is also governed by sub-tropical and humid climatic pattern which experiences hot summer in pre-monsoon from March to May, while receives 80% of total annual rainfall during monsoon from June to September. The pre-monsoon and winter is fed by lesser precipitation with coastal hazards and livelihood extremes (FAO, 2011) .

A well suited reconnaissance survey was performed prior to select the sampling station to optimize the consumption pattern of the people. The number tubewells were selected through the purposive (Non -probability) sampling technique from both bank of the Shailmari river basin. Water samples were collected for four successive seasons from each sampling locations following the guidelines of standard methods APHA, 1999 throughout the study period. For cationic and anionic analyses samples were collected in 500 ml PET bottles and preserved by adding HCl (to pH ~ 2) for cationic samples. Then samples were carried to the laboratory and conserved at 4˚C prior to laboratory analysis (APHA, 1999; Huang et al., 2014) . The number of samples in total 80 regarding four consecutive seasons namely, Pre-monsoon (May, 2014), Monsoon (August, 2014), Post-monsoon (October, 2014) and winter season (March, 2015) [10 from left bank and 10 from right bank for each season] (Figure 1). The sample station ID, owner of the tubewells, depth and number of households depend on those tubewells are listed in Table 1.

Note: Here LB₁ to LB₁₀ and RB₁ to RB₁₀ represent the left bank and right bank sampling station respectively and HH represent House Hold number depends on these tubewells.

In-situ and laboratory measurement were conducted following the rules prescribed in American Public Health Association and Standard Methods (APHA, 1999; Huang et al., 2014) . Some field parameters such as pH, DO, Temperature and EC were conducted instantly at the sampling site by a portable pH, DO meter(HACH sensION156 portable) and EC/TDS meter (HANNA H1-9635) after calibration (APHA, 1999) . In the laboratory, sodium (Na⁺) and potassium (K⁺) were measured using flame photometer (Model No. PEP 7) whereas titration method is used for measuring calcium (Ca²⁺), magnesium(Mg²⁺) and chloride (Cl⁻) (Ramesh & Anbu, 1996) . Bicarbonate ( HCO 3 − ) concentrations of the groundwater had measured by potentiometric titration method with H₂SO₄. Salinity and TDS were estimated in the laboratory by digital multi-meter (Model No. Bantee900p). Sulphate ( SO 4 2 − ), ortho-phosphate ( PO 4 3 − ), and nitrate ( NO 3 2 − ) of the water samples were analyzed using UV–visible spectrophotometer (Model No.4802 uv/visbouble beam Spectophotometer) (APHA, 1999) . Total hardness (TH) of the groundwater was calculated using the formula given by Sawyer et al. (2003) . For microbiological analysis, 500 ml water samples were aseptically collected in sterile Nalgen plastic bottles. All samples were placed in an insulated box filled with ice packs (Johnny Plastic Ice; Pelton Shepherd, Stockton, CA, USA) immediately after collection and transported to the Environmental Microbiology Laboratory of Environmental Science Discipline at Khulna University. We counted the concentration of Total Coliform (TC) and Faecal Coliform (FC) using the membrane filtration technique. Membrane filters, through which water sample was passed, were aseptically placed onto m-FC and m-Endo agar base plates followed by Hi-media procedures (APHA, 1999) . The m-Endo plates were incubated at 35˚C ± 0.2˚C for 24 h for enumeration of TC whereas for FC estimation, incubation was done at 44˚C - 45˚C. After incubation, characteristic pink colored colonies were counted as TC and blue colored colonies were counted as FC, expressed as colony forming units (cfu) per 100 ml.

Water quality monitoring is a vital part of any environmental monitoring program owing to reporting the results in both from societal and environmental point of view. The Canadian Council of Ministers of the Environment Water Quality Index (CCME WQI) produces a number between 0 (worst water quality) and 100 (best water quality) while numbers are divided into 5 descriptive categories: Excellent (CCME WQI Value 95 - 100), Good (CCME WQI Value 80 - 94), Fair (CCME WQI Value 65 - 79), Marginal (CCME WQI Value 45 - 64) and Poor (CCME WQI Value 0 - 44) in a mathematical framework (CCME, 2001) . For assessing ambient water quality conditions in respective water quality objectives, water samples were ranked to simplify the degree of vulnerability. Three vital elements are incorporated in this index: scope (F₁)—the number of variables not meeting water quality objectives; frequency (F₂)—the number of times these objectives are not met; and amplitude (F₃)—the amount by which the objectives are not met was calculated using the formula demonstrated in Table 2 (CCME, 2001; Rocchini & Swain, 1995) . To evaluate the suitability of water samples for drinking and domestic purposes were compared with both WHO (World Health Organization) and Bangladesh Drinking Water Quality standards corresponding the determined values of the physico-chemical parameters of the water samples (Ahmed & Rahman, 2000; WHO, 2011) .

To achieve the analytical insights and accuracy during the study period, the charge balance between cations and anions were found within ±10% by repetitive analysis after calculating the ionic balance errors (Vyas & Jethoo, 2015; Hounslow, 1995) . The analytical results exhibited inclusive hydrochemical variation among the seasons throughout the study period. Figure 3 shows the spatial and temporal variation, while Table 3 shows mean, the standard deviation, and the coefficient of variation between the left and right banks as well as within the seasons. The pH values range from 6.45 to 8.08 with a mean value 7.33 for left bank and range from 6.36 to 7.92 for right bank of Shailmari River. The lowest value is observed in winter and the highest value in monsoon season.

The observed average concentrations of Na⁺ and Cl⁻ were 225.76 ± 39.44 and 284.90 ± 68.00 for left bank and 218.36 ± 54.85 and 256.79 ± 57.58 for right bank respectively. The highest concentration of Na⁺ was found 298.35 mg/l in winter and for Cl⁻ 406.30 mg/l in pre-monsoon season that exhibits the increased salt concentration in the study area. The EC ranges from 1681 μS/cm (monsoon) to 2515 (winter) μS/cm with an average value of 2091.60 μS/cm for right bank. In case of left bank the average concentration of EC was 2121.30 ± 216.40 μS/cm while 2091.60 ± 240.80 μS/cm for right bank. TDS values ranges from 802 mg/l with average concentration within the range of 1106.60 ± 152.70 (Right bank) to 1598 mg/l (Left bank) with an average value of 1146.60 ± 149.90 mg/l. The lowest and highest concentration was found in monsoon and pre-monsoon season respectively. Among the anions, concentration of HCO 3 − was observed higher

Note: EC = Electrical conductivity, TDS = Total dissolved solids (TDS), DO = Dissolved oxygen, TH = Total hardness, FC = Faecal coliform, TC = Total coliform. *Maximum allowable limit for drinking in Bangladesh (BD) is used here for TH (200 - 500 mg/L), Mg²⁺ (30 - 35 mg/L) and Cl⁻ (150 - 600 mg/L).

in post-monsoon 590 mg/l with the variation of 482.40 ± 87.30 mg/l for left bank. For right bank the higher concentration was 594 mg/l ranges with mean variation 482.07 ± 54.30 mg/l. Variation of HCO 3 − could be recognized to the bicarbonate weathering coupled with wastewater mixing into the ground water during the sampling seasons (Chitradevi & Sridhar, 2011) . Ca²⁺ and Mg²⁺ were the second and third in cationic abundance mentioned in Table 3. During the study EC, TDS , Na⁺ and Cl⁻ were observed higher in pre-monsoon indicating potential salinization. This phenomenon might be attributed by reduced upstream freshwater inflow, geochemical interactions and presence of paleo brackish water in the river pathways while dense precipitation headed to dilute the concentrations in monsoon and consecutive post-monsoon seasons (Mondal et al., 2013) . Higher load of sulfate and nitrate in dry period might be coming from weathering, erosional deposits, leaching or leakage to adjacent sewage disposal to the ground water.

Hardness is one of the significant parameters for determining the type and nature of water classified in terms of degree of hardness as 1) soft (<75 mg/l), 2)

moderately hard (75 to 150 mg/l), 3) hard (150 to 300 mg/l) and 4) very hard (>300 mg/l) (Ahmed & Rahman, 2000) . According to aforementioned classification, 70% water samples of left bank and 40% water samples of right bank are fall under very hard category in respective pre-monsoon and winter season. The presence of high content of magnesium, calcium and associated ions accompanied by probable leaching of sewage disposal and rock mineral geochemical interaction, enhance the hardness of ground water (Chidambaram et al., 2009; Mondal et al., 2013) . Apart from water chemistry analysis the microbial analysis revealed the highest Faecal Coliform (FC) number is 17 MPN/100 ml and 16 MPN/100 ml in respective left and right bank of the river in the monsoon season. The highest numbers of Total Coliforms (TC) were 2200 MPN/100 ml also in monsoon for left bank and 2500 MPN/100 ml for right bank in post-monsoon season. In monsoon and post-monsoon 2, 5, 7 and 9 no. sampling spot faces significant microbial contamination. This contamination might be caused by peat latrines, sewages connections or entering contaminants into the wells through hydraulic connection which possesses diarrhea, dysentery, stomach problems and even severe health threats to the community people.

The correlation matrix was conducted by putting the mean value of water samples for the four consecutive seasons incorporated in Table 4. The computed analysis shows strong positive correlation of Mg²⁺ with EC, TDS and Ca²⁺ but negatively correlated with pH, EC and HCO 3 − relation indicate this ground water aquifer might be dominated by rock-weathering and geochemical interaction (Chidambaram et al., 2009; Rahman et al., 2011) . The strong positive relation between Ca²⁺ and Mg²⁺ may be due to calcite or dolomite weathering in the study area (Bahar & Reza, 2010; Rahman et al., 2011; Islam et al., 2016) . There are also more significant correlation pairs was noticed between Ca-Mg, Ca-Na, Ca- HCO 3 − and Mg- HCO 3 − where their association with TDS insight these ions probably derived from the same source of saline water as found in surrounding river water (Bahar & Reza, 2010; Islam et al., 2016) . As both NO 3 − and PO 4 3 −

Note: ** Correlation is significant at the 0.01 level; *Correlation is significant at the 0.05 level.

are derived from anthropogenic sources the strong association between this two indicate the anthropogenic influence on groundwater via leaching or hand pumps tubewells in this locality (Chidambaram et al., 2009) .

The correlation matrices of groundwater samples for pre-monsoon and winter also shows that pH have strong negative correlation with EC but strong positive correlation TDS whereas Ca²⁺ have strong positive correlation with EC and TDS. There is a perfect positive correlation of HCO 3 − also found with EC, TDS, Ca²⁺, Mg²⁺, Na and Cl⁻ where Na and Cl⁻ are moderately related with salinity in pre-monsoon but highly correlated in winter season indicating such ions are mostly derived from sea water. On the contrary, PO 4 3 − showed strong negative correlation with EC, Ca²⁺, Mg²⁺, HCO 3 − and SO 4 2 − whereas Cl⁻ shows positive correlation with EC, TDS, Ca²⁺, Mg²⁺, Na⁺ and K⁺ for both pre-monsoon and winter season. Ca²⁺ and Mg²⁺ are moderately correlated with HCO 3 − in pre-monsoon and winter season. On the contrary, PO 4 3 − showed no correlation as we found in the left bank in both season but Cl⁻ shows almost similar positive correlation with EC, TDS, Ca²⁺, Mg²⁺, Na⁺ and Mg²⁺ except K⁺ and pH for the right bank water samples. For monsoon and post-monsoon it is found that EC and TDS show a strong negative correlation with pH at the 1% level of significance, whereas EC and TDS shows a highly positive significant correlation in case of both left bank and right bank respectively. There is a strong positive correlation between Ca²⁺ and Mg²⁺ in case of left bank but no observed correlation in case of right bank for the monsoon season. In both season Na⁺ and Cl⁻ shows moderate positive correlation in left bank but perfectly positive correlation was found in case of right bank at the 1% level, indicating such ions are mostly derived from sea water. The computed analysis also indicated that EC and TDS shows strong positive correlation with Ca²⁺, NO 3 − , PO 4 3 − and HCO 3 − in case of left bank whereas in case of right bank it shows moderate positive correlation with Ca²⁺, HCO 3 − , Mg²⁺ and K⁺ but no relation with NO 3 − and PO 4 3 − . The strong correlation among EC, TDS, Na⁺ and Cl⁻ in pre-monsoon and winter is regarded as the salinization index which may indicate the sea water intrusion in this region while moderate correlation in monsoon season reflects the freshening of the water samples compared with pre-monsoon and winter season (Bahar & Reza, 2010) .

To reveal the cation and anion concentration within defines composition categories, the samples were plotted on trilinear diagram for indicating their types (Figure 4). It is a combination of distinct zones of hydrochemical facies that lie on common baseline (Hounslow, 1995) . The trilinear diagram of this study is classified into six hydrochemical facies: 1) Ca-Mg-HCO₃ type; 2) Na-Cl type; 3) Mixed Na-HCO₃-Cl type; 4) Mixed Ca-Mg-Cl type; 5) Ca-Mg-SO₄ type and 6) Na-HCO₃ type (Pipper, 1944) . Figure 4(a) shows the majority of the water samples were in type 2 (Na-Cl type), followed by the rest of the samples were in type 3 (Mixed Na-HCO₃-Cl type) in pre-monsoon and winter season. Based on the dominance of different cations and anions the order were Na⁺ > Cl⁻ > HCO 3 − >

Ca²⁺ > Mg²⁺ > K⁺ > SO 4 2 − > NO 3 − in pre-monsoon and winter season. This type of chemical composition of ground water samples were mostly controlled by sea water and tidal channels or structural interferences to the natural water flow via ground water layer which consist of recent alluvium deposit. This specifies that ground water becomes more Na-Cl types in these two seasons denoting the influence of sea-water mixing (Bahar & Reza, 2010) . In monsoon and post monsoon the trilinear diagram depicted that the majority of the water samples were intype 3 (Mixed Na-HCO₃-Cl type), while rest of the samples were in type 1 (Ca-Mg-HCO₃ type) and type 2 (Na-Cl type) depicted in Figure 4(b). The piper diagram advocates that majority of the water samples are dominated by alkali carbonate which has altered the concentration of sodium and chloride into fresh water by ion exchange method with exchange to Ca-Mg-HCO₃ (Pipper, 1944) . The overall result revealed the intrusion of saline water into coastal aquifers with significant percentages of sodium and chloride.

The analytical results of pH, EC, TDS, Salinity, DO, Total Hardness and major ions of ground water were put in CCME water quality index to understand the overall scenario and ongoing condition of the water (CCME, 2001; Rocchini & Swain, 1995) . Sixteen water quality parameters for four consecutive seasons were used to find out the CCME WQI value considering:

F₁) scope—the number of variables not meeting water quality objectives;

F₂) frequency—the number of times these objectives are not met and

F₃) amplitude—the amount by which the objectives are not met

The graph (Figure 5) and table (Table 5) in combination depicts that nearly three quarter of total water samples are felled under marginal category which means water quality is frequently threatened or impaired; conditions often depart from natural or desirable levels. Tiny fraction of water samples of left bank and almost a quarter portion of water samples of right bank were under fair category demonstrating that water quality is usually protected but occasionally threatened or impaired. From the calculated value it is evident that the water

Note: LB and RB represent the left bank and right of the river respectively.

quality of major tubewells often exceeds usual or allowable limits which elicit the ground water vulnerability and public health concern (CCME, 2001) . Salinity intrusion into the shallow aquifers, upstream reduction of tidal channel’s flow, excessive withdrawal of ground water for irrigation, sewage connection or leaching might be the prime cause of ground water vulnerability (CCME, 2001; Chidambaram et al., 2009) .

The water quality parameters were compared with the standard guideline values recommended by the both Bangladesh and WHO standards for drinking purposes (Rocchini & Swain, 1995; CCME, 2001) . Table 6 shows the comparison of analytical results with the prescribed guidelines and their percentages exceeding the recommended limits. The computed result indicates that 100% samples of EC surpassed the allowable limit recommended by WHO standard. Probably ground water inflows will contribute to the high electrical conductivity depending

Note: LB = Left Bank, RB = Right Bank, EC = Electrical conductivity, TDS = Total dissolved solids, DO = Dissolved oxygen, TH = Total hardness, FC = Faecal coliform, TC = Total coliform. ^a (Ahmed & Rahman, 2000) ; ^b (WHO, 2011) .

on the geology, rock erosion and leakage to sewage disposal or heavily ionized dissolved minerals that have been evaporated to dryness (Rahman et al., 2011) . Sodium and chloride are found in ground water naturally. However, increasing the concentration in pre-monsoon and winter season might be caused by underground salt deposition, pollution from septic and salt water intrusion due to proximity of tidal channels in coastal belt (EPA, 2001; Hasan et al., 2013) . TDS exceeds the permissible limit of both Bangladesh and WHO standard followed by the 100% for pre-monsoon. In respective monsoon, post-monsoon and winter season 70%, 90% and 80% water samples for left bank as well as 20%, 70% and 90% water samples for right bank surpassing the desirable limit. High values of TDS may be caused by stagnation of water in most of this tubewells due to the weight of residue left when a water sample has been evaporated to dryness and hydraulic connection with the river (Bahar & Reza, 2010) . This result indicates the degree of unsuitability for different consumption scheme (Chitradevi & Sridhar, 2011) . In view of HCO 3 − it was found that 100% water samples of all seasons are beyond the limit of WHO standard for the probability of carbonate weathering coupled with hydraulic connection into the aquifer. High concentration of bicarbonate also might be caused by dissolution of organic matter leading to an increase residual sodium carbonate with Na⁺ content in the soil (Zahid et al., 2008; Bahar & Reza, 2010) .

Faecal contamination in drinking water is a serious health threat due to the potential for contracting diseases from pathogenic organisms. From the microbial analysis of Faecal Coliform and Total Colifom it was found that 80% (pre-monsoon), 70% (monsoon), 70% (post-monsoon) and 30% (winter)water samples for left bank exceeded the limit of both Bangladesh and WHO standard. The right bank of the riverine union also showed almost similar pattern in all the successive seasons. Agricultural runoff, effluent from septic systems of sewage discharges, infiltration of domestic or wild animal fecal matter, poor well operation and construction (particularly shallow dug wells) are might be the prime source of microbial contamination in ground water (Ashbolt, 2015) . From overall comparison scheme of this study it has been demonstrated that majority of the water samples do not satisfy the limit of both Bangladesh and WHO standards signifying their degree of unsuitability for drinking purposes.