Section 3: Arsenic biogeochemistry in groundwater Copyright © 2005 Taylor & Francis Group plc, London, UK 133 Natural Arsenic in Groundwater: Occurrence, Remediation and Management – Bundschuh, Bhattacharya and Chandrasekharam (eds) © 2005, Taylor & Francis Group, London, ISBN 04 1536 700 X Natural enrichment of arsenic in groundwaters of Brahmanbaria district, Bangladesh: geochemistry, speciation modeling and multivariate statistics Ondra Sracek Institute of Geological Sciences, Faculty of Science, Masaryk University, Brno, Czech Republic Prosun Bhattacharya, Mattias von Brömssen, Gunnar Jacks Groundwater Arsenic Research Group, Department of Land and Water Resources Engineering, Royal Institute of Technology (KTH), Stockholm, Sweden Kazi Matin Ahmed Department of Geology, University of Dhaka, Dhaka, Bangladesh ABSTRACT: Groundwater with geogenic arsenic enrichment is commonly encountered in the Holocene sedimentary aquifers of the Bengal Delta Plain (BDP). The present study was carried out in Brahmanbaria district, covering an area of 18 km 2 in northeastern Bangladesh. The Chandina Formation is the main hydrostratigraphic unit of the area, which comprises silt and clay with high content of organic matter. Dissolved arsenic concentrations in groundwater are high, reaching Ͼ400 g/L in some wells. Groundwater is reducing with general lack of detectable dissolved oxy- gen (DO) and contains low concentrations of nitrate and sulfate. Concentrations of dissolved Fe are high, which is in general in agreement with the reductive dissolution of ferric oxide and hydroxide hypothesis. Results of speciation modeling indicated the possibility of precipitation of siderite, and to less extent, vivianite for many samples. The log P CO2 values were extremely high (ϾϪ1.0 atm), suggesting production of CO 2 in redox reactions involving the organic matter in the sediments. Redox potential values calculated on the basis of different redox couples and field Eh measurement indicated redox disequilibrium. Hierarchical cluster analysis (HCA) performed in paired groups mode using the program PAST indicated highest degree of similarity among redox- sensitive elements NO 3 , Mn, Fe, PO 4 , SO 4 , As, and pH. Na and Cl form a distinct group, which indicate the influence of sea water. Bicarbonate generated in several redox reactions and carbonate dissolution was linked to almost all parameters and this holds even more for the electrical conductiv- ity (EC). Principal components analysis (PCA) yielded Principal Component 1 (PC1) correspon- ding to sea water, and Principal Component 2 (PC2) corresponding to redox reactions with generally high arsenic concentrations. In summary, combination of speciation modeling and multivariate statistics proved to be useful in testing of conceptual model of geochemical evolution of arsenic- rich groundwater. 1 INTRODUCTION Natural arsenic in concentrations above the safe drinking water limits of World Health Organization (10 g/L; WHO 2001), and above the national drinking water standard (50 g/L) is present in groundwater of the Bengal Delta Plain (BDP) in many districts of Bangladesh (Mukherjee & Bhattacharya 2001, Smedley & Kinniburgh 2002, Ahmed et al. 2004). The source of arsenic is geogenic and is related to the sediments deposited by the rivers Ganges (Padma), Brahmaputra Copyright © 2005 Taylor & Francis Group plc, London, UK (referred as Jamuna in Bangladesh) and Meghna (Nickson et al. 2000, Bhattacharya et al. 2002a,b). Arsenic contaminated groundwater is common in the aquifers of alluvial lowlands, comprising the floodplains of Padma and Brahmaputra (Jamuna) rivers, and also the Ganges Delta. In this paper, we present applications of geochemical modeling and multivariate statistics in development and support of conceptual model of arsenic behavior. 2 GEOLOGICAL SETTING The BDP is a large sedimentary basin drained by the Ganges, Brahmaputra (Jamuna) and Meghna (GBM) rivers (Fig. 1). Huge amounts of sediments have been transported and converged at the lower reaches, forming the pro-grading delta at the head of the Bay of Bengal. In general, two broad physiographic units characterize the BDP – elevated Pleistocene Terraces such as the Barind and Madhupur tracts, floored with thick surficial oxidized clay and silty clay deposits, and the Holocene lowlands. The Holocene lowlands include piedmont plains, flood plains, delta plains and coastal plains (Umitsu 1987 and 1993, Brammer 1996, Ahmed et al. 2004). The area of present investigation is located in the Meghna Deltaic Plain comprising coarse-grained channel-fill deposits and fine grained overbank deposits. During the late Holocene period, in several parts of the BDP, sediments were deposited in marshy environments, as evidenced by occurrence of continuous layers of peat (Umitsu 1993, Ravenscroft et al. 2001, Ahmed et al. 2004). The present study was carried out in an area of 18 km 2 covering the Ashuganj and Brahmanbaria Sadar Upazilas (sub-districts) in Brahmanbaria district in eastern Bangladesh (Fig. 2). The Chandina Deltaic Plain (CDP), the major physiographic units covering the study area (Bakr 1977), is generally flat and occurring at relatively higher levels than the surrounding floodplains. The 134 Figure 1. Map of Bangladesh showing the network of the rivers Ganges (Padma), Jamuna (Brahmaputra) and Meghna rivers and the location of Brahmanbaria area. The major geomorphic domains, Barind and Madhupur tracts of Pleistocene age (lighter tone) are seen distinctly within the vast tract of Holocene alluv- ium. (Resolution: 625 meters; MODIS Data Type: MODIS-PFM; MODIS Band Combination: 1, 4, 3) (Source map: http://modis.gsfc.nasa.gov/MODIS/IMAGE_GALLERY/MODIS1000027_md.jpg). Copyright © 2005 Taylor & Francis Group plc, London, UK sediments of the CDP are composed of silt, silty loam, silty clay belonging to the Chandina Formation. The Chandina Formation is overlain by the Meghna alluvium and underlain by the Pleistocene Madhupur Clay and Pliocene Dupi Tila Formation. Neotectonic uplifts and course shifting of the Meghna and Old Brahmaputra rivers have influenced sedimentation in this area, particularly during the late Holocene time. The source ter- rains of most of these sediments were located in the areas around the Shillong Plateau in the north and Tripura Hills on the east. A thick sequence of fine to very fine Holocene sediments overlies the Pleistocene Madhupur Clay and Pliocene Dupi Tila sediments. The Holocene sediments are gen- erally gray in color and contain high amount of organic matter while underlying older sediments are characterized by reddish-brown, light brown and yellowish brown color and contain only low amounts of organic matter (Table 1). Groundwater from the Holocene sandy sedimentary aquifer is extracted by shallow hand tube wells. Water levels in the Holocene aquifers fluctuate with annual recharge/discharge conditions, with a maximum depth of 5–7 m bgs in pre-monsoon months. During the monsoon most of the area if flooded and the groundwater level reaches the ground surface. The multiple aquifer system 135 Figure 2. Geological map of a part of the Chandina Delta Plain (CDP) in the Brahmanbaria district, Bangladesh showing the location of the groundwater sampling. Table 1. Hydrostratigraphy of the study area. Age Unit Predominant Lithology Hydrogeological characteristics Recent/Holocene Meghna Alluvium Grey clay, silt and fine sand Upper unconfined aquifer, arsenic rich Late Pleistocene Chandina Formation Grey silt, silty loam, silty clay Aquitard Early Pleistocene Madhupur Clay Reddish brown clay Aquitard Pliocene Dupi Tila Sands Yellowish brown medium to Lower aquifer, low in arsenic fine sand Copyright © 2005 Taylor & Francis Group plc, London, UK in this area is characterized by variable hydraulic conductivity and water quality. Water quality is often good except for occurrences of pockets of brackish water, remnants of paleo-seawater. Occurrences of biogenic methane gas have also been reported from a number of places with the BDP and particularly in the vicinity of the study area (Ahmed et al. 1998, Ravenscroft et al. 2001). Arsenic and iron concentrations are frequently high in the Holocene aquifers. However, their conc- entrations are significantly lower in the underlying Dupi Tila aquifers (BGS & DPHE 2001, van Geen et al. 2003). This deeper aquifer probably receives recharge at their outcrops in the Tippera Hills region, outside the political boundary of Bangladesh. 3MATERIALS AND METHODS Samples of groundwater were collected during late November, 2000 from 30 domestic and governmental tube wells placed at varying depths of 18–150 m (Fig. 2). Parameters like pH, redox potential (Eh), temperature, and electrical conductivity (EC) were taken in the field. The pH was measured using a Radiometer Copenhagen PHM 80 instrument using a combination electrode (pH C2401-7). The Eh was measured in a flow-through cell using a combined platinum electrode (MC408Pt) equipped with a calomel reference cell. Samples collected for analyses included: (a) filtered (using Sartorius 0.45 m online filters) for major anions; (b) filtered and acidified with suprapure HNO 3 (14 M) for the cations and other trace elements including arsenic (Bhattacharya et al. 2002b). Arsenic speciation was performed with Disposable Cartridges(r) (MetalSoft Center, PA) in the field, Meng et al. (2001). The cartridges adsorb As(V), but allows As(III) to pass through. Sulfide was precipitated in the field by addition of Zn acetate. Major anions, Cl Ϫ , and SO 4 2Ϫ were analyzed in filtered water samples, with a Dionex DX-120 ion chromatograph with an IonPac As14 column. NO 3 Ϫ -N and PO 4 3Ϫ -P was analyzed spectrophotometrically with a Tecator Aquatec 5400. Ammonium (NH 4 ϩ ) was analyzed spectrophotometrically with a Tecator Aquatec 5400 at 540 nm wavelength. The major cations (Ca, Mg, Na and K) and minor and trace elements (Fe, Mn, As) were analyzed by inductively coupled plasma (ICP) emission spectrometry (Varian Vista-PRO Simultaneous ICP-OES) at Stockholm University. Dissolved organic carbon (DOC) in the water samples were determined on a Shimadzu 5000 TOC analyzer (0.5mg/L detection limit with a precision of Ϯ10% at the detection limit. Speciation modeling was performed by program PHREEQC (Parkhurst 1995). Thermodynamic data for arsenic were taken from data base of pro- gram MINTEQA2 (Allison et al. 1991). Multivariate statistics analysis was performed to verify the hydrogeochemical similarities among geochemical parameters. The data were analyzed using multivariate statistics implemented in the program PAST (Hammer et al. 2001). 4 GENERAL HYDROGEOCHEMISTRY Selected results of the groundwater chemical analyses are presented in Table 2. In addition, Bhattacharya et al. 2004 (in press) provide more detailed discussion on the trends of spatial variability of water chemistry. Shallow groundwater (Ͻ50 m) in Brahmanbaria region was of Ca-Mg-HCO 3 and Ca-Na-HCO 3 types (Fig. 3a). Groundwater samples had very variable HCO 3 Ϫ (74–562 mg/L) and SO 4 2Ϫ (bdl-32.9 mg/L) concentrations. In the intermediate and deeper aquifers groundwater of Na-Cl-HCO 3 type was also found (Fig. 3b). Groundwater pH values were between 6.2 and 7.6. Field Eh values corrected with respect to hydrogen electrode were from ϩ0.180 to ϩ0.29 V, indicating moderately reducing conditions. However, these results do not represent redox status of groundwater, possibly because of aeration of groundwater in hand-pump wells and during pumping as discussed later. Concentrations of total arsenic (As tot ) in shallow wells varied from 10 to 335 g/L and in interm- ediate wells reached up to 439 g/L. Concentrations of dissolved Fe were highly variable, from 0.28 mg/L to 10.3 mg/L. However, no correlation between dissolved iron (Fe tot ) and As tot in shal- low samples and only low correlation in intermediate depth samples were observed (Bhattacharya et al. 2004, in press). Most of dissolved arsenic (up to 99.5%) was present as As(III). Concentrations 136 Copyright © 2005 Taylor & Francis Group plc, London, UK 137 Copyright © 2005 Taylor & Francis Group plc, London, UK of NH 4 ϩ were high in some shallow wells, reaching 12.2 mg/L. High concentrations of dissolved organic matter (DOC, up to 21.8mg/L) were consistent with reducing character of groundwater. Dissolved sulfide with concentrations up to 2.1 mg/L was found in several wells, indicating the presence of sulfate reduction. Spatial variability of the distribution of arsenic in the shallow ground- waters and its relationship with other chemical parameters is discussed in detail in Bhattacharya et al. (2004, in press). Three domains were defined: Domain 1 with high concentrations As tot , and PO 4 3Ϫ and low Fe tot , and anomalous SO 4 2Ϫ concentrations; Domain 2 with high concentrations of As tot , and PO 4 3Ϫ , and low concentrations of Fe tot , and sulfate; Domain 3 with low concentrations of As tot , and PO 4 , and with high Fe tot , and sulfate concentrations (Fig. 4). 5 GEOCHEMICAL MODELLING Results of calculations of saturation indices (SI) together with calculated log P CO2 values are pres- ented in Table 3. There was no significant complexation of Fe with other inorganic anions and the 138 Figure 3. Major ion characteristics of Brahmanbaria groundwaters plotted on a piper diagram . (a) Shallow wells (Ͻ50 ) (b) Intermediate wells (50–150 m, black circles) and deep wells (Ͼ150 m, data not included in discussion). Figure 4. Spatial variability of total As (As tot ) in the shallow wells of Brahmanbaria in eastern Bangladesh. Copyright © 2005 Taylor & Francis Group plc, London, UK principal aqueous species of Fe was Fe 2ϩ and minor species was FeHCO 3 ϩ . Concentrations of Fe(III) were low, and the principal species were Fe(OH) 3 0 and, at lower pH, Fe(OH) 2 ϩ . Low Mn conc- entrations in groundwater within the reduced domains could possibly be due to precipitation of rhodochrosite (MnCO 3 ) (Sracek et al. 2000, McArthur et al. 2001, Ahmed et al. 2004). Calculated log P CO2 values were very high, reaching in some cases values higher than – 1.0. This is related to the generation of CO 2 in redox reactions like dissolution of ferric oxide and hydroxides in reaction with organic matter. This is consistent with high calculated values of DIC (up to 1.27 ϫ 10 Ϫ2 mol/L). Many samples (60%) were at equilibrium or supersaturated with respect to siderite (FeCO 3 ) sugg- esting that this mineral phase might have acted as a sink for dissolved iron. Some samples (30%) were also supersaturated with respect to vivianite Fe 3 (PO 4 ) 2 .8H 2 O, but the degree of saturation is generally lower than in the case of siderite. Few samples (13.3%) are also supersaturated with respect to rhodochrosite. Some samples are close to equilibrium with calcite and dolomite (not shown), but saturation is reached only in very limited number of samples. The speciation program was also used to calculate Eh values based on As(III)/As(V) couple and S(VI)/S(-II) couple determined analytically. Results of these calculations are presented in Table 4. Typical feature is the strong disequilibrium between measured field Eh values adjusted with respect to hydrogen electrode and values of Eh calculated on the basis of arsenic couple. Values of redox potential based on arsenate to arsenite ratios are lower than the field Eh values. This holds even more for sulfur redox couple, suggesting strong redox disequilibrium. However, the field Eh values truncated at ϩ0.180V are possibly unreliable and must be interpreted with caution. There 139 Table 3. Results of speciation calculations. Sample ID Depth (m) SI siderite SI vivianite SI rhodochrosite logP CO2 (atm) DIC (mol/L) 1 18.3 Ϫ0.66 Ϫ3.80 Ϫ0.50 Ϫ1.94 2.73 ϫ 10 Ϫ3 2 21.3 0.81 2.57 0.57 Ϫ2.37 2.37 ϫ 10 Ϫ3 3 21.3 Ϫ0.52 Ϫ2.31 Ϫ0.12 Ϫ2.41 2.09 ϫ 10 Ϫ3 4 21.3 0.33 0.46 Ϫ1.11 Ϫ1.10 1.10 ϫ 10 Ϫ2 5 21.3 Ϫ0.02 Ϫ0.72 Ϫ1.48 Ϫ0.90 1.09 ϫ 10 Ϫ2 6 22.9 Ϫ1.13 Ϫ4.29 Ϫ1.79 Ϫ1.63 2.31 ϫ 10 Ϫ3 7 24.4 0.47 Ϫ0.14 Ϫ0.70 Ϫ1.20 8.73 ϫ 10 Ϫ3 8 24.4 0.15 Ϫ1.23 Ϫ1.51 Ϫ1.19 4.82 ϫ 10 Ϫ3 9 25.9 Ϫ0.57 Ϫ2.72 Ϫ1.78 Ϫ0.93 1.01 ϫ 10 Ϫ2 10 27.4 0.92 2.36 0.17 Ϫ2.09 3.62 ϫ 10 Ϫ3 11 27.4 0.05 Ϫ1.65 Ϫ1.14 Ϫ1.67 1.01 ϫ 10 Ϫ2 12 27.4 0.72 Ϫ0.70 Ϫ0.73 Ϫ1.47 3.96 ϫ 10 Ϫ3 13 27.4 0.11 Ϫ0.38 Ϫ1.42 Ϫ1.30 6.90 ϫ 10 Ϫ3 14 28.0 Ϫ0.22 Ϫ1.56 Ϫ1.68 Ϫ0.82 1.14 ϫ 10 Ϫ2 15 29.0 0.59 2.45 0.51 Ϫ2.78 1.16 ϫ 10 Ϫ3 16 29.0 0.16 Ϫ1.11 Ϫ1.15 Ϫ1.00 1.01 ϫ 10 Ϫ2 17 32.0 Ϫ0.07 0.01 Ϫ0.47 Ϫ1.96 2.04 ϫ 10 Ϫ3 18 32.0 0.39 Ϫ0.43 Ϫ0.99 Ϫ1.48 8.23 ϫ 10 Ϫ3 19 32.0 0.34 0.40 Ϫ0.69 Ϫ1.37 8.66 ϫ 10 Ϫ3 20 32.0 0.05 Ϫ1.06 Ϫ1.44 Ϫ0.78 1.24 ϫ 10 Ϫ2 21 39.6 0.68 2.04 Ϫ0.52 Ϫ1.68 9.74 ϫ 10 Ϫ3 22 41.1 Ϫ0.46 Ϫ1.92 Ϫ1.56 Ϫ1.13 1.03 ϫ 10 Ϫ2 23 54.9 0.94 1.80 Ϫ0.68 Ϫ1.48 8.26 ϫ 10 Ϫ3 24 59.4 0.08 Ϫ0.03 Ϫ1.51 Ϫ1.31 1.24 ϫ 10 Ϫ2 25 62.5 Ϫ0.55 Ϫ1.91 Ϫ1.41 Ϫ1.42 1.13 ϫ 10 Ϫ2 26 64.0 0.82 2.66 0.48 Ϫ2.18 2.45 ϫ 10 Ϫ3 27 64.0 0.12 Ϫ0.61 Ϫ1.62 Ϫ0.98 1.27 ϫ 10 Ϫ2 28 68.6 Ϫ0.73 Ϫ4.52 Ϫ1.10 Ϫ1.17 8.01 ϫ 10 Ϫ3 29 73.2 Ϫ0.10 Ϫ1.86 Ϫ2.12 Ϫ1.05 1.27 ϫ 10 Ϫ2 30 75.6 Ϫ0.15 Ϫ0.73 Ϫ1.49 Ϫ1.28 8.61 ϫ 10 Ϫ3 Copyright © 2005 Taylor & Francis Group plc, London, UK also is a possibility of precipitation of secondary sulfide minerals like mackinawite. This mineral is a precursor of pyrite and may incorporate some arsenic, acting as a sink for arsenic in ground- water where sulfate reduction takes place. 6 MULTIVARIATE STATISTICS Results of Hierarchical Cluster Analysis (HCA) performed in Ward’s mode using the program PAST are given in Figure 5. They indicate highest degree of similarity among NO 3 Ϫ , Mn, Fe tot , PO 4 3Ϫ , SO 4 2Ϫ , As tot , and pH. Most of them are redox sensitive species, except for PO 4 3Ϫ which is linked to Fe due to its release during reductive dissolution of ferric oxide and hydroxides. The effect of pH is however not clear, although this parameter plays a role in precipitation of minerals like siderite and vivianite. Na ϩ and Cl Ϫ form a distinct group, which most likely indicates relict sea water entrapped in the aquifers. Ca 2ϩ and Mg 2ϩ are separated from Na ϩ and Cl Ϫ because they are related not only to relict sea water in the sediments, but also to processes like dissolution of carbonates, and weathering of silicates enhanced by production of CO 2 in redox reactions. HCO 3 Ϫ 140 Table 4. Comparison of redox data. Sample ID/ Eh (V) Eh (V) Parameter Field Eh (V) As(V)/As(III) S(VI)/S(-II) SI mackinawite 3 ϩ0.180 0.029 Ϫ0.252 3.09 16 ϩ0.180 0.059 Ϫ0.205 1.66 19 ϩ0.180 Ϫ0.011 Ϫ0.225 1.96 21 ϩ0.180 Ϫ0.073 Ϫ0.247 2.14 24 ϩ0.180 Ϫ0.018 ϽϪ0.254 1.55 Similarity HCO 3 - CI - Na + NO - 3 Mn Fe tot PO 4 3- K + pH SO 4 2- Mg 2+ Ca 2+ EC As tot 0 -100 -200 -300 -400 -500 -600 -700 -800 12345678 91011121314 Figure 5. Results of Hierarchical Cluster Analysis (HCA). Copyright © 2005 Taylor & Francis Group plc, London, UK generated in several redox reactions and carbonate dissolution is linked to almost all parameters and this holds even more for EC. Results of Principal Components Analysis (PCA) presented in Figure 6 indicate two principal components and identified as: PC1 with high loadings for EC, Na ϩ , Cl Ϫ , and HCO 3 Ϫ , and PC2 with high loading for HCO 3 Ϫ and with relatively low, but significant loading for arsenic. The PC1 corresponds to the influence of relict seawater entrapped in the sediments to the groundwater, and PC2, which corresponds to the impact of redox reactions. These principal components explain 92.01% and 6.71%, respectively, of total variance in sample set. The samples at the bottom right (27 and 28, Fig. 6) indicate strong influence of saline water, most likely reflect to be relict seawa- ter entrapped in the sediments. On the other hand, samples at the top of the graph (4, 11, 21, 22, 25 etc., Fig. 6) are strongly influenced by redox reactions and they generally have high arsenic concentrations. Samples at the bottom left (1, 2, 3, 6, 17 etc., Fig. 6) are relatively less influenced by both processes and have low arsenic concentrations. It seems that redox processes relatively less influence the samples showing signature of relict saline water from the marine sources. However, the impact of palaeo-seawater relicts in the aquifers seems to be a local phenomenon in the Brahmanbaria area, which is also seen in many other areas of Bangladesh. 7 DISCUSSION AND CONCLUSIONS The conceptual model of arsenic and iron behavior in groundwater in Bangladesh can be summar- ized as follows: (a) reductive dissolution of ferric oxide and hydroxides in reaction with organic matter like peat after consumption of more favored electron acceptors in a reaction like: 141 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 400 300 200 100 1000 Component 1 Component 2 2000 3000 0 -100 -200 Figure 6. Results of Principal Components Analysis (PCA). Copyright © 2005 Taylor & Francis Group plc, London, UK [...]... contaminated groundwater in alluvial aquifers from Delta Plains, Eastern India: Options for safe drinking water supply Int Jour Water Res Management 13(1): 79–82 Bhattacharya, P., Frisbie, S.H., Smith, E., Naidu, R., Jacks, G & Sarkar B 2002a Arsenic in the Environment: A Global Perspective In B.Sarkar (ed.) Handbook of Heavy Metals in the Environment (Chapter 6), New York: Marcell Dekker Inc., pp 145–215... 2001 Arsenic in groundwater: testing pollution mechanism for sedimentary aquifers in Bangladesh Water Resour Res 37: 109–117 Meng, X., Korfiatis, G.P., Christodoulatos, C & Bang, S 2001 Treatment of arsenic in Bangladesh well water using a household co-precipitation and filtration system Wat Res 35: 2805–2810 Mukherjee, A.B & Bhattacharya, P 2001 Arsenic in groundwater in the Bengal Delta Plain: Slow... Routh, J 2002b Arsenic in groundwater of the Bengal Delta Plain aquifers in Bangladesh Bull Env Cont Toxicology 69: 538–545 Bhattacharya, P., Ahmed, K.M., Broms, S., Fogelström, J., Jacks, G., Sracek, O & Routh, J 2004 Mobility of arsenic in groundwater in a part of Brahmanbaria district, NE Bangladesh In: Naidu, R., Smith, E., Smith, L., Smith, J and Bhattacharya, P (Eds.) Managing Arsenic in the Environment:... geochemical overview In: A.L Ramanathan, V Subramanian & R Ramesh (eds.) Proc International Seminar on Applied Hydrogeochemistry’ Annamalan University, Tamil Nadu, India, pp 47–56 Umitsu, M 1987 Late Quaternary sedimentary environment and landform evolution in the Bengal Lowland Geog Rev Japan (Ser B) 60: 164 –178 Umitsu, M 1993 Late Quaternary sedimentary environments and landforms in the Ganges Delta... no important minerals of arsenic in groundwater with low sulfate concentrations and adsorption of arsenic is limited in alkaline pH region (Langmuir 1997) Results of speciation modeling are consistent with the conceptual model (Bhattacharya et al 1997, Nickson et al 2000, Bhattacharya et al 2002a,b, Ahmed et al 2004) because there are very high PCO2 and DIC values in samples with high arsenic concentrations... high loadings for typical sea water ions such as Naϩ and ClϪ Sulfate was less useful indicator because its concentration was very low The combination of geochemical modeling and multivariate statistics has been proved to be a useful tool in testing the hypothesis about arsenic release mechanism ACKNOWLEDGEMENTS The authors would like to acknowledge the Swedish Research Council (VR) and Sida-SAREC for... respect to siderite, and, to less extent with respect to vivianite Presence of ferrous carbonate and phosphate minerals was also confirmed by sequential extraction (Ahmed et al 2004) Multivariate statistics groups together redox-sensitive parameters like Fe, Mn, As, and DOC Phosphate is also in the same group because it is indirectly linked to redox processes through its release in reductive dissolution... sulfate reduction takes place, minerals like mackinawite also act as active sinks for dissolved iron However, sulfate concentrations are generally low and this process does not seem to be very significant This means that correlation between dissolved iron and arsenic frequently observed at sites contaminated by arsenic may be disturbed because behavior of dissolved arsenic is more conservative than... and Sida-SAREC for supporting the research on High arsenic groundwater in Bangladesh since January 1997 The authors would like to thank Andreas Mende for his valuable suggestions to improve the manuscript REFERENCES Ahmed, K.M., Hoque, M., Hasan, M.K., Ravenscroft, P & Chowdhury, L.R 1998 Origin and occurrence of water well methane gas in Bangladesh aquifers Jour Geol Soc India 51: 697–708 Ahmed, K.M.,... Bengal Delta Plain: Slow Poisoning in Bangladesh Environmental Reviews 9(3): 189–220 Nickson, RT, McArthur, J.M., Ravenscroft, P., Burgess, W.G & Ahmed, K.M 2000 Mechanism of arsenic release to groundwater, Bangladesh and West Bengal Appl Geochem 15(4): 403–413 Parkhurst, D.L 1995 Users Guide to PHREEQC-A Computer Program for Speciation, Reaction-Path, Advective-Transport, and Inverse Geochemical Calculations, . precipitation of secondary sulfide minerals like mackinawite. This mineral is a precursor of pyrite and may incorporate some arsenic, acting as a sink for arsenic in ground- water where sulfate reduction. Ϫ0.018 ϽϪ0.254 1.55 Similarity HCO 3 - CI - Na + NO - 3 Mn Fe tot PO 4 3- K + pH SO 4 2- Mg 2+ Ca 2+ EC As tot 0 -1 00 -2 00 -3 00 -4 00 -5 00 -6 00 -7 00 -8 00 12345678 91011121314 Figure 5. Results of Hierarchical. Barind and Madhupur tracts, floored with thick surficial oxidized clay and silty clay deposits, and the Holocene lowlands. The Holocene lowlands include piedmont plains, flood plains, delta plains