Hydrogeochemistry and Usability of Groundwater at the Tista River Basin in Northern Bangladesh

Objectives: A study on groundwater of Lalmonirhat Sadar Upazila, Bangladesh has been conducted to explore the groundwater hydrochemistry and suitability for potable water and agricultural uses. Methods/statistical analysis: During the dry season, fifty groundwater samples were collected from different shallow tube wells adjacent to Tista River, Bangladesh. Water samples were analysed according to the standard procedures. We calculated some parameters required for assessing irrigation water quality. The correlation matrix, hierarchical cluster analysis (HCA), and principal component analysis (PCA) were used for better understanding of groundwater chemistry and categorizing the water samples. Findings: The Ca–Mg–CO3–HCO3 was the dominant water type, and carbonate weathering was the main hydro geochemical process in the study area. The contributions of ion exchange and reverse ion exchange were almost equal. Groundwater characteristics were influenced by precipitation and not affected by seawater intrusion. The results of all the measured and calculated parameters clearly indicated the suitability of all ground water samples for irrigation. Based on all analysed parameters including hardness, all waters were suitable for drinking and domestic uses. Applications: The groundwater of the study area might be suitable for irrigation, drinking and domestic uses until the waters are otherwise polluted.


Introduction
For sustainable development, groundwater is a ubiquitous source of water for agricultural, domestic, and drinking purposes. At present, the uses of groundwater in the Earth for irrigation and domestic purposes are 40% and 70%, respectively. The declining quality of surface water resources has forced the people to become more dependent on groundwater. Consequently, the overexploitation and deterioration of groundwater has been observed worldwide. 1,2 In Bangladesh, groundwater is the prime source for 75% irrigation 3 and 90% drinking uses 4 and these demands could be increased day by day. However, groundwater chemistry is controlled by the geological make-up, weathering of rocks, recharge water quality, and other sources. 5,6 Water quality related to geology, climate, sources, and uses can be explained by the chemical analysis of water. 7 Especially, major ions in water can help explain the geochemical processes influencing groundwater quality. 8,9 For successful crop production, water quality is an important component for irrigation. Because the ionic concentrations in irrigation water have a direct or indirect effect on plant nutrition or soil characteristics. The concentrations of different anions and captions may increase in groundwater due to chemical fertilizers, industrial, and domestic wastes. 10 The changes in groundwater quality can be easily explained by its chemistry. Some research in other regions has investigated the irrigation water quality on crop production or Keywords: Hydrogeochemistry, Groundwater, Irrigation, Drinking, Bangladesh soil properties. 11 Studies in Bangladesh 3,4 showed that HCO 3 − , Na + , Ca 2+ , Mg 2+ , and Cl − were the prevalent ions in groundwater collected from different regions. In Bangladesh, farmers face acute shortage of irrigation water during dry season and use groundwater sources for irrigation. Besides this, water of desirable quality is necessary for drinking and domestic purposes. Thus the water quality assessment is the most significant aspect of water management irrespective of its utility.
Lalmonirhat Sadar Upazila is a densely populated area compared to the other Upazilas of Lalmonirhat district. There is no systematic study on the groundwater quality of Lalmonirhat Sadar Upazila for irrigation and drinking purposes. The investigation on groundwater quality is an important task for safety and survival of human population. In Bangladesh, arsenic contamination of groundwater at shallow aquifer is from geogenic sources 12 and irrigation with this water increases the arsenic uptake in field crops causing ultimate deposition of arsenic in the human body. But the people are using the groundwater for both the irrigation and drinking purposes. A systematic investigation of the groundwater of this Upazila is, however, necessary for assessing the geochemistry of groundwater and its suitability for various purposes. Therefore, the objectives were to characterise the physicochemical properties of groundwater, assess the processes controlling the groundwater chemistry, and evaluate suitability for irrigation, drinking, and domestic uses.

Geological Settings
In Bangladesh, the aquifer systems have not been divided stratigraphically. However, the aquifers are 1) upper shallower aquifer, 2) main aquifer, and 3) deep aquifer. 13 We collected the waters from the main aquifer. This aquifer occurs at depths from 5 m to 75 m and is the main water bearing zone. This zone is mainly semi-confined or leaky, stratified interconnected and unconfined. Medium and coarse grained sediments inter bedded with gravel is around 140 m from the ground surface.

Collection of Groundwater Samples
The groundwater samples were collected in the month of March-April 2016 from Lalmonirhat Sadar Upazila in the district of Lalmonirhat adjacent to Tista River, Bangladesh. Water levels are very low in these months. Fifty samples were collected from different shallow tube wells in the study area. Figure 1 shows the information of water sampling sites. Water samples were collected in 2 L plastic bottles. The bottles and glassware were washed with P-free detergent, and rinsed three times with ultrapure water. After soaking in 10% (v/v) HCl for 24 h, the bottles and glassware were again rinsed three times with ultra-pure water. All water samples were collected directly at the wellhead. Before sampling, plastic bottles were again rinsed 3-4 times with water to be sampled. The samples were analysed as quickly as possible on arrival at the laboratory.

Measurements of Physicochemical Parameters
Using portable meters (Thermo, Orion), we measured electrical conductivity (EC), pH, and temperature. After filtration (0.45 µm), alkalinity was measured by titration with 0.02 N H 2 SO 4 until the pH reached at 4.3. Hydrochloric acid was used to acidify the samples to maintain the pH <2 for analysis of cations and silica, and all the samples were stored at a temperature of 4 °C. Total dissolved solids (TDS) was measured by drying and Md. Jahidul Islam, AKM Mosharof Hossain, Md. Shazadur Rahman, Md. Hasel Khandoker and Mst. Nasrin Zahan weighing method. Zn 2+ , Cu 2+ , Mn 2+ , As 3+ , and Fe 3+ were determined by atomic absorption spectrophotometry. 14 Ca 2+ and Mg 2+ were measured by complexometric titration. K + and Na + were analysed by flame emission spectrophotometry. Boron (B) and silicon dioxide (SiO 2 ) were analysed using spectrophotometer (DR 2800). SO 4 2− was determined turbid metrically. CO 3 2− and HCO 3 − were analysed titrimetrically. PO 4 3− and NO 3 − were determined colorimetrically. 14 Chloride was estimated by argentometric titration. 14 Fluoride (F − ) was estimated by ion chromatography. After treating the sample with KMnO 4 and NaOH, and digestion, the chemical oxygen demand (COD Mn ) was estimated titrimetrically.

Quality Assurance and Quality Control
For any analysis, accuracy is very important for explaining the experimental result. Standard procedures were used for measuring all parameters. The accuracies of chemical analyzes were checked by means of externally supplied standards with known standards added to samples and reagent blanks, and also calibration check standards. The precision of measurements was checked taking three replicates of the sample. All reagents used in chemical analysis were from Merck Company. The differences between the sum of cations and the sum of anions and the measured or calculated TDS to EC ratio were used for detecting discrepancies. The sum of anions and cations expressed as meq/L must be balanced because all potable waters are electrically neutral. The difference between the anion and caption sums fell between acceptable limits (5−10%) and the percentage of difference is calculated on the basis of the following equation: The ratios of calculated or measured TDS to EC were from 0.55 to 0.70 which are quite acceptable for chemical analysis of water.

Evaluation of Water Quality
The equations related to the irrigation water class rating and hydrochemistry of groundwater were as follows: a) Sodium adsorption ratio (SAR) 2

Multivariate Statistical Analysis
The correlation matrix, hierarchical cluster analysis (HCA), and principal component analysis (PCA) were used for better understanding of groundwater chemistry. The Kaiser-Mayer-Olkin (KMO) measure evaluates the applicability of PCA. In this study, the KMO was 0.646. To interpret the factors of high significance, factor rotation using varimax was employed. HCA links the similar observations systematically and can easily identify different hydrochemical characteristics. For cluster analysis, Ward's method with squared Euclidean distance was applied and is the best option for dendrogram preparation. 15 The relations among different water parameters were detected by correlation studies. All the statistical calculations were carried out by SPSS 22.

Quality Characteristics of Groundwater
The summary results of groundwater of our study are shown in Table 1. The water temperature ranged from 22.5 to −27.1 °C during the study period. The pH of water varied from 6.4 to 8.0 and indicated that the waters were slightly acidic to slightly alkaline ( Table 1). The COD of the study water ranged from 1.2 to 3.9 mgO 2 /L. The alkalinity and hardness values varied from 12 to 81 mg/L  Table 1). The concentration of K was comparatively low than that of Na in all waters (Table 1), because K minerals have low ability in migration 16 and are resistant to decomposition. 17 However, Na + ranged from 5 to 20 mg/L with a mean of 9.60 mg/L. Na + usually comes from the dissolution of evaporites and silicates. 8 Ca 2+ and Mg 2+ are the results of carbonate dissolution in general and varied from 5.61 to 37.68 and 5.83 to 32.08 mg/L, respectively. The concentrations of As 3+ , Fe 3+ ,Zn 2+ , Mn 2+ and Cu 2+ ranged from 0.011 to 0.042 mg/L, 0.064 to 1.190 mg/L, 0.025 to 0.067 mg/L, 0.030 to 0.710 mg/L, and 0.011 to 0.148 mg/L, respectively.
The concentrations of CO 3 2− were negligible compared with HCO 3 − . The sources of CO 3 2− and HCO 3 − are the dissolution of carbonate rocks resulting in eventual precipitation of CO 2 . 16 Sedimentary rocks and chloride salts are the sources of Cl − . 17 The dominant natural sources of SO 4 2− include rock weathering and biochemical processes. 18 Weathering of muscovite, biotite, fluorite, and fluoroapatite is the main cause of fluoride in groundwater. However, the concentrations of HCO 3 − , Cl − , SO 4 2−, NO 3 − , PO 4 3− , F − , B, and SiO 2 ranged from 49 to 207 mg/L, 4 to 53 mg/L, 0.18 to 11.72 mg/L, 1.25 to 7.21 mg/L, 0.003 to 0.019 mg/L, 0.13 to 0.52 mg/L, 0.028 to 0.096 mg/L, and 5 to 33 mg/L, respectively. Fertilizer application can alter the major ion concentrations in groundwater in the study area.

Mechanisms Controlling Groundwater Chemistry
Relations among the different anions and cations are able to explain the mechanisms that control groundwater chemistry. It can be explained by the following headings.

Carbonate Weathering
In carbonate weathering, the molar ratio of Ca 2+ + Mg 2+ : HCO 3 − (Figure 2Af) had greater than unity suggesting the dominancy of carbonate weathering and the source of high HCO 3 − . The possible causes are the dissolutions of gypsum, anorthite, and calcium montmorillonite and cation exchange. On the other hand, the formation of CaCO 3 can decrease Ca 2+ concentration with a proportional increase of Na + . Consequently, the Mg 2+ :Ca 2+ ratio were greater than 1 (Figure 2Ae). With elevated Na + ions, Mg 2+ is dominant over Ca 2+ in the increased clay-rich soil. 19 Dominance of Na + + Mg 2+ over Na + + Ca 2+ in groundwater can have a greater Na + hazard. 19 The HCO 3 − :Na + > 1 indicates carbonate weathering while its lower value suggests silicate weathering. 20 In this study, the ratios of HCO 3 − + CO 3 2− : Ca 2+ (Figure 2Bi) and HCO 3 − + CO 3 2− : Mg 2+ (Figure 2Bj) were close to the unity reflecting the dominancy of Ca and Mg-containing minerals. However, the influences of Na-containing minerals are less because the ratio of HCO 3 − + CO 3 2− : Na + (Figure 2Ad) was far below the unity. In most cases, the ratios of Ca 2+ + Mg 2+ : total cations were close to unity (Figure 2Ab). The ratios of Na + + K + : total cations indicate lower concentrations of these two cations over Ca 2+ and Mg 2+ (Figure 2Aa

Silicate Weathering
The ratio of Na + : total cations (TC) indicates the levels of silicate weathering process. 22 Groundwater samples had Na + : TC< 0.25 (Figure 2Bk) indicating that there was less silicate weathering. The ratio of Na + + K + : TC was 0.25 ( Figure 2Aa). This implies that silicate weathering is less responsible to the supply of cations in groundwater. When carbonate and silicate minerals are the major contributors for groundwater chemistry, the HCO 3 − : TC value would be one. 23 This study did not correspond well to the abovementioned mechanism (Figure 2Bh). The plotting of Ca 2+ +Mg 2+ versus HCO 3 − (Figure 2Af) further infers that the groundwater did not have an excess of HCO 3 − . This HCO 3 − was not balanced by Na + (Figure 2Ad), as the silicate weathering was not a prime mechanism to release the Na + and HCO 3 − into the groundwater. The ratio of Na + :Cl − in most samples was lower than unity ( Figure  2Ac). The dissolution of NaCl generates a 1:1 ratio of Na + :Cl − while the release of Na + from silicate weathering produces a wide ratio. 24

Saline Water Intrusion
Salinity and saline water intrusions are related to Na-Cl relationship. 25 The equations of Cl − /Σ anions > 0.8 and Na + /(Na + + Cl − ) < 0.5 suggest seawater intrusion into groundwater. 26 In this study, the average value of Cl − /Σ anions ratio was 0.26 and also Na + /(Na + + Cl − ) ratio was 0.42 (Table 1). A significant correlation (r = 0.56) exists between Na + and Cl − suggesting that they might be originating from the same sources (Table 3). The cation exchange process may also increase Na + concentration in addition to the dissolution of halite. 27 Figure 2Ac shows the deviations of the expected Na + :Cl − (1:1) relation indicating that a fraction of Na is associated with another anion. In the study area, the ratio of Na + /Cl − < 1 meaning another source is contributing chloride to the groundwater.

Ion Exchange
In aquifer, the clay minerals in unconsolidated deposits adsorb cations in their pore spaces. Ion exchange with clay materials is responsible for ionic concentration in groundwater. 28 In this study, the majority of the samples showed an excess of Mg 2+ over Ca 2+ (Figure 2Ae), which could be the dissolution of Mg-containing minerals in addition to dolomite. The exchange of Na + in water by Ca 2+ or Mg 2+ in clay material can cause an excess of Ca 2+ or Mg 2+ in groundwater and vice versa. 29 An excess of HCO 3 − + SO 4 2− over Ca 2+ + Mg 2+ indicates ion exchange process, while an excess of Ca 2+ + Mg 2+ over HCO 3 -+ SO 4 2supports reverse ion exchange. 30 In this study, the ratio of (Ca 2+ + Mg 2+ )/(HCO 3 − + SO 4 2− ) > 1 indicates the reverse ion exchange process (Figure 2Bg). If the dissolutions of calcite, dolomite, and gypsum are dominant in water, the relationship between Ca 2+ +Mg 2+ and SO 4 2− + HCO 3 − would be close to 1:1. 28 The role of ion exchange in groundwater can be evaluated by plotting (Ca 2+ + Mg 2+ ) -(HCO 3 − + SO 4 2− ) against (Na + + K + ) -Cl − . The (Ca 2+ + Mg 2+ )-(HCO 3 − + SO 4 2− ) relation implies that excess Ca and Mg were originated by the influences of other processes, whereas the (Na + + K + )-Cl − indicates that K and Na were from the sources other than their respective chlorides. In case of the dominancy of ion exchange process, the graphical representations of the samples will form a line with a slope of -1. 31 Figure  2Bl shows that samples had a slope of -1.326, indicating cation exchange influences the groundwater chemistry in the study area.

Chloroalkaline Index
Two chloroalkaline indices (CAI-I and CAI-II) suggested by Schoeller 32 were used to identify the specific ion exchange between the groundwater and its surroundings. 33 In general, Na + and K + in groundwater are exchanged with Ca 2+ and Mg 2+ in aquifer matrix. An exchange between Na + or K + in water with Mg 2+ or Ca 2+ in rock indicates a direct ion exchange. Na + or K + will decrease in water and both the above indices will be positive. Conversely, the reverse ion exchange will produce negative indices. 33 The negative and positive values for both indices suggest the contribution of direct ion exchange and reverse ion exchange in the system (Figure 3). Results show that 68% is negative and 32% is positive in CAI-I while in CAI-II, 22% is negative and 78% is positive. On an average, the contribution of direct ion exchange and reverse ion exchange reaction is almost equal. Gibbs 34 proposed two diagrams known as Gibbs diagrams that explain the major mechanisms governing water chemistry. These diagrams are widely used in groundwater studies 8,33 in addition to surface water. In our study, the water samples fell in the precipitation dominance area of Gibbs diagram (Figure 4). Due to high evapotranspiration and restricted fresh water exchange in the study area, salt layers may form in the evaporating surface. 35 On the other hand, human activities can also change the groundwater chemical composition.

Piper Diagram
Piper diagram 36 can easily classify and compare the water types. The concentrations of majoranions and cations build this diagram. Groundwater samples are classified as Ca 2+ -Mg 2+ -CO 3 2− -HCO 3 − type ( Figure 5). This signifies that the dissolution of carbonates is a predominant process releasing substantial amount of Ca 2+ and Mg 2+ into the groundwater of the study area. Piper's diagram confirms that all the groundwater's are characterised as alkaline earth's (Ca 2+ + Mg 2+ ) exceeds alkalies (Na + + K + ) and are described as weak acids (CO 3 2− + HCO 3 − ) exceed strong acids (SO 4 2− + Cl − + F − ).

Multivariate Statistical Analysis
In this study, a Pearson correlation matrix was used to determine the relationships between variables ( Table  2). There were few significant relationships among the measured concentrations with depth. The high correlations between Cl − and HCO 3 − (r = 0.55), Cl − and Na + (r = 0.56), K + and HCO 3 − (r = 0.54), Na + and K + (r = 0.76), and Na + and HCO 3 − (r = 0.73) indicate that they most likely derive from the same sources of water ( Table  2). The relation between Ca 2+ -HCO 3 − (r = 0.87) and Ca 2+ -Na + (r = 0.66) might represent the contributions of carbonate and silicate weathering.
PCA was applied to the 24 physicochemical parameters, but only eight PCs with eigenvalue greater than 1 are considered to be the most important ( Table  3). The high eigenvalue indicates the greater contribution to the variability of the groundwater chemistry. The PC loadings were categorised into strong, moderate, and weak corresponding to the absolute loading values of >0.75, 0.75−0.50, and 0.50−0.30, respectively. 37 A total of eight factors were accredited responsible for groundwater chemistry in the study area accounting 76.80% of the total variance (Table 3). Factor-1 has the highest loading (33.33%) and the contributors include Ca 2+ , HCO 3 -, Cl -, Mg 2+ , Na + , K, EC, TDS, hardness, and alkalinity. These major ions might be due to bedrock dissolution arising from rock-water interaction.
Examination of Figure 6 led to the identification of three major clusters in the dendrogram labeled as cluster 1, cluster 2, and cluster 3. The majority of samples (28) were grouped in cluster 1. Their similarity could support the hypothesis of a common evolutional trend to the study area. The cluster 2 and cluster 3 include 12 samples and 10 samples, respectively. The linkage distance is highest in cluster 3 followed by cluster 1 and cluster 2.

Irrigation Suitability
For agricultural use, the suitable pH ranges from 6.0 to 8.5 38 and all waters were suitable. The permeability index (PI) ranging from 48 to 116% locates under Class I and Class II (Table 1). These two classes allow 75% or more permeability of irrigation waters. Groundwater with Kelly's ratio greater than one is not good for irrigation. In this study, Kelly's ratio ranged from 0.11 to 0.43 indicating the suitability of all the samples for irrigation (Table 1).
All the waters were rated as 'fresh' (TDS < 1000 mg/L) suggested by Freeze and Cherry. 41 Based on major cations, all the study waters were below the permissible limits (Ca 2+ = 80 mg/L, Mg 2+ = 35 mg/L, Na + = 200 mg/L, K + = 30 mg/L). 42 The chloride concentration should be below 140 mg/L in irrigation water, 38 and all waters were suitable for irrigation. The groundwater NO 3 --N did not exceed the maximum concentration (30 mg/L) as because most crops are relatively unaffected below this value. 38 The lower concentrations of Cu 2+ , Zn 2+ , Fe 3+ , As 3+ , and Mn 2+ might be suitable for crop production and soil environment.

Drinking and Public Health Suitability
In this study, the pH (6.4−8.0) of all groundwater samples was safe (safe limit 6.5−8.5) for drinking water prescribed by WHO. 43 The recommended concentration of TDS is 500 mg/L, 43 and all waters were suitable for drinking. The Na + concentration should be below 200 mg/L. 43 People can feel  4 2− , and NO 3 − of the studied groundwater samples were under the permissible limits. WHO 43 suggests the maximum permissible limit of Cl − , SO 4 2− , and NO 3 − in drinking water should be 250 mg/L, 150 mg/L, and 10 mg/L, respectively. The content of As is above the WHO guideline for drinking water (<0.01 mg/L) 43 but below the Bangladesh limit (<0.05 mg/L). Concentration of Fe of almost all the samples exceeded the WHO permissible limit (<0.1 mg/L). The mean value of Mn (0.298 mg/L) was above the WHO guideline for drinking water (<0.05 mg/L). Exposure to high levels of manganese can cause a disease manganism characterised by a Parkinson-like syndrome showing weakness, muscle pain, apathy, and slow clumsy movement of limbs. Hard water may cause high incidence of urolithiosis, 43 anencephaly, parental mortality, and cardio-vascular disorders. This water also develops scales in distribution pipes and cooking utensils, and requires more soap for washing clothes. 35

Conclusion
Groundwater quality of Lalmonirhat Sadar Upazila was measured for drinking and agricultural purposes. The findings of this study are as follows: • The major cations in the study area were in the order of Mg 2+ > Ca 2+ > Na + > K + . The anions were also arranged in decreasing order as HCO 3 − > Cl − > NO 3 − > SO 4 2− . • The dominant water type was Ca 2+ -Mg 2+ -CO 3 2− -HCO 3 − and the study area is a precipitation dominance area.
• Based on Wilcox requirement and SAR, the groundwater might be suitable for irrigation without salinity hazard.
• Carbonate weathering was the main component for controlling groundwater chemistry followed by silicate weathering and ion exchange processes.  Anthropogenic activities might cause the excess concentrations of some anions and cations. • Accumulated salts on soil surfaces generated by irrigation can leach by recharge water and reach the water table, and could become a cause of groundwater pollution. And recycling of groundwater for irrigation can increase soil and groundwater salinity.
• According to the measured parameters such as pH, TDS, Na + , Cl − , SO 4 2− , NO 3 − , and semi-heavy metals (Cu, Zn, Fe, Mn, and As), all waters were suitable for drinking and domestic purposes.
Based on observed patterns, it can be concluded that all the groundwater samples of the Upazila, Bangladesh were suitable for irrigation, drinking, and domestic uses. However, continuous monitoring of groundwater in the study area would be useful for predicting the suitability for irrigation and drinking purposes in the upcoming days.