Geochemical characterisation of shallow aquifer sediments of Matlab Upazila, Southeastern Bangladesh — Implications for targeting low-As aquifers

Journal of Contaminant Hydrology 99 (2008) 137–149

Contents lists available at ScienceDirect

Journal of Contaminant Hydrology j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j c o n h yd

Geochemical characterisation of shallow aquifer sediments of Matlab Upazila, Southeastern Bangladesh — Implications for targeting low-As aquifers Mattias von Brömssen a,h,⁎, Sara Häller Larsson a, Prosun Bhattacharya a, M. Aziz Hasan a,b, Kazi Matin Ahmed b, M. Jakariya c, Mohiuddin A. Sikder d, Ondra Sracek e, Annelie Bivén a, Barbora Doušová f, Claudio Patriarca g, Roger Thunvik a, Gunnar Jacks a a KTH-International Groundwater Arsenic Research Group, Department of Land and Water Resources Engineering, Royal Institute of Technology (KTH), Teknikringen 76, SE-100 44 Stockholm, Sweden b Department of Geology, University of Dhaka, Curzon Hall Campus, Dhaka, 1000, Bangladesh c NGO Forum for Drinking Water Supply and Sanitation, 4/6 Block E, Lalmatia, Dhaka-1206, Bangladesh d Department of Environmental Sciences, Stamford University, Dhanmondi, Dhaka-1209, Bangladesh e Institute of Geological Sciences, Faculty of Science, Masaryk University, Kotlářská 2, 611 37 Brno, Czech Republic f Department of Solid State Chemistry, Institute of Chemical Technology in Prague, Technicka 5, 166 28 Prague 6, Czech Republic g Sapienza università di Roma, Dipartimento Scienze della Terra, P.le A.Moro 5, 00185 Roma, Italy h Department of Soil and Water Environment, Ramböll Sweden AB, Box 4205, SE-102 65 Stockholm, Sweden

a r t i c l e

i n f o

Article history: Received 18 March 2008 Received in revised form 12 May 2008 Accepted 14 May 2008 Available online 24 May 2008 Keywords: Arsenic Groundwater Bangladesh Redox Hydrogeochemistry Shallow sedimentary aquifer

a b s t r a c t High arsenic (As) concentrations in groundwater pose a serious threat to the health of millions of people in Bangladesh. Reductive dissolution of Fe(III)-oxyhydroxides and release of its adsorbed As is considered to be the principal mechanism responsible for mobilisation of As. The distribution of As is extremely heterogeneous both laterally and vertically. Groundwater abstracted from oxidised reddish sediments, in contrast to greyish reducing sediments, contains significantly lower amount of dissolved arsenic and can be a source of safe water. In order to study the sustainability of that mitigation option, this study describes the lithofacies and genesis of the sediments within 60 m depth and establishes a relationship between aqueous and solid phase geochemistry. Oxalate extractable Fe and Mn contents are higher in the reduced unit than in the oxidised unit, where Fe and Mn are present in more crystalline mineral phases. Equilibrium modelling of saturation indices suggest that the concentrations of dissolved Fe, Mn and PO3− 4 -tot in groundwater is influenced by secondary mineral phases in addition to redox processes. Simulating AsIII adsorption on hydroferric oxides using the Diffuse Layer Model and analytical data gave realistic concentrations of dissolved and adsorbed AsIII for the reducing aquifer and we speculate that the presence of high PO3− 4 -tot in combination with reductive dissolution results in the high-As groundwater. The study confirms high mobility of As in reducing aquifers with typically dark colour of sediments found in previous studies and thus validates the approach for location of wells used by local drillers based on sediment colour. A more systematic and standardised colour description and similar studies at more locations are necessary for wider application of the approach. © 2008 Elsevier B.V. All rights reserved.

⁎ Corresponding author. Department of Soil and Water Environment, Ramböll Sweden AB, Box 4205, SE-102 65 Stockholm, Sweden. E-mail addresses: [email protected] (M. von Brömssen), [email protected] (P. Bhattacharya), [email protected] (K.M. Ahmed), [email protected] (M. Jakariya), [email protected] (M.A. Sikder), [email protected] (O. Sracek), [email protected] (B. Doušová), [email protected] (C. Patriarca). 0169-7722/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.jconhyd.2008.05.005

1. Introduction Groundwater from shallow, Holocene alluvial aquifers of Bangladesh contains elevated levels of geogenic arsenic (As) (BGS and DPHE, 2001; Bhattacharya et al., 2002a,b; Ahmed et al., 2004; Bhattacharya et al., 2006a,b, 2007). Several

138

M. von Brömssen et al. / Journal of Contaminant Hydrology 99 (2008) 137–149

million tubewells have been installed in these aquifers during the last decades, in order to provide safe water to the vast number of people exposed to waterborne diseases derived from the use of bacterial contaminated surface water. Today, shallow hand tubewells are the principal source of drinking water in rural Bangladesh, and a population within a range from 35 to 77 million is at risk of drinking groundwater with elevated As concentrations (Smith et al., 2000; Chakraborti et al., 2004; Kapaj et al., 2006). Most of the tubewells (~90%) have been installed on private initiatives by local drillers or masons, presumably by the simple hand percussion drilling method, a low cost technique for well installation in the soft unconsolidated alluvial sediments of Bangladesh. Reductive dissolution of Fe(III)-oxyhydroxides under strongly reducing conditions in the young sediments seems to be the prime cause of As mobilization (Bhattacharya et al., 1997; Nickson et al., 1998; BGS and DPHE, 2001; Harvey et al., 2002; Stuben et al., 2003; Ahmed et al., 2004; McArthur et al., 2004; Swartz et al., 2004). The distribution of As is extremely heterogeneous, both laterally and vertically, and consequently, the “patchy distribution” has often been explained in the terms of “local variations in sedimentary characteristics, hydrogeological and hydrogeochemical conditions” (BGS and DPHE, 2001; Ravenscroft, 2001; Bhattacharya et al., 2002a,b; McArthur et al., 2004; Bhattacharya et al., 2006a,b; Bhattacharya et al., 2008). Targeting shallow and low-As aquifers in Bangladesh has been difficult. Nickson

et al. (2000) argued that identification of the mechanism of As release to groundwater may provide a framework to guide the placement of new wells for abstracting groundwater with acceptable concentrations of As following the drinking water standards of Bangladesh (50 µg/L) and more specifically to comply with the WHO drinking water guideline (10 µg/L; WHO, 2004). Because of the extremely heterogeneous distribution of As, BGS and DPHE (2001) reported that detailed mapping and investigation of As concentrations in tubewells, as an option for the placement of As-safe tubewells, would be too difficult. Nonetheless, switching to arsenic safe tubewells that have been tested As safe and reinstallation of shallow tubewells to alternate depths have emerged from people’s own initiative (Jakariya et al., 2007). These so called well-switching schemes have been encouraged by various researchers, including van Geen et al. (2002). In recent studies (van Geen et al., 2004; von Brömssen et al., 2005; von Brömssen et al., 2007; Stollenwerk et al., 2007; Hasan et al., in press; Pal and Mukherjee, in press) it is argued that groundwater from red (or orange/brown) sediments, within shallower depth (b150 m), are As safe. These sediments were deposited in alluvial settings under reducing conditions and were subsequently exposed to weathering and oxidation during the last glacial maximum (Umitsu, 1987; Umitsu, 1993). At that time, the surface sediments were eroded and weathered and then flushed (Goodbred et al., 2003). Von Brömssen et al. (2005, 2007)

Fig. 1. Location of Matlab Upazila (demarcated by embossed line) in southeastern Bangladesh (inset) and regional surface geology.

M. von Brömssen et al. / Journal of Contaminant Hydrology 99 (2008) 137–149

demonstrated a distinct correlation between the colour of the aquifer sands, groundwater redox-conditions and the relative risk for mobilization of geogenic As. It was suggested that the colours of the sediments could be used as a simple tool by the local drillers to target As-safe aquifers. This study indicated that it is possible to assess the relative risk of elevated concentrations of As in aquifers if the colour characteristics of the sediments are known which would help the local drillers to target safe aquifers for the placement of new tubewells. Thus following a simplified use of this sediment colour concept, a sustainable mitigation approach can be established in Matlab Upazila, and other areas in Bangladesh with similar geological settings, as well as elsewhere in the world, for improving the safe water coverage based on the initiatives of the local drillers. On the other hand, if the sustainability cannot be proved, such initiatives and the current practice of the local drillers needs to be stopped immediately before a large investment is made for installation of new wells. The objective of this study was to characterise the different sediments from shallow aquifers of Matlab Upazila, describe the lithofacies and genesis of sediments and to establish the relationship between aqueous and solid phase geochemistry and colour of sediments which is a prerequisite for further studies on the sustainability of the oxidised low-As aquifers for drinking water supplies. 2. The study area Matlab Upazila is located approximately 60 km south-east of Dhaka to the east of the confluence of the rivers of Ganges (Padma), Brahmaputra (Jamuna) and Meghna (Fig. 1). Bangla-

139

desh has a tropical climate with a mild winter season between October and March, a hot humid summer from March to June and a warm and humid monsoon season from June to October when most of the precipitation occurs. The study area is situated within the Meghna flood plain and is characterised by meander channels and scrolls, natural levees and back swamps formed by the river system. The low-lying landscape is naturally flooded each year during the monsoon and the surface sediments are represented by Holocene alluvial silt. The shallow Holocene aquifers of Matlab Upazila are heavily affected by geogenic As that is very heterogeneously distributed over the Upazila (Jakariya and Bhattacharya, 2007). At the study site a thick layer of black to grey sediments, down to approximately 40 m, overlies an oxidised unit of yellowish-grey to reddish-brown sediments (von Brömssen et al., 2007). The groundwater level is within a depth of 5 m below ground surface and fluctuates 3–4 m annually in response to recharge–discharge conditions (Bangladesh Water Development Board, unpublished data). 3. Materials and methods 3.1. Groundwater and sediment sampling Groundwater samples were collected from hand tubewells (n = 40) in two villages, Dighaldi and Mubarakdi of Matlab. The depth of the tubewells and colour of the sediments (as described by the drillers) at the screened depth were recorded by interviewing the drillers who installed the wells. Subsequently, the groundwater composition was linked to the colour characteristics of the sediments. Four colours were used to describe washed sediments characteristic in the area: black,

Fig. 2. Photographs of the 10 sediment core samples that were used for extraction tests and mineralogical studies, the numbers on top of the cores represent the depth (in meter), these numbers are used to name the samples. The colour descriptions correspond to the colours described in von Brömssen et al. (2007) followed by local drillers in Matlab Upazila.

140

M. von Brömssen et al. / Journal of Contaminant Hydrology 99 (2008) 137–149

described by von Brömssen et al. (2007), i.e. black, white, offwhite and red sediments (Fig. 2) were used for the mineralogical and geochemical investigations. 3.2. Analytical work

Fig. 3. A close-up photograph of the bioturbated shale at the depth of 36 m (118 ft), the circle mark the vegetation material sampled for 14C-dating. The diameter (∅) of the core is 5 cm.

white, off-white (buff), and red (orange/brown). Sampling, colour classification, analysis of tubewell water and evaluation of the respective hydrogeochemical characteristics for the different aquifer sediments were carried out during 2004 as described in von Brömssen et al. (2007). During January– February 2005, a new set of samples were collected from the same tubewells. The samples from the tubewells represent typical ground water chemistry from both oxidised and reduced units, as confirmed by previous investigation at the site (von Brömssen et al. 2007). One undisturbed core drilling was conducted at Mubarakdi village and samples were collected by a modified split spoon method down to the depth of 60 m; 0.3 m core samples were collected in 5 cm diameter plastic tubes. The drilling was performed with a combination of a hammer technique and a donkey pump. Core samples were taken every 1.5 m down to the depth of 30 m and between 30 and 60 m at 0.6 m interval. The core samples were sealed at both ends with wax immediately after recovery in the field to avoid any oxidation. Later, the core samples were split for lithological studies (Fig. 2), one half was preserved for lithological studies and the other half was used for sequential extraction and mineralogical studies. When the core samples were split, a coal-like vegetation remains (presumably from a root) was found in the bioturbated shale unit at the depth of 36 m (Fig. 3). The vegetation material was picked with clean inorganic tools and wrapped in aluminium foil for 14C-dating. Ten sediment samples representing the whole depth and colour spectra

3.2.1. Groundwater analyses Alkalinity (HCO−3) was determined according to the standard method SS-EN ISO 9963-2 (SIS, 1996) by titrating the filtered unacidified samples with 0.02 M HCl to pH 4.5. Other major anions, F−, Cl−, and SO2− 4 were analysed using the filtered unacidified water samples, with a Dionex DX-120 ion chroma+ tograph with an IonPac As14 column. NO−3, PO3− 4 -tot and NH4 was analysed spectrophotometrically with a Tecator Aquatec 5400. The major cations (Ca, Mg, Na and K) and minor and trace elements (Fe, Mn, As) were analyzed analysed by inductively coupled plasma emission spectrometry (ICP-OES) at the Department of Geology, Stockholm University. Dissolved organic carbon (DOC) in the water samples was determined on a Shimadzu 5000 TOC analyser (0.5 mg/L detection limit with a precision of ±10%). The 14C analysis was performed at the Radiocarbon Dating Laboratory in Lund using Single Stage Accelerator Mass Spectrometry (SSAMS). 3.2.2. Mineralogical investigations Selected sediment samples were analysed under the stereomicroscope in order to study the bulk minerals responsible for the colours of the sediments. X-ray fluorescence (XRF) and X-ray diffraction (XRD) analyses were carried out at the Institute of Chemical Technology in Prague, Czech Republic. 3.2.3. Sequential extractions Sequential extraction was carried out to quantify the amount of reactive components such as Fe and Mn in the sediments and their relationship with As. Ten selected core sediment samples were sequentially leached using i) deionized water (DIW) for quantification of the water soluble fraction of As and other trace elements; ii) 0.01 M NaHCO3 for the release of elements under high pH conditions; iii) 1 M Naacetate (C2H3NaO2, NaAc) for elements bound to carbonates and phosphates phases (Dodd et al., 2000; Ahmed et al., 2004); iv) 0.2 M oxalate (NH4C2O4, Oxalate) for quantification of Fe, Al, and Mn bound to amorphous oxides and hydroxides in the sediments; v) 0.2 M oxalate (NH4C2O4) +0.1 M ascorbic acid (Oxalate+AA) for residual amount of Fe, Al and Mn bound to oxides and hydroxides including crystalline phases; and vi) 7 M HNO3 for determining As and other elements associated with the non-silicate minerals (Table 1). The scheme followed

Table 1 Sequential extraction procedure adopted for the study Fraction

Extractant

Extracting condition

SSR

Wash step

1 (DIW) 2 (NaHCO3) 3 (NaAc)

DIW 0.01 M NaHCO3 (pH 8.65) 1 M Na-acetate (C2H3NaO2)

1:25 1:25 1:50

25 mL DIW 25 mL DIW 25 mL DIW

4 (Oxalate) 5 (Oxalate+AA)

0.2 M oxalate (NH4C2O4) 0.2 M oxalate (NH4C2O4) + 0.1 M ascorbic acid 7 M HNO3

2 h shaking, 20 °C, pH adjusted to 6.95 with NaOH 2 h shaking, 20 °C 2 h shaking, 20 °C, pH adjusted to 5 with acetic acid glacial (C2H4O2) 4 h shaking, 20 °C, in the dark 30 min in water bath at 96 °C ± 3 °C in the light

1:25 1:25

2 h on sand bed, boiling.

1:15

25 mL DIW 25 mL 0.2 M oxalate (NH4C2O4), 10 min shaking in the dark None

6 (HNO3)

M. von Brömssen et al. / Journal of Contaminant Hydrology 99 (2008) 137–149

141

Table 2 Analytical data used for adsorption simulations using Visual MINTEQ v. 2.53 and diffusive layer model by Dzombak and Morel (1990) Analytical data FeOx As PO3− 4 AsOx

Min

Max

Corresponding input data for simulation

mg/kg µg/L mg/L mg/kg

700 1 0.5 0.5

15 000 350 8 1.5

Hfo As (dissolved) PO3− 4 (dissolved) As (assumed to be adsorbed)

g/L mg/L

5 0.5 6.3 25 2.1

100 8.0 7.1 25 7.77

Total amount (solid phase only) Fixed dissolved concentration Fixed Fixed at 25 Total amount (adsorbed + dissolved) Used for unit conversion for the model

Input data for simulation Hfo PO3− 4 pH T As L/S ratio aquifer

Boundary condition

°C mg/L 0.238

the methods described by Wenzel et al. (2001) and Bhattacharya et al. (2006a,b) and was performed on 1 g air-dried homogenised sediment sample in a 50 mL centrifuge tube. Between each step, the sediment was washed. The extracts were preserved by acidification with 0.5 mL ultrapure 14 M HNO3/100 mL. Blanks were used in each step so that impurities could be subtracted from the extractants. 3.3. Geochemical modelling Thermodynamic relationships between species in solution and aquifer solid phases have been investigated through the degree of saturation with respect to minerals defined as saturation indices (SI): SI ¼ log IAP=Ksp




where IAP is the ion activity product and Ksp is the solubility product for a given temperature. When SI = 0 (IAP = Ksp) the solution is at thermodynamic equilibrium with respect to a specific mineral and when SI N 0 the water is supersaturated with respect to a mineral and vice versa if SI b 0. Chemical equilibrium modelling of SI was thus used to identify possible sinks, sources and reactions in the aquifer system (Sracek et al., 2004). Saturation indices calculations were performed using PHREEQC version 2.14.2 (Parkhurst and Appelo, 1999) with the WATEQ4F thermodynamic database. Eh values measured in field and corrected with respect to standard hydrogen electrode (SHE) were used for speciation of redoxcouples. This means that the model calculates the activities of the different species of each element and then uses these activities for calculation of saturation indices. As redox potential (Eh) as a proxy for redox condition for modelling is valid only under conditions when there is redox equilibrium, a sensitivity analysis was done to evaluate the impact of Eh on the resulting SI values by altering Eh in the model. Adsorption of As for aquifer conditions was simulated as surface complexation reactions on hydroferric oxides (Hfo) with the Diffuse Layer Model (DLM) in Visual MINTEQ 2.53 (Dzombak and Morel, 1990; Allison et al., 1991). The system hydroferric oxides, adsorbed and dissolved AsIII and PO3− 4 were simulated and compared with analytical data. Default parameter based on Dzombak and Morel (1990) was used for specific surface (600 m2/g). The modelling approach was based on the assumptions that As adsorption was controlled by pH, Hfo content, species of As (AsIII or AsV) and presence of available

competing ions (Smedley and Kinniburgh, 2002; Sracek et al., 2004; Gustafsson and Bhattacharya, 2007). Input data for the modelling came from both this and previous studies (Table 2). The amounts of Hfo and As were calculated based on oxalate extractions data for Fe and As respectively (FeOx, AsOx) and the L/S ratio of alluvial aquifer properties. Dissolved As concentration was added to AsOx amount giving total available As in the system. Dissolved PO3− 4 -tot concentration was fixed assuming dissolved PO3− 4 -tot was controlled by secondary mineral phases, and thus, only adsorbed amount was simulated. This assumption was based on the simulations of SI of mineral phases including PO3− 4 -tot (see below). 4. Results 4.1. Sequence of aquifer sediments The colour of the sediments from the uppermost sequence is black to greyish while the sediments from the lower sequence are reddish, yellowish to whitish (Fig. 2). The major units are separated by 5 m thick hard clayey shale (Figs. 3 and 4). The lithological log prepared from the core samples with the diagnostic sedimentological characteristics of the major units are presented in Fig. 4. The uppermost sequence (Unit 1, Fig. 4) consist of a partly oxidised topsoil of 2 m covering 5 m dark coloured clay and 28 m of sand, silt and clay in a fining upward sequence down to a depth of 35 m below the surface. Cross beddings and trough cross beddings identified in the core samples reveal that the deposition of the sediments occurred in a tide influenced flood plain environment. This uppermost sequence is of Holocene age and has not undergone any extensive weathering or oxidation and represents reducing aquifer conditions. Biotite and other dark coloured ferromagnesian and opaque minerals are responsible for the dark colours. These minerals are found in bands and together with organic matters at the depths of 15–20 m. The bioturbated hard shale (Unit 2A, Fig. 4) separating the two major units is grey in colour and indicate shallow marine depositional environment. The vegetation matter (root) incorporated in the shale gave an age of approximately 8000 yr BP from 14C-analysis and suggests that the marine transgression took place around that time which is consistent with other published studies (Goodbred et al., 2003). During 7000–9800 yr BP bioturbated sediments were deposited in the Bengal basin. An unconformity has been identified between the shale and the

142

M. von Brömssen et al. / Journal of Contaminant Hydrology 99 (2008) 137–149

Fig. 4. Lithological log with the diagnostic sedimentological characteristics of the sequence of sedimentary aquifers of the study area based on the recovered sediment cores.

uppermost reduced sequence confirming that the underlying sediments were eroded and exposed to weathering and oxidation. Thus the two major units correspond to two interglacial periods. However, no dating could be done for the underlying sequence due to the lack of organic matter.

Investigations of the core samples derived from the lower oxidised unit (Unit 2B, Fig. 4) show that these sediments were also exposed to weathering and oxidation. They have lower abundance of biotite. Fe(III)-oxyhydroxides coatings on quartz, feldspars and other mineral grains are responsible

Table Table 33 Mineralogical composition of of sediment sediment samples samples investigated investigated from from the the sediment sediment core coreprofile profile Mineralogical composition

* See Figure 4 for the position of the respective sediment samples in the sedimentary sequence * See Figure 4 for the position of the respective sediment samples in the sedimentary sequence

M. von Brömssen et al. / Journal of Contaminant Hydrology 99 (2008) 137–149

for the reddish colour of the sediments. This unit was also deposited in a fluvial and flood plain environment and is represented by fining upward sequences of sand, silt and clay. Clayballs incorporated in the sand (see Fig. 2, sample 54) indicate that some of sediments were deposited in river channels. 4.2. Geochemical characterisation of the aquifers 4.2.1. Groundwater chemical composition The groundwater samples are mostly of Ca–Mg–HCO3 and Ca–Na–HCO3 types, with HCO−3 as the dominant anion. − Reducing conditions with low SO2− 4 and NO3 coupled with high NH+4 and DOC typically characterise the aquifers associated with high dissolved concentrations of As. The groundwater from the reducing unit is characterised by high + 2− concentrations of As, PO3− 4 -tot, NH4, DOC, Fe and by low SO4 while the groundwater from the oxidised unit is characterised by high Mn, low NH+4, DOC, Fe, PO3− 4 -tot and As. For further details of the groundwater chemical composition of the study area, we refer to von Brömssen et al. (2007).

143

represented by the white and off-white sediments. Maximum contents of Fetotal and Mntotal were 32,800 and 300 mg/kg and minimum contents were 1580 and 14 mg/kg, respectively. Fetotal and Mntotal correlate very well (Fig. 7) and Fetotal/Mntotal are almost constant (average 112, std 35) for all samples. Also FeOx and MnOx correlate well for all sediments although FeOx/ MnOx is distinctively different between reducing (average 107, std 24) and oxidised sediments (average 35, std 19). AsOx and FeOx correlate (R = 0.95) for the reducing samples although the number of samples (n = 4) is low. The FeOxAA/FeOx ratio for the reducing unit is much lower (average ratio = 3) than for the oxidised unit (average ratio = 26) and in the oxidised unit FeOx is much lower than in the reducing unit, approximately 10 to 20 times. While performing the sequential extraction we noted that the colour of the red sediments changed during the OxAA extraction step from red to whitish. The blank concentration for As relative to the extraction concentration was 50% (std = 25) due to low amounts of As in respective fraction. For Fe and Mn the blank concentration relative to the extraction concentration was 2% and 1% respectively (std = 6% and 1% respectively).

4.2.2. Mineralogical investigations XRD analysis shows that the mineralogy of the sands is dominated by quartz, K-feldspar (orthoclase) and plagioclase (anorthite and albite) with a substantial content of biotite and ferro-hornblende in the reducing Holocene sediments. Magnetite and rutile were also identified in this sequence (Table 3). Biotite identified at the depth of 21 and 23 m corresponds to the dark bands found in the core samples (Fig. 4). The lower weathered and oxidised sediment sequence contains relatively lower quantities of feldspars. In this sequence neither biotite nor ferro-hornblende were identified, however both rutile and magnetite were found. Semi-quantitative analysis using XRF shows that the most abundant metal is Al followed by Fe, Ti, and Mn, respectively (Fig. 5). Iron content is approximately 65 times higher than Mn content (average value) for both the reducing and oxidised sediments. Relatively higher contents of Fe2O3 and MnO coincide with the reducing sediments, particularly the dark banded core samples at the depth of 21 m and 23 m, and the red samples from the depth of 40 and 44 m. Low CaO, MgO and P2O5 coincide with the samples from the oxidised sequence. The very red sample from 44 m depth is distinctively lower in Na2O, MgO and CaO. The SiO2 content is relatively lower in the reducing sequence, ranging from 63 to 82 wt.%. 4.2.3. Sequential extractions The sequential extraction shows that Astotal contents are low, maximum total content is 3500 µg/kg and was found in the reducing sequence at the depth of 23 m (Fig. 6). AsDIW and AsNaHCO3 contents were all below detection limit (bdl: b130 µg/kg), AsNaAc contents ranged from bdl to 400 µg/kg, AsOx from bdl to 1060 µg/kg. Highest AsNaAc as well as AsOx values coincide with reducing sample from 23 m depth, but except that sample AsNaAc and AsOx contents were of the same magnitude in both sequences. The highest content of As (1900 µg/kg) was found in the OxAA-fraction of reducing sample 21 and AsoxAA fraction holds most of the As in 50% of the samples. The concentrations of extractable Fe and Mn in all fractions were high in the reducing sequence. The corresponding concentrations were low for the oxidised sediments

Fig. 5. Major element composition (plotted as oxide wt.%) of the sediments representing the two distinct sedimentary units based on the XRF analyses. a) Fe, Mn, Al and Ti; b) Si, Ca, K, Na, Mg and P.

144 M. von Brömssen et al. / Journal of Contaminant Hydrology 99 (2008) 137–149 Fig. 6. Depth-wise distribution of a) As, b) Fe, and c) Mn, based on the results of the sequential extraction of the aquifer sediments. Please note that the DIW and NaHCO3 fractions are not shown as they were very low compared to the other fraction.

M. von Brömssen et al. / Journal of Contaminant Hydrology 99 (2008) 137–149

145

Fig. 7. Correlations between selected species and sequential extraction fractions, a) Mntotal vs Fetotal, b) MnOx vs FeOx, c) Astotal vs Fetotal, d) AsOx vs FeOx.

4.2.4. Geochemical modelling The sensitivity analysis of the simulations of SI values shows that mineral phases including Fe(III), e.g. Fe(III)-oxyhydroxides, are sensitive to an alteration of Eh. The other mineral phases discussed in this paper, e.g. siderite and vivianite are not sensitive to the alteration of Eh. SI calculations indicate that groundwater in both reducing (black samples) and oxidising aquifers (white, off-white and red samples) is near saturation with respect to Fe(III)-oxyhydroxides. When performing the sensitivity analysis with lower Eh values, ground water becomes unsaturated with respect to Fe (III)-oxyhydroxides. Ground water in the reducing unit is slightly supersaturated with respect to siderite (FeCO3) and vivianite (Fe3(PO4)2·8(H2O)) (SImax ~ 1 and 3, respectively) but undersaturated in the oxidised unit. SI values for siderite and vivianite correlate well with concentration of dissolved Fe. Groundwater in both aquifer units is near saturation with respect to rhodochrosite (MnCO3) and supersaturated with respect to MnHPO4. Groundwater from the reducing unit is near saturation with respect to hydroxyapatite (Ca5(PO4)3(OH)), but it is undersaturated with respect to the mineral phase in the oxidised unit (Fig. 8).

The surface complexation modelling of AsIII adsorption in the reducing aquifer unit suggests that AsIII adsorption is mainly controlled by the amount of Hfo in solid phase. pH is less important within the pH-interval found in the aquifer units of the study area. Only if PO3− 4 -tot is included as a competing ion in the simulations they result in reasonable concentrations of dissolved and adsorbed AsIII. If PO3− 4 -tot is excluded all AsIII would be adsorbed onto the Hfo. According to the modelling results H2PO−4 is the major P specie with an 3− activity 2 to 4 times higher than HPO2− 4 . PO4 is a minor species only. Fig. 9 shows the simulated dissolved concentrations of AsIII vs dissolved PO3− 4 -tot concentrations. In the diagrams actual groundwater samples are overlain (grey dots) and a correlation between As and PO3− 4 -tot is observed for the reduced aquifer unit, the R2 value is 0.4826. 5. Discussion The lithofacies and genesis of the reddish oxidised sediments in Matlab Upazila seems to correlate well with sediments reported from other areas in Bangladesh and West-

146

M. von Brömssen et al. / Journal of Contaminant Hydrology 99 (2008) 137–149

Fig. 8. a) Bivariate plots showing the dependency of the modelled SI values for a) Fe(OH)3(a) vs Fe, b) siderite vs Fe, c) vivianite vs Fe, d) hydroxyapatite vs PO3− 4 , e) MnHPO4 vs Mn and f) rhodochrosite vs Mn.

Bengal India. Oxidised reddish (orange/brown) sediments have been reported from various parts of West-Bengal, India (Pal and Mukherjee, in press; McArthur et al., 2004) and Bangladesh (Umitsu, 1993; Goodbred and Kuehl, 2000; van Geen et al., 2004; Stollenwerk et al., 2007; Hasan et al., in press). Further characterisation of these oxidised As-low aquifers should be given high priority as this appears to be a viable mitigation option. For drinking water management purposes a common classification system of these shallow aquifers should be established so that the different studies can be systematically organized. Even though the reported aquifer sequences appear to be very heterogeneous, similar sediments have been found in nature elsewhere under similar geological settings. This implies the knowledge of the sedimentary sequences is important for the local drillers for the identification of the safe sediments for the installation of safe wells. The elevated aqueous concentrations of Fe and Mn are expected to be a product of reductive dissolution of Fe(III)-

oxyhydroxides and reduction of solid phase Mn(IV)-oxides as expressed by the simplified equations (for a detailed evaluation of these processes we refer to Mukherjee et al., 2008-this issue): • 8Fe(OH)3(s) + CH3OO− + 21H+ = 8Fe2+ + CO2 + 24H2O • 2MnO2(s) + CH3OO− + 5H+ = 2Mn2+ + 4H2O + CO2 As distinctively higher contents of amorphous Fe and Mn (FeOx and MnOx) are found in the reduced unit and redoxstatus is sufficiently low for iron-reduction, high dissolved concentrations of both these elements would be expected in groundwater from the reducing aquifer unit. However, dissolved Fe concentration is relatively higher in groundwater of the reduced unit while Mn concentration is relatively higher in the oxidised unit. As both rhodochrosite and MnHPO4 are near saturation and/or slightly supersaturated in both units, these two minerals may exert an important role for controlling dissolved Mn concentrations, in addition to redox processes. The SI values for MnHPO4 are approximately 2 and, thus, these calculations may represent a crystalline

M. von Brömssen et al. / Journal of Contaminant Hydrology 99 (2008) 137–149

147

Fig. 9. Simulation of adsorbed and dissolved AsIII for the system with hydroferric oxides (Hfo), AsIII and PO3− 4 using the Diffuse Layer Model. Figure a) shows results with maximum amount As found in the sediments and b) with minimum amount.

phase while a non- or semi-crystalline phase may be present in the aquifer. For dissolved Fe, siderite, vivianite, and/or Fe (III)-oxyhydroxides would be the minerals that are likely to control dissolved Fe concentration in the reducing unit, in addition to redox processes. In the oxidised unit Fe seems to be more crystalline and less accessible to reductive dissolution. It appears that As is bound to more crystalline phases in the oxidised zone as the fractions of amorphous phases of Fe are low. Both oxalate and NaAc extractable fractions of Fe are low for the oxidised samples. As we found that both amorphous Fe-oxyhydroxides (FeOx) and solid phase Mn(IV) oxides (MnOx) contents are higher (approximately 10 to 20 times) in the reduced unit dissolved As concentration should be lower here than in the oxidised unit assuming i) that Fe-oxyhydroxides are the main sorbent, ii) the same species of As are present, and that iii) pH, the available amount of As and concentrations of competing ions are comparable. The species of As were only known for the dissolved phase in the reduced unit (98% was found to be in the form of AsIII; von Brömssen et al. 2007) and thus AsIII was assumed to be the adsorbed species as well. Modelling adsorption of AsIII shows that competing ion(s) was needed to explain the high dissolved AsIII concentrations or else all AsIII was adsorbed onto the Hfo. Including fixed dissolved concentration of PO3− 4 -tot in the model gave reasonable activities

within the range of AsIII concentrations found in the reducing aquifer unit demonstrating that high PO3− 4 -tot may be responsible for mobilization of As, in addition to the reductive dissolution of Fe(III)-oxyhydroxides. Fixed dissolved concentra3− tions of PO3− 4 -tot could be used as a condition for PO4 -tot in the 3− model, assuming that PO4 -tot concentrations are regulated by equilibriums with other secondary mineral phases such as hydroxyapatite and vivianite. Furthermore, a correlation between dissolved As and PO3− 4 -tot was also found and, thus, based on the surface complexation modelling results we speculate that PO3− 4 -tot and/or other competing ions are responsible for the high dissolved As concentrations, in addition to the reductive dissolution of Fe(III)-oxyhydroxides. Validation of the results of surface complexation modelling has not been done, however, we suggest that the role of PO3− 4 -tot for mobilizing As should be further studied. 6. Conclusions The results presented here describe the lithofacies and genesis of the sediments in the shallow aquifer and establish the relationship between aqueous and solid phase geochemistry. The findings from these investigations are consistent with the results reported earlier from the region (von Brömssen

148

M. von Brömssen et al. / Journal of Contaminant Hydrology 99 (2008) 137–149

et al. 2007). The findings also validate the approach of installation of new safe wells on the basis of sediment colour as used by the local drillers. However, the colour description needs to be standardised using scientifically acceptable charts improvising the colour codes following the Munsell Colour Chart (Munsell Color, 2000). Locally adoptable colour charts may be produced and provided to the drillers for use in the field. Groundwater in the contaminated reducing aquifers is + characterised by high concentrations of PO3− 4 -tot, NH4, DOC, Fe 2− and low SO4 ,whereas the targeted oxidised low-As aquifers are characterised by high Mn, low NH+4, DOC, Fe, and PO3− 4 -tot. Sediment chemistry of the high-and low-As aquifers show that Fetotal and Mntotal correlates well for both the high- and the low-As unit. However, the FeOx/MnOx (oxalate extraction) is distinctively higher for the reducing high-As unit indicating that amorphous Fe oxides and hydroxides are more inclined to weathering and oxidation than amorphous Mn oxides and hydroxides. Geochemical modelling indicates that the concentrations of Fe, Mn and PO3− 4 -tot in groundwater are also influenced by the formation of secondary minerals in addition to redox processes. Simulation of AsIII adsorption within the reducing aquifer unit system showed that AsIII was largely influenced by the amount of Hfo, pH and by competing ion(s). In addition to reductive dissolution of Fe(III)-oxyhydroxides we speculate that high PO3− 4 -tot concentrations are responsible for mobilization of As as simulation using the Diffuse Layer Model showed a good fit with the observed groundwater and adsorbed concentrations of AsIII. The study confirms high mobility of dissolved As concentration in reducing aquifers with typically dark colour of sediments found in previous studies and, thus, validates the approach for location of wells used by local drillers, which is based on sediment colour that hold promise for future implementation as a sustainable arsenic mitigation. The findings presented here have implications for targeting low-As aquifers that prompt further studies of the sustainability of As-low aquifers of Matlab Upazila, SE-Bangladesh. Further characterisation of the oxidised As-low aquifers should be done as this appears to be a viable mitigation option. For drinking water management purposes a common classification system for the aquifers and colour description of the sediments should be established. Even though reported aquifer sequences are heterogeneous, local drillers may target the As-low aquifer units and thus they play a central part for implementation of the proposed mitigation option. Acknowledgements The work was supported by the research grants from the Swedish International Development Agency (Sida-SAREC) dnr:SWE-2002-129, Swedish Research Council (VR-Sida) dnr: 348-2003-4963 and the Strategic Environmental Research Foundation (MISTRA) dnr: 2005-035-137. The authors thank BRAC and NGO Forum for Drinking Water Supply and Sanitation, Bangladesh for the facilities at Matlab and to Karim and Alec for assistance during the fieldwork. Ann Fylkner, Magnus Mörth, Monica Löwén, Marianne Ahlbom helped us with the analytical work. We acknowledge Sigrun Santesson at the KTH-International Office for providing our students (SLH and AB) with Minor Field Study (MFS) grant as

well as Giesela at Svalorna in Bangladesh for local hospitality during the fieldwork. References Ahmed, K.M., Bhattacharya, P., Hasan, M.A., Akhter, S.H., Alam, S.M.M., Bhuyian, M.A.H., Imam, M.B., Khan, A.A., Sracek, O., 2004. Arsenic enrichment in groundwater of the alluvial aquifers in Bangladesh: an overview. Applied Geochemistry 19 (2), 181–200. Allison, J.D.D., Brown, S., Novo-Gradac, K.J., 1991. MINTEQA2, a Geochemical Assessment Data Base and Test Cases for Environmental Systems. EPA, Athens, GA, U.S. BGS and DPHE, 2001. Arsenic Contamination of Groundwater in Bangladesh. British Geological Survey. WC/00/19. Bhattacharya, P., Chatterjee, D., Jacks, G., 1997. Occurrence of arseniccontaminated groundwater in alluvial aquifers from Delta Plains, Eastern India: options for safe drinking water supply. Water Resources Development 13 (1), 79–92. Bhattacharya, P., Frisbie, S.H., Smith, E., Naidu, R., Jacks, G., Sarkar, B., 2002a. Arsenic in the environment: a global perspective. In: Sarkar, B. (Ed.), Handbook of Heavy Metals in the Environment. Marcell Dekker Inc., New York, New York, USA, pp. 147–215. Bhattacharya, P., Jacks, G., Ahmed, K.M., Khan, A.A., Routh, J., 2002b. Arsenic in groundwater of the Bengal Delta Plain aquifers in Bangladesh. Bulletin of Environmental Contaminants and Toxicology 69, 538–545. Bhattacharya, P., et al., 2006a. Distribution and mobility of arsenic in the Rio Dulce alluvial aquifers in Santiago del Estero Province, Argentina. Science of The Total Environment 358 (1–3), 97–120. Bhattacharya, P., Ahmed, K.M., Hasan, M.A., Broms, S., Fogelström, J., Jacks, G., Sracek, O., von Brömssen, M., Routh, J., 2006b. Mobility of arsenic in groundwater in a part of Brahmanbaria district, NE Bangladesh. In: Naidu, R., Smith, E., Owens, G., Bhattacharya, P., Nadebaum, P. (Eds.), Managing Arsenic in the Environment: From Soil to Human Health. CSIRO Publishing, Melbourne, Australia, pp. 95–115. ISBN: 0643068686. Bhattacharya, P., Welch, A.H., Stollenwerk, K.G., McLaughlin, M.J., Bundschuh, J., Panaullah, G., 2007. Arsenic in the environment: biology and chemistry. Science of the Total Environment 379, 109–120. doi:10.1016/j. scitotenv.2007.02.037. Bhattacharya, P., von Brömssen, M., Hasan, M.A., Jacks, G., Ahmed, K.M., Sracek, O., Jakariya, M., Huq, S.M.I., Naidu, R., Smith, E. & Owens, G. (2008) Arsenic mobilisation in the Holocene flood plains in South-central Bangladesh: evidences from the hydrogeochemical trends and modeling results. In: Bhattacharya, P., Ramanathan, A.L., Mukherjee A.B., Bundschuh, J., Chandrasekharam, D. & Keshari, A.K. (eds.) Groundwater for Sustainable Development: Problems, Perspectives and Challenges. Taylor and Francis/ A. A. Balkema, The Netherlands: 283–299 (ISBN 978-0-415-40776-2). Chakraborti, D., Sengupta, M.K., Rahaman, M.M., Ahamed, S., Chowdhury, U.K., Hossain, M.A., Mukherjee, S.C., Pati, S., Saha, K.C., Dutta, R.N., Quamruzzaman, Q., 2004. Groundwater arsenic contamination and its health effects in the GangaMeghna-Brahmaputra plain. Journal of Environmental Monitoring 6 (6), 74N–83N. Dodd, J., Large, D.J., Fortey, N.J., Milodowski, A.E., Kemi, S., 2000. A petrographic investigation of two sequential extraction techniques applied to anaerobic canal bed mud. Environmental Geochemistry and Health 22, 281–296. Dzombak, D.A., Morel, F.M.M., 1990. Surface Complexation ModellingHydrous Ferric Oxide. John Wiley, New York, USA. 393 pp. Goodbred, S.L., Kuehl, S.A., 2000. The significance of large sediment supply, active tectonism, and eustasy on margin sequence development: late quaternary stratigraphy and evolution of the Ganges–Brahmaputra delta. Sedimentary Geology 133, 227–248. Goodbred, J., Steven, L., Kuehl, S.A., 2003. The production, transport, and accumulation of sediment: a cross-section of recent developments with an emphasis on climate effects. Sedimentary Geology 162 (1–2), 1–3. Gustafsson, J.P., Bhattacharya, P., 2007. Geochemical modelling of arsenic adsorption to oxide surfaces. In: Bhattacharya, P., Mukherjee, A.B., Bundschuh, J., Zevenhoven, R., Loeppert, R.H. (Eds.), Arsenic in Soil and Groundwater Environment: Biogeochemical Interactions, Health Effects and Remediation, Trace Metals and other Contaminants in the Environment, vol. 9. Elsevier B.V., Amsterdam, The Netherlands, pp. 153–200. doi:10.1016/S0927-5215(06)09006-1. Harvey, C.F., Swartz, C.H., Badruzzaman, A.B.M., Keon-Blute, N., Yu, W., Ali, M.A., Jay, J., Beckie, R., Niedan, V., Brabander, D., Oates, P.M., Ashfaque, K.N., Islam, S., Hemond, H.F., Ahmed, M.F., 2002. Arsenic mobility and groundwater extraction in Bangladesh. Science 298, 1602–1606. Hasan, M.A., von Brömssen, M., Bhattacharya, P., Ahmed, K.M., Sikder, A.M., Jacks, G., Sracek, O., in press. Geochemistry and mineralogy of shallow alluvial aquifers in Daudkandi upazila in the Meghna flood plain, Bangladesh. Environmental Geology. doi:10.1007/s00254-008-1319-8.

M. von Brömssen et al. / Journal of Contaminant Hydrology 99 (2008) 137–149 Jakariya, M., Bhattacharya, P., 2007. Use of GIS in local level participatory planning for arsenic mitigation: a case study from Matlab Upazila, Bangladesh. Journal of Environmental Sciences and Health, Part A Toxic/ Hazardous Substances and Environmental Engineering 42 (12), 1933–1944. Jakariya, M., von Brömssen, M., Jacks, G., Chowdhury, A.M.R., Ahmed, K.M., Bhattacharya, P., 2007. Searching for a sustainable arsenic mitigation strategy in Bangladesh: experience from two upazilas. International Journal of Environment and Pollution 31 (3/4), 415–430. Kapaj, S., Peterson, H., Liber, K., Bhattacharya, P., 2006. Human health effects from chronic arsenic poisoning – A Review. J. Environ. Sci. Health, Part A., 41 (10), 2399–2428. doi:10.1080/10934520600873571. McArthur, J.M., Banerjee, D.M., Hudson-Edwards, K.A., Mishra, R., Purohit, R., Ravenscroft, P., Cronin, A., Howarth, R.J., Chatterjee, A., Talukder, T., Lowry, D., Houghton, S., Chadha, D.K., 2004. Natural organic matter in sedimentary basins and its relation to arsenic in anoxic ground water: the example of West Bengal and its worldwide implications. Applied Geochemistry 19 (8), 1255–1293. Mukherjee, A., von Brömssen, M., Scanlon, B.R., Bhattacharya, P., Fryar, A.E., Hasan, M.A., Ahmed, K.M., Chatterjee, D., Jacks, G., Sracek, O., 2008. Hydrogeochemical comparison and effects of overlapping redox zones on groundwater arsenic near the Western (Bhagirathi sub-basin, India) and Eastern (Meghna sub-basin, Bangladesh) margins of the Bengal Basin. Journal of Contaminant Hydrology 99, 31–48 (this issue). doi:10.1016/j.jconhyd.2007.10.005. Munsell Color, 2000. Munsell® Soil Colour Charts. Year 2000 Revised Washable Edition. Munsell Color, X-rite, MI, USA. Nickson, R., McArthur, J., Burgess, W., Ahmed, K.M., Ravenscroft, P., Rahman, M., 1998. Arsenic poisoning of Bangladesh groundwater. Nature 395, 338. Nickson, R.T., McArthur, J.M., Ravenscroft, P., Burgess, W.G., Ahmed, K.M., 2000. Mechanism of arsenic release to groundwater, Bangladesh and West Bengal. Applied Geochemistry 15 (4), 403–413. Pal, T., Mukherjee, K., in press. Study of subsurface geology in locating arsenic-free groundwater in Bengal delta, West Bengal India. Environmental Geology. doi:10.1007/s00254-008-1221-4. Parkhurst, D.L., Appelo, C.A.J., 1999. Users's Guide to PHREEQC (Version 2) — A Computer Program for Speciation, Batch-reaction, One-dimensional Transport and Inverse Geochemical Calculations. U.S. GEOLOGICAL SURVEY, Denver, Colorado, USA. 99-4259. Ravenscroft, P., 2001. Distribution of groundwater arsenic in Bangladesh related to geology. In: Jacks, G., Bhattacharya, P., Khan, A.A. (Eds.), Groundwater Arsenic Contamination in the Bangal Delta Plain of Bangladeh. KTH, Dhaka, p. 122. SIS, 1996. Water Quality. Determination of Alkalinity—Part2: Determination of Carbonate Alkalinity (SS-EN ISO 9963-2), 10p. Smedley, P.L., Kinniburgh, D.G., 2002. A review of the source, behaviour and distribution of arsenic in natural waters. Applied Geochemistry 17 (5), 517–568.

149

Smith, A.H., Lingas, E.O., Rahman, M., 2000. Contamination of drinking-water by arsenic in Bangladesh: a public health emergency. Bulletin of the World Health Organization 78 (9), 1093–1103. Sracek, O., Bhattacharya, P., Jacks, G., Gustafsson, J.-P., Bromssen, M.v., 2004. Behavior of arsenic and geochemical modeling of arsenic enrichment in aqueous environments. Applied Geochemistry 19 (2), 169–180. Stollenwerk, K.G., et al., 2007. Arsenic attenuation by oxidized aquifer sediments in Bangladesh. Science of The Total Environment 379 (2–3), 133–150. Stuben, D., Berner, Z., Chandrasekharam, D., Karmakar, J., 2003. Arsenic enrichment in groundwater of West Bengal, India: geochemical evidence for mobilization of As under reducing conditions. Applied Geochemistry 18 (9), 1417–1434. Swartz, C.H., Blute, N.K., Bodruzzaman, B., Ali, M.A., Brabander, D., Jay, J., Besabcon, J., Islam, S., Hemond, H.F., Harvey, C., 2004. Mobility of arsenic in a Bangladesh aquifer: inferences from geochemical profiles, leaching data, and mineralogical characterization. Geochimica et Cosmochimica Acta 68 (22), 4539–4557. Umitsu, M., 1987. Late Quaternary sedimentary environment and landform evolution in the Bengal Lowland. Geographical Review of Japan B 60,164–178. Umitsu, M., 1993. Late Quaternary sedimentary environments and landforms in the Ganges Delta. Sedimentary Geology 83, 177–186. van Geen, A., Ahasan, H., Horneman, A.H., Dhar, R.K., Zheng, Y., Hussain, I., Ahmed, K.M., Gelman, A., Stute, M., Simpson, H.J., Wallace, S., Small, C., Parvez, F., Slakovich, V., Lolacono, N.J., Becker, M., Cheng, Z., Momtaz, H., Shahnewaz, M., Seddique, A.A., Graziano, J., 2002. Promotion of wellswitching to mitigate the current arsenic crisis in Bangladesh. Bulletin of the World Health Organization 80 (9), 732–737. van Geen, A., Rose, J., Thoral, S., Garnier, J.M., Zheng, Y., Bottero, J.Y., 2004. Decoupling of As and Fe release to Bangladesh groundwater under reducing conditions. Part II: evidence from sediment incubations. Geochimica et Cosmochimica Acta 68 (17), 3475–3486. von Brömssen, M. et al., 2005. Targeting safe aquifers in regions with elevated arsenic in groundwater of Matlab Upazila, SE Bangladesh, 8th ICOBTE, Adelaide, Australia. von Brömssen, M., Jakariya, M., Bhattacharya, P., Ahmed, K.M., Hasan, M.A., Sracek, O., Jonsson, L., Lundell, L., Jacks, G., 2007. Targeting low-arsenic in Matlab Upazila, Southeastern Bangladesh. 2007. Targeting low-arsenic aquifers in Matlab Upazila, Southeastern Bangladesh. Science of The Total Environment 379 (2–3), 121–132. Wenzel, W.W., Kirchbaumer, N., Prohaska, T., Stingeder, G., Lombi, E., Adriano, D.C., 2001. Arsenic fractionation in soils using an improved sequential extraction procedure. Analytica Chimica Acta 436 (2), 309–323. WHO, 2004. Guidelines for drinking-water quality, Third ed. Recommendations, vol. 1. World Health Organisation, Geneva. 515p.