Science of the Total Environment 485–486 (2014) 12–22
Contents lists available at ScienceDirect
Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv
Shallow hydrostratigraphy in an arsenic affected region of Bengal Basin: Implication for targeting safe aquifers for drinking water supply Ashis Biswas a,b,⁎, Prosun Bhattacharya a, Abhijit Mukherjee c, Bibhash Nath d, Helena Alexanderson e, Amit K. Kundu b, Debashis Chatterjee b, Gunnar Jacks a a KTH-International Groundwater Arsenic Research Group, Division of Land and Water Resources Engineering, Department of Sustainable Development, Environmental Science and Engineering, KTH Royal Institute of Technology, Teknikringen 76, SE-100 44 Stockholm, Sweden b Department of Chemistry, University of Kalyani, Kalyani 741235, West Bengal, India c Department of Geology and Geophysics, Indian Institute of Technology-Kharagpur, Kharagpur 721302, West Bengal, India d School of Geosciences, The University of Sydney, Sydney, NSW 2006, Australia e Department of Geology, Lund University, Sölvegatan 12, SE-223 62 Lund, Sweden
H I G H L I G H T S • • • • •
Shallow hydrostratigraphic framework over an area of 100 km2 is investigated. Two types of aquifer viz. brown and grey sand aquifer (BSA and GSA) are identified. Concentration of As and Fe is high (N 10 µg/L) at GSA, while low at BSA. Despite low concentration of As, the concentration of Mn is high (N 400 µg/L) at BSA. BSA should be targeted only when no other As safe drinking source is available.
a r t i c l e
i n f o
Article history: Received 16 January 2014 Received in revised form 7 March 2014 Accepted 7 March 2014 Available online 1 April 2014 Editor: D. Barcelo Keywords: Bengal Basin Shallow hydrostratigraphy Groundwater Drinking water supply Arsenic Manganese
a b s t r a c t To delineate arsenic (As) safe aquifer(s) within shallow depth, the present study has investigated the shallow hydrostratigraphic framework over an area of 100 km2 at Chakdaha Block of Nadia District, West Bengal. Drilling of 29 boreholes and subsequent hydrostratigraphic modeling has identified three types of aquifer within 50 m below ground level (bgl). Aquifer-1 represents a thick paleochannel sequence, deposited parallel to the River Hooghly and Ichamati. Aquifer-2 is formed locally within the overbank deposits in the central floodplain area and its vertical extension is strictly limited to 25 m bgl. Aquifer-3 is distributed underneath the overbank deposits and represents an interfluvial aquifer of the area. Aquifer-3 is of Pleistocene age (~70 ka), while aquifer-1 and 2 represent the Holocene deposits (age b9.51 ka), indicating that there was a major hiatus in the sediment deposition after depositing the aquifer-3. Over the area, aquifer-3 is markedly separated from the overlying Holocene deposits by successive upward sequences of brown and olive to pale blue impervious clay layers. The groundwater quality is very much similar in aquifer-1 and 2, where the concentration of As and Fe very commonly exceeds 10 μg/L and 5 mg/L, respectively. Based on similar sediment color, these two aquifers have jointly been designated as the gray sand aquifer (GSA), which constitutes 40% (1.84 × 109 m3) of the total drilled volume (4.65 × 109 m3). In aquifer-3, the concentration of As and Fe is very low, mostly b2 μg/L and 1 mg/L, respectively. This aquifer has been designated as the brown sand aquifer (BSA) according to color of the aquifer materials and represents 10% (4.8 × 108 m3) of the total drilled volume. This study further documents that though the concentration of As is very low at BSA, the concentration of Mn often exceeds the drinking water guidelines. © 2014 Elsevier B.V. All rights reserved.
1. Introduction
⁎ Corresponding author at: Department of Geological Sciences, University of Saskatchewan, 114 Science Place, Saskatoon, SK S7N 5E2, Canada. Tel.: +1 306 966 5764; fax: +1 306 966 8593. E-mail addresses: [email protected], [email protected] (A. Biswas).
http://dx.doi.org/10.1016/j.scitotenv.2014.03.045 0048-9697/© 2014 Elsevier B.V. All rights reserved.
The Bengal Basin, located at the head of Bay of Bengal, is the world's largest fluvio-deltaic system (Alam et al., 2003; Coleman, 1981; Mukherjee et al., 2009; Neidhardt et al., 2013). It is being drained by the River Ganges, Brahmaputra and Meghna and their tributaries and distributaries, also known as the Ganges–Brahmaputra–Meghna (GBM)
A. Biswas et al. / Science of the Total Environment 485–486 (2014) 12–22
basin (Mukherjee et al., 2007). It covers an area of ~200,000 km2 (most of Bangladesh and neighboring part of West Bengal, currently Paschim Banga) and hosts more than 2% (120 millions) of the world's population (Mukherjee et al., 2007). As people living in this region mostly relies on groundwater for domestic and agricultural purpose, the occurrence of elevated dissolved arsenic (As) in groundwater has put millions of people at the risk of chronic As toxicity (Chakraborti et al., 2004; Fendorf et al., 2010; Mukherjee and Bhattacharya, 2001). The extent of severity has been termed as the world's largest mass poisoning in human history (Smith et al., 2000). Since the first reporting of As in the aquifers of West Bengal in 1978 (Saha, 1984), numerous investigations have been carried out in different parts of the Bengal Basin (mostly in Bangladesh) to characterize source, regulating processes and distribution of As in the aquifers and development of possible mitigation alternatives (Bhattacharya et al., 1997; Charlet and Polya, 2006; Fendorf et al., 2010; Smedley and Kinniburgh, 2002; Winkel et al., 2008). It is revealed that the occurrence of high dissolved As in groundwater of this region is mostly confined to shallow aquifers (depth b 60 m) of Holocene age (BGS and DPHE, 2001; Horneman et al., 2004; Ravenscroft et al., 2005). On the contrary, the deep aquifers (depth N150 m) of Pleistocene age are mostly safe (b10 μg/L) (Burgess et al., 2010; Michael and Voss, 2008). However, because of high installation costs, the construction of tubewells (TWs) at the deep aquifers for domestic purpose is hardly affordable by rural villagers and it is argued that the identification of safe aquifer(s) within the shallow depth is necessary to improve the As mitigation scenario in rural Bengal (Biswas et al., 2012a; Hug et al., 2011; van Geen et al., 2007; von Brömssen et al., 2007). In recent years, one of the most highlighted aspects of As contamination in groundwater is its spatial distribution within the shallow aquifers. The concentration of dissolved As in the aquifers varies spatially (from unsafe to safe and vice versa) within a scale of tens of meters (van Geen et al., 2002, 2003). In the last few years numerous studies have been undertaken to explore the sedimentological control on the spatial distribution of dissolved As in shallow aquifers. The important finding is that the As contamination of groundwater is restricted to the gray sand aquifer and the reddish-brown sand aquifer is very rarely contaminated (Biswas et al., 2012a; Bundschuh et al., 2010; Pal and Mukherjee, 2008, 2009; McArthur et al., 2008, 2011; von Brömssen
Nepal
et al., 2007). Consequently, the reddish-brown sand aquifer has been proposed as an alternative drinking water source, which can be targeted by locally available low-cost drilling technology (McArthur et al., 2011; von Brömssen et al., 2007). However, before advocating for targeting this aquifer for safe drinking water supply it is necessary to investigate its regional distribution and subsurface geologic architecture at different geographic locations of the basin. In the present study, we have undertaken a detailed subsurface lithological investigation to explore the shallow hydrostratigraphic framework in an As-affected region of the central western Bengal Basin. In order to extend the current understanding of basin fill history and explain the evolution of the hydrostratigraphic framework in the study area, dating of different litho-units has been conducted and local stratigraphic units have been correlated to the regional late Quaternary stratigraphic units reported from other parts of the Bengal Basin. This study has further documented the distribution of dissolved As, Fe and Mn in different aquifer types to justify the potentiality of targeting safe aquifer within the shallow depth in As affected regions of the Bengal Basin.
2. Materials and methods 2.1. Study area The study area is located at the Chakdaha Block of Nadia District, West Bengal, 60 km north of Kolkata city and approximately 170 km inland from the present coastline of Bay of Bengal (Fig. 1). Geographically the study area represents the stable shelf part of the central western Bengal Basin (Fig. 1) (Mukherjee et al., 2009). The River Hooghly (distributaries of River Ganges) and Ichamati bound the study area in the west and east, respectively. The area is very flat; elevation varies within ~0–20 m above mean sea level (Neidhardt et al., 2013). Throughout the area prominent geomorphological features are abandoned channels, meandering scars, oxbow lakes and ponds (Fig. 1), which highlight the frequent changes in the channel courses during the recent past (Nath et al., 2005; Neidhardt et al., 2013). Hydrogeologically, the aquifers of the study area represent part of the “Sonar Bangla Aquifer” (Mukherjee et al., 2007).
a
Himalayan Foothills Bhutan
b BH-18
India
26°N
Investigation area 22°N
Indo-Burmese Fold Belt
Latitude
24°N
Bengal Basin
Bangladesh
BH-21
(A)
Dauki Fault India West Bengal
13
BH-24
BH-23
BH-17
(B)
Chakdaha BH-15 RS BH-2
(C)
BH-26
(A')
BH-1 BH-28 BH-4
(B')
BH-29 BH-22
BH-27
BH-7
BH-20 BH-25
BH-3
BH-13
(D)
(C')
BH-14 BH-10
BH-12
(E)
BH-19
BH-16
(D')
BH-11 BH-5
BH-8
BH-9 BH-6
Bay of Bengal
(E')
4.75 km 100 km
88°E
90°E
92°E
Longitude
Borehole Location
Fig. 1. Study area maps showing a. physiographic settings of Bengal Basin and location of Chakdaha Block of Nadia District, West Bengal, India in Bengal Basin (adopted from Neidhardt et al., 2013). b. Location of 29 boreholes and 6 west–east traverses along which lithologic cross-section has been prepared (showing in Fig. 2) over the study area. The satellite images were acquired from Google Earth 6.0.2.
14
A. Biswas et al. / Science of the Total Environment 485–486 (2014) 12–22
2.2. Subsurface sedimentological investigation In order to investigate the distribution of different aquifers and aquitards within the shallow depth, 29 boreholes with almost equal spacing (Fig. 1) were drilled over 100 km2 area during the period of March to May 2011, by locally available hand suction drilling technology (Ali, 2003), which usually can reach up to 50 m below ground level (bgl). The washed sediment samples were collected from each 1.5 m (5 ft) interval. The sampling frequency was higher when a visible change in the lithology and color was noticed during drilling. The colors (black, gray, olive to pale blue and brown) of the sediments were determined just after recovery, prior to atmospheric oxidation. Sediments were classified into the three major lithotypes of clay, sandy clay and sand. The sediment of the lithotype clay and sand represents aquitard and aquifer, respectively, while sandy clay represents aquitard with higher permeability. Based on these lithotypes and colors, lithologs for the boreholes were prepared (Appendix Fig. A.1). Using the 29 lithologs, a 3D lithologic model for the area was developed using the software RockWorks ver. 15 (RockWare, Golden, CO, USA). The lithological data of each borehole were interpolated in three dimensions by the algorithm of “lithoblending”, where each voxel is assigned a lithology value corresponding to a particular lithotype of closest known data point and then the processes continue both horizontally and vertically to the next voxel and so on, until a voxel with a specified lithology value (that is next known data point) is encountered. The model was optimized by changing the node spacing and the final model was selected that showed least sensitivity towards the change in grid size (Mukherjee et al., 2007). The final model has the resolution of 100 m (X) × 100 m (Y) × 0.5 m (Z), which results in a node spacing of 99 (X) × 95 (Y) × 99 (Z), with the total number of 931,095 solid grids, each having the volume of 5000 m 3. Finally, using this 3D lithologic model, 6 west–east 2D lithologic cross-sections and a number of plan-view maps that represent the distribution of aquifers and aquitards over the study area at a specific depth were developed (Mukherjee et al., 2007). 2.3. Dating of the aquifer sediments A total of 4 (borehole: BH-13 depth: 15.2 m and 25.8 m bgl; borehole: BH-2 depth: 10.6 m bgl and borehole: BH-6 depth: 34.8 m bgl) organic rich peaty sediments were collected during drilling for 14 C age dating. The samples were collected in airtight polyethylene zipper bags and preserved at − 18 °C to minimize the bacterial activity, prior to sending for analysis. The samples were analyzed within six months of collection at the Lund University Radiocarbon Dating Laboratory by Single Stage Accelerator Mass Spectrometer (SSAMS) (National Electrostatic Corp. (NEC), Wisconsin, USA). The precision of the analysis was ±50 14C years. The 14C ages were calibrated with reference to the atmospheric data from Reimer et al. (2009) by the software OxCal v3.10 (Bronk Ramsey and Lee, 2013). The range of calibrated ages was reduced to the point estimate by taking the midpoint of two terminal ages that limit the 68.2% (one standard deviation) range of the distribution and taking the half of the difference between this midpoint and one terminal age as the single standard deviation of estimate (Blaauw and Heegaard, 2012). A new drilling was conducted in April 2013, at just 1 m apart from the old drilled hole of BH-13 for the coring of OSL dating. Two samples were collected at the depths of 37.3 m bgl (BH-13_OSL_1) and 43.6 m bgl (BH-13_OSL_2). The targeted depth was reached by normal hand suction drilling technology (Ali, 2003), then a PVC tube (length: 30 cm, thickness: 2 mm) was fitted at the top of the drilling pipe and hammered down to collect the samples. After recovery, the core tubes were wrapped with multiple layers of aluminum foil and black colored insulating tape and preserved at complete darkness. The samples were
analyzed within two months of sample collection at the Lund Luminescence Laboratory, Lund University. Only the central part of each core tube was sampled for the analysis. Large single aliquots of 180–250 μm quartz grains in cups were analyzed in a Risø TL/OSL reader model DA-20. A Single Aliquot Regeneration (SAR) protocol was used (Murray and Wintle, 2000, 2003), with the respective pre-heat and cut-heat at 220 °C and 200 °C for sample BH-13_OSL_1 and 260 °C and 240 °C for sample BH-13_OSL_2, as determined by preheat plateau and dose recovery tests. The samples were optically stimulated by blue light sources (470 ± 30 nm; ~50 mW/cm2). Since the second sample suffered from feldspar contamination (mean IR/blue ratio N10%), it was additionally stimulated with infrared light (post-IR blue protocol; Banerjee et al., 2001). The detection was made through 7 mm of U340 glass filter (Bøtter-Jensen et al., 2000), the first 1.6 s of the signal was integrated for the peak and the following 1.6 s for the background. Aliquots were accepted if they had a test dose error of b 10%, a signal more than three times the background and a recycling ratio within 10% of unity (b 20% for sample BH-13_OSL_2). The dose recovery ratio of the two samples (0.96 ± 0.06, n = 6 and 0.99 ± 0.11, n = 4, respectively) showed that the selected protocols were able to correctly measure a given dose. The environmental dose rate was determined by gamma spectrometry and by estimating the contribution of cosmic radiation as described by Murray et al. (1987) and Prescott and Hutton (1994), respectively. Average water content was estimated from the values of natural and saturated water contents as measured on the sediment. 2.4. Groundwater sampling and analysis A total of 423 groundwater samples were collected for the determination of As, Fe and Mn, during March to May 2011 from the existing TWs, which were installed at the depth b 70 m bgl (Biswas et al., 2012b). After purging the TWs for few minutes groundwater sample was collected in a prewashed high density polyethylene vials (Tarson) and acidified on-site by HNO 3 (1% v/v, Suprapur, Merck). The concentration of As in the samples was determined by a hydride generation atomic absorption spectrophotometer (HG-AAS, detection limit b 1 μg/L), while the concentration of Fe and Mn was determined by a graphite furnace atomic absorption spectrophotometer (GF-AAS, GTA-120, detection limit 1.5 μg/L and 1.0 μg/L for Fe and Mn, respectively) (Varian, AA240). For further details of groundwater sampling, analysis and quality assurance reader is referred to Biswas et al. (2012b). This data set of 423 groundwater analyses was then compiled with other available data sets for As, Fe and Mn from the study area (e.g. data set of Biswas et al., 2012a). Out of the total number of TWs in the combined data set, 325 TWs were installed in the area enclosed by the drilled boreholes and had screen position within the maximum drilling depth (50 m) and thus, the lithologic information around the screen position was known. In order to assure true representativeness of a TW to a specific aquifer type, only these TWs have been used for subsequent statistical analysis to investigate the accuracy of the lithologic model and the distribution of As, Fe and Mn in different aquifer types of the study area. 3. Results 3.1. Accuracy of the lithologic model The accuracy of the lithologic model was tested by assuming that the screen of all 325 TWs was placed in the sand and the depth of TWs reported by the TW owner during sampling was correct. For the purpose of accuracy test, the depth of the TWs up to midpoint of the screen was considered. Our field experience indicates that the screen length of the TWs with depth b 30 m is usually 1.8 m, while for the TWs of depth N 30 m, it is 3.6 m in the study area. Consequently, in order to
A. Biswas et al. / Science of the Total Environment 485–486 (2014) 12–22
(A) BH-18
BH-24
BH-26
(A )
BH-19 0
0
BH-21
15
Aqu-1 -25
-25
Aqu-1 Aqu-3
0
-45
-45
Aqu-1
3 BH-28
BH-1
BH-4
-25
Aqu-1 Aqu-1
-45
-45
Aqu-3 3
0
0
BH-2
BH-22
BH-29
BH-7
-25
-45 0
Aqu-3 4.5
0
BH-14
BH-10
BH-16
-25 -45
-45
3.5
(D ) Depth (m bgl)
Aqu-1 Aqu-3 7.0
BH-8
BH-5
Brown sand
-25
Aqu-1 0
Brown clay
9.0
Aqu-2
(E) BH-12 BH-11
Grey sand
0
BH-3
Olive clay
-45
Aqu-1
(D )
-25
-45 -25
BH-16
0
Depth (m bgl)
BH-25
BH-20
Aqu-1
(D) BH-13
Brown sandy clay
8
BH-27
0
Grey clay -45
Aqu-3 4
(C) BH-15
Black clay
-25
Aqu-1
0
(C )
Grey sandy clay
Aqu-2 Aqu-1
Lithology Legend
6
0
(C) BH-15
(B ) Depth (m bgl)
Aqu-2
-25
Depth (m bgl)
0
BH-23
0
(B) BH-17
6
BH-9
BH-6
(E )
0
0
-45 -25
Aqu-1 0
5
-25 -45
Aqu-2
10
Length of the Traverse (km) Fig. 2. Six west–east lithologic cross-sections along the traverses shown in Fig. 1. Aqu-1, 2 and 3 represents aquifer-1, 2 and 3 respectively.
determine the depth of the TWs up to the midpoint of screen, 1 m and 2 m were subtracted from the total well depth of b 30 m and N30 m, respectively. In the next step, the TWs at each specific depth were grouped together and placed on the lithologic plan-view map of the corresponding depth to test whether the screen position of TWs was located at the sand units. The accuracy analysis reveals that according to the proposed lithologic model 277, 11 and 37 TWs are placed in sand, sandy clay and clay respectively, which indicates 85% accuracy of the model.
3.2. Shallow hydrostratigraphic framework in the study area Six 2D west–east lithologic cross-sections over the study area have been displayed in Fig. 2. Additionally, a few plan-view maps representing the distribution of aquifer and aquitard at different specific depths over the study area have been presented in Appendix Fig. A.2. These cross sections and plan view maps represent the distribution of a very complex aquifer–aquitard framework within the drilling depth. The surface sandy clay layer extends over the entire study area and the thickness
16
A. Biswas et al. / Science of the Total Environment 485–486 (2014) 12–22
a
c
b
d
f
h
e
g
i
j
k
Fig. 3. Photographs of few representative sediment samples collected during drilling over the study area. Hand written letters represent depth in feet from where samples were collected. (a) BH-17: represents typical channel deposits, where a thick layer of sand (aquifer-1) is extended from just below of surface sandy clay unit to the depth, where drilling was stopped, (b) BH-22: represents interfluvial sequence, where aquifer-2 is distributed between 50 and 70 ft within overbank deposits, below aquifer-3 is distributed and separated from the overbank deposits by successive sequences of olive to pale blue and brown clay layers, (c) BH-17 (depth: 3.03 m bgl): light brown surface sandy clay layer with mottles of Fe and Mn oxyhydroxides, (d) BH-2 (10.6 m bgl): black peaty materials dated as 2.99 ± 0.04 ka, (e) BH-13 (25.8 m bgl): black peaty materials dated as 9.51 ± 0.02 ka, (f) BH-4 (24.5 m bgl): black peaty materials with shell fragments, (g) BH-4 (27.3 m bgl): typical dark gray colored sand found in aquifer-1 and 2, (h) BH-22 (28.9 m bgl): pale blue colored clay layer, (i) BH-22 (30.3 m bgl): olive colored clay layer, (j) BH-13 (31.8 m bgl): brown colored clay layer with mottles of Fe and Mn oxyhydroxides and (k) BH-13 (37.9 m bgl): brown colored sand of aquifer-3, dated as ~70 ka.
varies spatially between 2 and 9 m, becoming thin along the east margin of the study area (BH-4, BH-7, BH-16) (Fig. 2). In the upper portion of the surface sandy clay layer the presence of plant debris was common and the sediment was mostly light brown in color (Fig. 3). The mottles of brown to black colored Fe and Mn oxyhydroxides were visible in the water table fluctuation zone (2–5 m bgl) (Fig. 3). According to the type of formation, three types of aquifer can be distinguished within the drilling depth. Aquifer-1 is unconfined, extending continuously just underneath the surface sandy clay layer to a depth of 50 m bgl, where drilling was stopped. Geographically the aquifer-1 is distributed along the northwest and eastern margin of the study area, parallel to the River Hooghly and Ichamati, respectively (Figs. 2, A.2). Very fine to medium sized sand of dark gray color with a very little portion of silt are prevailing in the aquifer-1 (Fig. 3), representing the paleochannel (PC) sequences (McArthur et al., 2008), deposited by River Hooghly and Ichamati. Occasionally, very thin films of organic carbon rich clay layers are inter-bedded within the aquifer-1, particularly along the eastern margin of the study area (BH-6, BH-7, BH-16) (Fig. 2). The overbank deposition of dark gray to black colored soft sandy clay to clay layer of thickness 15–30 m is extensive along the central floodplain (north–south transect) that widens towards southwestern part of the study area (Fig. 2). Within these overbank deposits, two sublayers of black peaty sediment (Fig. 3) are observed within the depth ranges of 5–15 m bgl and 14–32 m bgl (Fig. 2). The thickness of the peaty sediment layer in the lower part varies between 0.5 and 9 m (average: 3.42 m) and is more extensively distributed over the study
area, compared to that observed in the upper depth interval (thickness: 1.5–3.0 m, average: 1.78 m) (Fig. 2). Sometimes an aquifer is formed locally within these overbank deposits; however, its vertical extension is very much limited within the 25 m bgl and has been classified as aquifer-2 (Fig. 2). Throughout the study area, the aquifer-2 is laterally connected to the aquifer-1, except at the borehole BH-8 along the cross section E-E (Fig. 2). Fine to medium sized sand of dark gray color is prevailing in this aquifer also (Fig. 3). However, visibly, the silt content in this aquifer sediment is higher, compared to that in aquifer1. As discussed below, the chemical composition of groundwater in these two aquifers is also similar. Thus, because of similar color of the aquifer materials, henceforth the aquifer-1 and 2 are collectively designated as gray sand aquifers (GSA) and constitute 40% (1.84 × 109 m3) of the total drilled volume (4.65 × 109 m3). Aquifer-3 is distributed along the central (north–south transect) and southwestern part of the study area, underneath the overbank deposits, representing the paleointerfluve (PI) aquifer (McArthur et al., 2008) of the study area and is vertically extended up to the depth where drilling was stopped (Fig. 2). Since, in all the locations drilling was stopped before reaching the next clay aquitard, currently we do not have exact information of the thickness of this aquifer. However, according to the driller's opinion, this aquifer generally extends up to a depth of around 60 m bgl. Aquifer-3 represents 10% (4.8 × 108 m3) of the total drilled volume and is mainly composed of brown colored medium to coarse sand. Occasionally, the presence of fine to medium sized gravel was also observed during drilling (Fig. 3). The grain size and the intensity
A. Biswas et al. / Science of the Total Environment 485–486 (2014) 12–22
from other parts of the Bengal Basin (Pal and Mukherjee, 2008, 2009; von Brömssen et al., 2007, 2008).
Table 1 Dates of aquifer sediments from the study area. Sample ID
Lab. ID
Borehole
Depth (m bgl)
Age (ka) 2.86 4.90 6.25 8.51
Calibrated age (ka)
14
C dating B2_35 B6_115 B13_50 B13_85
LuS 10074 LuS 10075 LuS 10076 LuS10077
BH-2 BH-6 BH-13 BH-13
10.6 34.8 15.2 25.8
OSL dating BH-13_OSL_1 BH-13_OSL_2
Lund 13011 Lund 13012
BH-13 BH-13
37.3 43.6
± ± ± ±
0.05 0.05 0.06 0.06
2.99 5.63 7.21 9.51
± ± ± ±
0.04 0.02 0.03 0.02
72 ± 7 66 ± 7
of brown color of the aquifer sediment were noticed to decrease with increasing depth of the aquifer (Fig. 3), as also noticed by Hoque et al. (2012) at Barasat, North 24-Parganas, West Bengal (45 km south of the present study area). Because of the color of the aquifer materials, hereafter this aquifer is designated as brown sand aquifer (BSA). Though, throughout the area, this aquifer is laterally connected to the aquifer-1, in all the boreholes it is markedly separated from the overlying overbank deposits by the successive sequences of olive to pale blue and brown colored hard clay layers (Figs. 2 and 3). The thickness of the olive to pale blue clay layer spatially varies between 0.9 and 6.1 m (average: 3.3 m) within the depth range of 23–35 m bgl. While, the thickness of the brown clay layer varies between 1.5 and 7.5 m (average: 3.89 m) in the depth interval of 24–36 m bgl, except in two boreholes (BH-11 and 12) in the southwestern part of the study area. In these two boreholes the brown clay layer continued from 21 m bgl to the depth of 50 m bgl, where drilling was stopped (Fig. 2). However, the driller confirmed the presence of BSA at the base of brown clay layer around the depth of 70–80 m bgl. Brown to black patches of Fe and Mn oxyhydroxides were also common in the brown colored clay layers (Fig. 3). Goodbred and Kuehl (2000) have designated this clay as a paleosol of lateritic uplands and McArthur et al. (2008) have further classified it as the Last Glacial Maximum Paleosol (LGMP). A similar shallow hydrostratigraphic framework has also been reported as close as at Barasat (Hoque et al., 2012; McArthur et al., 2008, 2011) and
3.3. Ages of the aquifer sediment and assessment of the Holocene sedimentation rate The results of OSL dating of the two sand samples and 14C ages of the four peaty sediments, collected from the different drilling depths of the study area have been given in Table 1. The results indicate that the sand samples, collected from the aquifer-3 at the depth of 37.3 m bgl and 43.6 m bgl of BH-13 correspond to the age of 72 ± 7 ka and 66 ± 7 ka respectively. Since the luminescence characteristics of sample BH-13_OSL_2 were not as good as the first sample, the age of this sample (66 ± 7 ka) should be considered as minimum age of the sample. However, the close agreements of ages of the two samples confirm that these sediments were deposited around 70 ka. Out of the four samples dated for 14 C, three represent the overbank deposits and the fourth one was retrieved from the inter-bedded clay layer in channel fill sands. For the overbank deposits, the oldest age of 9.51 ± 0.02 ka is observed for the sediment at the depth of 25.8 m bgl in BH-13 (Table 1). The other peaty sediments retrieved from the depth of 15.2 m bgl in the same borehole and 10.6 m bgl in BH-2 correspond to the age of 7.21 ± 0.03 ka and 2.99 ± 0.04 ka, respectively (Table 1). For the channel fill deposit, the 14C dating of peaty sediment collected from the depth of 34.8 m bgl in BH-6 has yielded the age of 5.63 ± 0.02 ka. It is interesting to note that despite the deeper depth of sample collection for the channel deposits (34.8 m bgl), the age is ~ 4 ka younger compared to the deepest sample collected from the overbank deposits (25.8 m bgl) (Table 1). The three 14C dates available for the sediment of overbank deposits have been used to assess the Holocene sedimentation rate. The linear interpolation of the age-depth relation indicates that within the period of early Holocene to mid Holocene (9.50–7.00 ka) the rate of sedimentation was 4.6 m/ka, which was decreased to 1.1 m/ka in the mid Holocene to late Holocene period (7.00–3.00 ka). It is worthwhile to mention that these estimations very closely agree to the estimations by Sarkar et al. (2009) (4.4 m/ka for early Holocene to mid Holocene
40
Concentration of Fe (mg/L)
500 400 300 200 100
32 24 16 8 0
0 Aqu-1 Aqu-2 Aqu-3 (n = 128)(n = 35) (n = 114)
Aqu-1 Aqu-2 Aqu-3 (n = 128)(n = 35)(n = 114)
4000
Concentration of Mn (µg/L)
Concentration of As (µg/L)
17
3200
Legend Max. 75 percentile Median 25 percentile Min.
2400 1600 800 0 Aqu-1 Aqu-2 Aqu-3 (n = 128)(n = 35)(n = 114)
Fig. 4. Distribution of As, Fe and Mn in groundwater of different aquifers in the study area.
18
A. Biswas et al. / Science of the Total Environment 485–486 (2014) 12–22
and 0.5–1.2 m/ka for mid Holocene to late Holocene period) for the southern part of the western Bengal Basin. The close agreements of our estimations to that of Sarkar et al. (2009) probably indicate that these Holocene sedimentation rates are regionally valid for the whole western Bengal Basin. 3.4. Distribution of As, Fe and Mn in groundwater of different aquifers The aquifer type-wise classification indicates that out of the 277 TWs, for which the screen is placed within sand units according to the proposed lithologic model, 128, 35 and 114 TWs are installed in the aquifer-1, aquifer-2 and aquifer-3 respectively. The results of the determination of As, Fe and Mn in groundwater collected from these TWs have been provided in Appendix Table A.1 and their aquifer type-wise distributions have been presented in Fig. 4. The results indicate that the concentrations of dissolved As and Fe are very high in aquifer-1 and 2, compared to aquifer-3. Only 12% (n = 15) and 17% (n = 6) of the TWs installed in aquifer-1 and 2 respectively are safe for As compared to the WHO provisional drinking water guideline and the Indian national drinking water standard for acceptable limit of 10 μg/L. In contrast, in 75% (n = 85) of the TWs of aquifer-3 the concentration of As is within the safe limit and more specifically the concentration is b2 μg/L in 55% (n = 63) of the wells. Similarly, in aquifer-1 and 2 the concentration of Fe is b1 mg/L in only 6 (5%) and 2 (6%) TWs respectively and the concentration is N 5 mg/L in 66 (52%) and 11 (31%) TWs respectively.
However, in aquifer-3 the concentration of Fe in 82 TWs (72%) is b1 mg/L and only in 9 (8%) TWs the concentration exceeds 5 mg/L (Fig. 4). The distribution of dissolved Mn in aquifers shows opposite trend to that of As and Fe, the concentration being high in aquifer-3 and low in aquifer-1 and 2 (Fig. 4). In 80% of the TWs of aquifer-1 (n = 102) and aquifer-2 (n = 28) the concentration of Mn is below the value of the previous health-based WHO drinking water guideline (400 μg/L), while the concentration exceeds this cut-off value for 85% (n = 97) of the TWs installed in aquifer-3. If the national drinking water standard of India for the permissible limit of Mn (300 μg/L) is considered as the cut-off value, the extent of contamination in the TWs of aquifer-3 becomes 89%. More specifically, in 16% (n = 18) of these TWs the concentration of Mn is N1500 μg/L with a maximum concentration of 3396 μg/L. The results indicate that the aquifer-1 and 2 are very much similar in terms of water quality and thus collectively considered as GSA because of gray color of the aquifer sand as mentioned before. Henceforth, the aquifer-wise hydrogeochemical comparison is limited in terms of color of the aquifer sand such as BSA and GSA only. The spatial distributions of As, Fe and Mn in BSA and GSA have been shown in Fig. 5, which indicates that the TWs in BSA with exceptionally high concentration of As, Fe and Mn are mostly located along the margin, which supports the findings of McArthur et al. (2012) at Barasat, West Bengal. However, no such pattern in the spatial distributions is observed for the TWs installed in GSA (Fig. 5). Similar aquifer type-wise
Fe
As
Mn
Latitude
GSA
BSA
Longitude Fe
As
Contents lists available at ScienceDirect
Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv
Shallow hydrostratigraphy in an arsenic affected region of Bengal Basin: Implication for targeting safe aquifers for drinking water supply Ashis Biswas a,b,⁎, Prosun Bhattacharya a, Abhijit Mukherjee c, Bibhash Nath d, Helena Alexanderson e, Amit K. Kundu b, Debashis Chatterjee b, Gunnar Jacks a a KTH-International Groundwater Arsenic Research Group, Division of Land and Water Resources Engineering, Department of Sustainable Development, Environmental Science and Engineering, KTH Royal Institute of Technology, Teknikringen 76, SE-100 44 Stockholm, Sweden b Department of Chemistry, University of Kalyani, Kalyani 741235, West Bengal, India c Department of Geology and Geophysics, Indian Institute of Technology-Kharagpur, Kharagpur 721302, West Bengal, India d School of Geosciences, The University of Sydney, Sydney, NSW 2006, Australia e Department of Geology, Lund University, Sölvegatan 12, SE-223 62 Lund, Sweden
H I G H L I G H T S • • • • •
Shallow hydrostratigraphic framework over an area of 100 km2 is investigated. Two types of aquifer viz. brown and grey sand aquifer (BSA and GSA) are identified. Concentration of As and Fe is high (N 10 µg/L) at GSA, while low at BSA. Despite low concentration of As, the concentration of Mn is high (N 400 µg/L) at BSA. BSA should be targeted only when no other As safe drinking source is available.
a r t i c l e
i n f o
Article history: Received 16 January 2014 Received in revised form 7 March 2014 Accepted 7 March 2014 Available online 1 April 2014 Editor: D. Barcelo Keywords: Bengal Basin Shallow hydrostratigraphy Groundwater Drinking water supply Arsenic Manganese
a b s t r a c t To delineate arsenic (As) safe aquifer(s) within shallow depth, the present study has investigated the shallow hydrostratigraphic framework over an area of 100 km2 at Chakdaha Block of Nadia District, West Bengal. Drilling of 29 boreholes and subsequent hydrostratigraphic modeling has identified three types of aquifer within 50 m below ground level (bgl). Aquifer-1 represents a thick paleochannel sequence, deposited parallel to the River Hooghly and Ichamati. Aquifer-2 is formed locally within the overbank deposits in the central floodplain area and its vertical extension is strictly limited to 25 m bgl. Aquifer-3 is distributed underneath the overbank deposits and represents an interfluvial aquifer of the area. Aquifer-3 is of Pleistocene age (~70 ka), while aquifer-1 and 2 represent the Holocene deposits (age b9.51 ka), indicating that there was a major hiatus in the sediment deposition after depositing the aquifer-3. Over the area, aquifer-3 is markedly separated from the overlying Holocene deposits by successive upward sequences of brown and olive to pale blue impervious clay layers. The groundwater quality is very much similar in aquifer-1 and 2, where the concentration of As and Fe very commonly exceeds 10 μg/L and 5 mg/L, respectively. Based on similar sediment color, these two aquifers have jointly been designated as the gray sand aquifer (GSA), which constitutes 40% (1.84 × 109 m3) of the total drilled volume (4.65 × 109 m3). In aquifer-3, the concentration of As and Fe is very low, mostly b2 μg/L and 1 mg/L, respectively. This aquifer has been designated as the brown sand aquifer (BSA) according to color of the aquifer materials and represents 10% (4.8 × 108 m3) of the total drilled volume. This study further documents that though the concentration of As is very low at BSA, the concentration of Mn often exceeds the drinking water guidelines. © 2014 Elsevier B.V. All rights reserved.
1. Introduction
⁎ Corresponding author at: Department of Geological Sciences, University of Saskatchewan, 114 Science Place, Saskatoon, SK S7N 5E2, Canada. Tel.: +1 306 966 5764; fax: +1 306 966 8593. E-mail addresses: [email protected], [email protected] (A. Biswas).
http://dx.doi.org/10.1016/j.scitotenv.2014.03.045 0048-9697/© 2014 Elsevier B.V. All rights reserved.
The Bengal Basin, located at the head of Bay of Bengal, is the world's largest fluvio-deltaic system (Alam et al., 2003; Coleman, 1981; Mukherjee et al., 2009; Neidhardt et al., 2013). It is being drained by the River Ganges, Brahmaputra and Meghna and their tributaries and distributaries, also known as the Ganges–Brahmaputra–Meghna (GBM)
A. Biswas et al. / Science of the Total Environment 485–486 (2014) 12–22
basin (Mukherjee et al., 2007). It covers an area of ~200,000 km2 (most of Bangladesh and neighboring part of West Bengal, currently Paschim Banga) and hosts more than 2% (120 millions) of the world's population (Mukherjee et al., 2007). As people living in this region mostly relies on groundwater for domestic and agricultural purpose, the occurrence of elevated dissolved arsenic (As) in groundwater has put millions of people at the risk of chronic As toxicity (Chakraborti et al., 2004; Fendorf et al., 2010; Mukherjee and Bhattacharya, 2001). The extent of severity has been termed as the world's largest mass poisoning in human history (Smith et al., 2000). Since the first reporting of As in the aquifers of West Bengal in 1978 (Saha, 1984), numerous investigations have been carried out in different parts of the Bengal Basin (mostly in Bangladesh) to characterize source, regulating processes and distribution of As in the aquifers and development of possible mitigation alternatives (Bhattacharya et al., 1997; Charlet and Polya, 2006; Fendorf et al., 2010; Smedley and Kinniburgh, 2002; Winkel et al., 2008). It is revealed that the occurrence of high dissolved As in groundwater of this region is mostly confined to shallow aquifers (depth b 60 m) of Holocene age (BGS and DPHE, 2001; Horneman et al., 2004; Ravenscroft et al., 2005). On the contrary, the deep aquifers (depth N150 m) of Pleistocene age are mostly safe (b10 μg/L) (Burgess et al., 2010; Michael and Voss, 2008). However, because of high installation costs, the construction of tubewells (TWs) at the deep aquifers for domestic purpose is hardly affordable by rural villagers and it is argued that the identification of safe aquifer(s) within the shallow depth is necessary to improve the As mitigation scenario in rural Bengal (Biswas et al., 2012a; Hug et al., 2011; van Geen et al., 2007; von Brömssen et al., 2007). In recent years, one of the most highlighted aspects of As contamination in groundwater is its spatial distribution within the shallow aquifers. The concentration of dissolved As in the aquifers varies spatially (from unsafe to safe and vice versa) within a scale of tens of meters (van Geen et al., 2002, 2003). In the last few years numerous studies have been undertaken to explore the sedimentological control on the spatial distribution of dissolved As in shallow aquifers. The important finding is that the As contamination of groundwater is restricted to the gray sand aquifer and the reddish-brown sand aquifer is very rarely contaminated (Biswas et al., 2012a; Bundschuh et al., 2010; Pal and Mukherjee, 2008, 2009; McArthur et al., 2008, 2011; von Brömssen
Nepal
et al., 2007). Consequently, the reddish-brown sand aquifer has been proposed as an alternative drinking water source, which can be targeted by locally available low-cost drilling technology (McArthur et al., 2011; von Brömssen et al., 2007). However, before advocating for targeting this aquifer for safe drinking water supply it is necessary to investigate its regional distribution and subsurface geologic architecture at different geographic locations of the basin. In the present study, we have undertaken a detailed subsurface lithological investigation to explore the shallow hydrostratigraphic framework in an As-affected region of the central western Bengal Basin. In order to extend the current understanding of basin fill history and explain the evolution of the hydrostratigraphic framework in the study area, dating of different litho-units has been conducted and local stratigraphic units have been correlated to the regional late Quaternary stratigraphic units reported from other parts of the Bengal Basin. This study has further documented the distribution of dissolved As, Fe and Mn in different aquifer types to justify the potentiality of targeting safe aquifer within the shallow depth in As affected regions of the Bengal Basin.
2. Materials and methods 2.1. Study area The study area is located at the Chakdaha Block of Nadia District, West Bengal, 60 km north of Kolkata city and approximately 170 km inland from the present coastline of Bay of Bengal (Fig. 1). Geographically the study area represents the stable shelf part of the central western Bengal Basin (Fig. 1) (Mukherjee et al., 2009). The River Hooghly (distributaries of River Ganges) and Ichamati bound the study area in the west and east, respectively. The area is very flat; elevation varies within ~0–20 m above mean sea level (Neidhardt et al., 2013). Throughout the area prominent geomorphological features are abandoned channels, meandering scars, oxbow lakes and ponds (Fig. 1), which highlight the frequent changes in the channel courses during the recent past (Nath et al., 2005; Neidhardt et al., 2013). Hydrogeologically, the aquifers of the study area represent part of the “Sonar Bangla Aquifer” (Mukherjee et al., 2007).
a
Himalayan Foothills Bhutan
b BH-18
India
26°N
Investigation area 22°N
Indo-Burmese Fold Belt
Latitude
24°N
Bengal Basin
Bangladesh
BH-21
(A)
Dauki Fault India West Bengal
13
BH-24
BH-23
BH-17
(B)
Chakdaha BH-15 RS BH-2
(C)
BH-26
(A')
BH-1 BH-28 BH-4
(B')
BH-29 BH-22
BH-27
BH-7
BH-20 BH-25
BH-3
BH-13
(D)
(C')
BH-14 BH-10
BH-12
(E)
BH-19
BH-16
(D')
BH-11 BH-5
BH-8
BH-9 BH-6
Bay of Bengal
(E')
4.75 km 100 km
88°E
90°E
92°E
Longitude
Borehole Location
Fig. 1. Study area maps showing a. physiographic settings of Bengal Basin and location of Chakdaha Block of Nadia District, West Bengal, India in Bengal Basin (adopted from Neidhardt et al., 2013). b. Location of 29 boreholes and 6 west–east traverses along which lithologic cross-section has been prepared (showing in Fig. 2) over the study area. The satellite images were acquired from Google Earth 6.0.2.
14
A. Biswas et al. / Science of the Total Environment 485–486 (2014) 12–22
2.2. Subsurface sedimentological investigation In order to investigate the distribution of different aquifers and aquitards within the shallow depth, 29 boreholes with almost equal spacing (Fig. 1) were drilled over 100 km2 area during the period of March to May 2011, by locally available hand suction drilling technology (Ali, 2003), which usually can reach up to 50 m below ground level (bgl). The washed sediment samples were collected from each 1.5 m (5 ft) interval. The sampling frequency was higher when a visible change in the lithology and color was noticed during drilling. The colors (black, gray, olive to pale blue and brown) of the sediments were determined just after recovery, prior to atmospheric oxidation. Sediments were classified into the three major lithotypes of clay, sandy clay and sand. The sediment of the lithotype clay and sand represents aquitard and aquifer, respectively, while sandy clay represents aquitard with higher permeability. Based on these lithotypes and colors, lithologs for the boreholes were prepared (Appendix Fig. A.1). Using the 29 lithologs, a 3D lithologic model for the area was developed using the software RockWorks ver. 15 (RockWare, Golden, CO, USA). The lithological data of each borehole were interpolated in three dimensions by the algorithm of “lithoblending”, where each voxel is assigned a lithology value corresponding to a particular lithotype of closest known data point and then the processes continue both horizontally and vertically to the next voxel and so on, until a voxel with a specified lithology value (that is next known data point) is encountered. The model was optimized by changing the node spacing and the final model was selected that showed least sensitivity towards the change in grid size (Mukherjee et al., 2007). The final model has the resolution of 100 m (X) × 100 m (Y) × 0.5 m (Z), which results in a node spacing of 99 (X) × 95 (Y) × 99 (Z), with the total number of 931,095 solid grids, each having the volume of 5000 m 3. Finally, using this 3D lithologic model, 6 west–east 2D lithologic cross-sections and a number of plan-view maps that represent the distribution of aquifers and aquitards over the study area at a specific depth were developed (Mukherjee et al., 2007). 2.3. Dating of the aquifer sediments A total of 4 (borehole: BH-13 depth: 15.2 m and 25.8 m bgl; borehole: BH-2 depth: 10.6 m bgl and borehole: BH-6 depth: 34.8 m bgl) organic rich peaty sediments were collected during drilling for 14 C age dating. The samples were collected in airtight polyethylene zipper bags and preserved at − 18 °C to minimize the bacterial activity, prior to sending for analysis. The samples were analyzed within six months of collection at the Lund University Radiocarbon Dating Laboratory by Single Stage Accelerator Mass Spectrometer (SSAMS) (National Electrostatic Corp. (NEC), Wisconsin, USA). The precision of the analysis was ±50 14C years. The 14C ages were calibrated with reference to the atmospheric data from Reimer et al. (2009) by the software OxCal v3.10 (Bronk Ramsey and Lee, 2013). The range of calibrated ages was reduced to the point estimate by taking the midpoint of two terminal ages that limit the 68.2% (one standard deviation) range of the distribution and taking the half of the difference between this midpoint and one terminal age as the single standard deviation of estimate (Blaauw and Heegaard, 2012). A new drilling was conducted in April 2013, at just 1 m apart from the old drilled hole of BH-13 for the coring of OSL dating. Two samples were collected at the depths of 37.3 m bgl (BH-13_OSL_1) and 43.6 m bgl (BH-13_OSL_2). The targeted depth was reached by normal hand suction drilling technology (Ali, 2003), then a PVC tube (length: 30 cm, thickness: 2 mm) was fitted at the top of the drilling pipe and hammered down to collect the samples. After recovery, the core tubes were wrapped with multiple layers of aluminum foil and black colored insulating tape and preserved at complete darkness. The samples were
analyzed within two months of sample collection at the Lund Luminescence Laboratory, Lund University. Only the central part of each core tube was sampled for the analysis. Large single aliquots of 180–250 μm quartz grains in cups were analyzed in a Risø TL/OSL reader model DA-20. A Single Aliquot Regeneration (SAR) protocol was used (Murray and Wintle, 2000, 2003), with the respective pre-heat and cut-heat at 220 °C and 200 °C for sample BH-13_OSL_1 and 260 °C and 240 °C for sample BH-13_OSL_2, as determined by preheat plateau and dose recovery tests. The samples were optically stimulated by blue light sources (470 ± 30 nm; ~50 mW/cm2). Since the second sample suffered from feldspar contamination (mean IR/blue ratio N10%), it was additionally stimulated with infrared light (post-IR blue protocol; Banerjee et al., 2001). The detection was made through 7 mm of U340 glass filter (Bøtter-Jensen et al., 2000), the first 1.6 s of the signal was integrated for the peak and the following 1.6 s for the background. Aliquots were accepted if they had a test dose error of b 10%, a signal more than three times the background and a recycling ratio within 10% of unity (b 20% for sample BH-13_OSL_2). The dose recovery ratio of the two samples (0.96 ± 0.06, n = 6 and 0.99 ± 0.11, n = 4, respectively) showed that the selected protocols were able to correctly measure a given dose. The environmental dose rate was determined by gamma spectrometry and by estimating the contribution of cosmic radiation as described by Murray et al. (1987) and Prescott and Hutton (1994), respectively. Average water content was estimated from the values of natural and saturated water contents as measured on the sediment. 2.4. Groundwater sampling and analysis A total of 423 groundwater samples were collected for the determination of As, Fe and Mn, during March to May 2011 from the existing TWs, which were installed at the depth b 70 m bgl (Biswas et al., 2012b). After purging the TWs for few minutes groundwater sample was collected in a prewashed high density polyethylene vials (Tarson) and acidified on-site by HNO 3 (1% v/v, Suprapur, Merck). The concentration of As in the samples was determined by a hydride generation atomic absorption spectrophotometer (HG-AAS, detection limit b 1 μg/L), while the concentration of Fe and Mn was determined by a graphite furnace atomic absorption spectrophotometer (GF-AAS, GTA-120, detection limit 1.5 μg/L and 1.0 μg/L for Fe and Mn, respectively) (Varian, AA240). For further details of groundwater sampling, analysis and quality assurance reader is referred to Biswas et al. (2012b). This data set of 423 groundwater analyses was then compiled with other available data sets for As, Fe and Mn from the study area (e.g. data set of Biswas et al., 2012a). Out of the total number of TWs in the combined data set, 325 TWs were installed in the area enclosed by the drilled boreholes and had screen position within the maximum drilling depth (50 m) and thus, the lithologic information around the screen position was known. In order to assure true representativeness of a TW to a specific aquifer type, only these TWs have been used for subsequent statistical analysis to investigate the accuracy of the lithologic model and the distribution of As, Fe and Mn in different aquifer types of the study area. 3. Results 3.1. Accuracy of the lithologic model The accuracy of the lithologic model was tested by assuming that the screen of all 325 TWs was placed in the sand and the depth of TWs reported by the TW owner during sampling was correct. For the purpose of accuracy test, the depth of the TWs up to midpoint of the screen was considered. Our field experience indicates that the screen length of the TWs with depth b 30 m is usually 1.8 m, while for the TWs of depth N 30 m, it is 3.6 m in the study area. Consequently, in order to
A. Biswas et al. / Science of the Total Environment 485–486 (2014) 12–22
(A) BH-18
BH-24
BH-26
(A )
BH-19 0
0
BH-21
15
Aqu-1 -25
-25
Aqu-1 Aqu-3
0
-45
-45
Aqu-1
3 BH-28
BH-1
BH-4
-25
Aqu-1 Aqu-1
-45
-45
Aqu-3 3
0
0
BH-2
BH-22
BH-29
BH-7
-25
-45 0
Aqu-3 4.5
0
BH-14
BH-10
BH-16
-25 -45
-45
3.5
(D ) Depth (m bgl)
Aqu-1 Aqu-3 7.0
BH-8
BH-5
Brown sand
-25
Aqu-1 0
Brown clay
9.0
Aqu-2
(E) BH-12 BH-11
Grey sand
0
BH-3
Olive clay
-45
Aqu-1
(D )
-25
-45 -25
BH-16
0
Depth (m bgl)
BH-25
BH-20
Aqu-1
(D) BH-13
Brown sandy clay
8
BH-27
0
Grey clay -45
Aqu-3 4
(C) BH-15
Black clay
-25
Aqu-1
0
(C )
Grey sandy clay
Aqu-2 Aqu-1
Lithology Legend
6
0
(C) BH-15
(B ) Depth (m bgl)
Aqu-2
-25
Depth (m bgl)
0
BH-23
0
(B) BH-17
6
BH-9
BH-6
(E )
0
0
-45 -25
Aqu-1 0
5
-25 -45
Aqu-2
10
Length of the Traverse (km) Fig. 2. Six west–east lithologic cross-sections along the traverses shown in Fig. 1. Aqu-1, 2 and 3 represents aquifer-1, 2 and 3 respectively.
determine the depth of the TWs up to the midpoint of screen, 1 m and 2 m were subtracted from the total well depth of b 30 m and N30 m, respectively. In the next step, the TWs at each specific depth were grouped together and placed on the lithologic plan-view map of the corresponding depth to test whether the screen position of TWs was located at the sand units. The accuracy analysis reveals that according to the proposed lithologic model 277, 11 and 37 TWs are placed in sand, sandy clay and clay respectively, which indicates 85% accuracy of the model.
3.2. Shallow hydrostratigraphic framework in the study area Six 2D west–east lithologic cross-sections over the study area have been displayed in Fig. 2. Additionally, a few plan-view maps representing the distribution of aquifer and aquitard at different specific depths over the study area have been presented in Appendix Fig. A.2. These cross sections and plan view maps represent the distribution of a very complex aquifer–aquitard framework within the drilling depth. The surface sandy clay layer extends over the entire study area and the thickness
16
A. Biswas et al. / Science of the Total Environment 485–486 (2014) 12–22
a
c
b
d
f
h
e
g
i
j
k
Fig. 3. Photographs of few representative sediment samples collected during drilling over the study area. Hand written letters represent depth in feet from where samples were collected. (a) BH-17: represents typical channel deposits, where a thick layer of sand (aquifer-1) is extended from just below of surface sandy clay unit to the depth, where drilling was stopped, (b) BH-22: represents interfluvial sequence, where aquifer-2 is distributed between 50 and 70 ft within overbank deposits, below aquifer-3 is distributed and separated from the overbank deposits by successive sequences of olive to pale blue and brown clay layers, (c) BH-17 (depth: 3.03 m bgl): light brown surface sandy clay layer with mottles of Fe and Mn oxyhydroxides, (d) BH-2 (10.6 m bgl): black peaty materials dated as 2.99 ± 0.04 ka, (e) BH-13 (25.8 m bgl): black peaty materials dated as 9.51 ± 0.02 ka, (f) BH-4 (24.5 m bgl): black peaty materials with shell fragments, (g) BH-4 (27.3 m bgl): typical dark gray colored sand found in aquifer-1 and 2, (h) BH-22 (28.9 m bgl): pale blue colored clay layer, (i) BH-22 (30.3 m bgl): olive colored clay layer, (j) BH-13 (31.8 m bgl): brown colored clay layer with mottles of Fe and Mn oxyhydroxides and (k) BH-13 (37.9 m bgl): brown colored sand of aquifer-3, dated as ~70 ka.
varies spatially between 2 and 9 m, becoming thin along the east margin of the study area (BH-4, BH-7, BH-16) (Fig. 2). In the upper portion of the surface sandy clay layer the presence of plant debris was common and the sediment was mostly light brown in color (Fig. 3). The mottles of brown to black colored Fe and Mn oxyhydroxides were visible in the water table fluctuation zone (2–5 m bgl) (Fig. 3). According to the type of formation, three types of aquifer can be distinguished within the drilling depth. Aquifer-1 is unconfined, extending continuously just underneath the surface sandy clay layer to a depth of 50 m bgl, where drilling was stopped. Geographically the aquifer-1 is distributed along the northwest and eastern margin of the study area, parallel to the River Hooghly and Ichamati, respectively (Figs. 2, A.2). Very fine to medium sized sand of dark gray color with a very little portion of silt are prevailing in the aquifer-1 (Fig. 3), representing the paleochannel (PC) sequences (McArthur et al., 2008), deposited by River Hooghly and Ichamati. Occasionally, very thin films of organic carbon rich clay layers are inter-bedded within the aquifer-1, particularly along the eastern margin of the study area (BH-6, BH-7, BH-16) (Fig. 2). The overbank deposition of dark gray to black colored soft sandy clay to clay layer of thickness 15–30 m is extensive along the central floodplain (north–south transect) that widens towards southwestern part of the study area (Fig. 2). Within these overbank deposits, two sublayers of black peaty sediment (Fig. 3) are observed within the depth ranges of 5–15 m bgl and 14–32 m bgl (Fig. 2). The thickness of the peaty sediment layer in the lower part varies between 0.5 and 9 m (average: 3.42 m) and is more extensively distributed over the study
area, compared to that observed in the upper depth interval (thickness: 1.5–3.0 m, average: 1.78 m) (Fig. 2). Sometimes an aquifer is formed locally within these overbank deposits; however, its vertical extension is very much limited within the 25 m bgl and has been classified as aquifer-2 (Fig. 2). Throughout the study area, the aquifer-2 is laterally connected to the aquifer-1, except at the borehole BH-8 along the cross section E-E (Fig. 2). Fine to medium sized sand of dark gray color is prevailing in this aquifer also (Fig. 3). However, visibly, the silt content in this aquifer sediment is higher, compared to that in aquifer1. As discussed below, the chemical composition of groundwater in these two aquifers is also similar. Thus, because of similar color of the aquifer materials, henceforth the aquifer-1 and 2 are collectively designated as gray sand aquifers (GSA) and constitute 40% (1.84 × 109 m3) of the total drilled volume (4.65 × 109 m3). Aquifer-3 is distributed along the central (north–south transect) and southwestern part of the study area, underneath the overbank deposits, representing the paleointerfluve (PI) aquifer (McArthur et al., 2008) of the study area and is vertically extended up to the depth where drilling was stopped (Fig. 2). Since, in all the locations drilling was stopped before reaching the next clay aquitard, currently we do not have exact information of the thickness of this aquifer. However, according to the driller's opinion, this aquifer generally extends up to a depth of around 60 m bgl. Aquifer-3 represents 10% (4.8 × 108 m3) of the total drilled volume and is mainly composed of brown colored medium to coarse sand. Occasionally, the presence of fine to medium sized gravel was also observed during drilling (Fig. 3). The grain size and the intensity
A. Biswas et al. / Science of the Total Environment 485–486 (2014) 12–22
from other parts of the Bengal Basin (Pal and Mukherjee, 2008, 2009; von Brömssen et al., 2007, 2008).
Table 1 Dates of aquifer sediments from the study area. Sample ID
Lab. ID
Borehole
Depth (m bgl)
Age (ka) 2.86 4.90 6.25 8.51
Calibrated age (ka)
14
C dating B2_35 B6_115 B13_50 B13_85
LuS 10074 LuS 10075 LuS 10076 LuS10077
BH-2 BH-6 BH-13 BH-13
10.6 34.8 15.2 25.8
OSL dating BH-13_OSL_1 BH-13_OSL_2
Lund 13011 Lund 13012
BH-13 BH-13
37.3 43.6
± ± ± ±
0.05 0.05 0.06 0.06
2.99 5.63 7.21 9.51
± ± ± ±
0.04 0.02 0.03 0.02
72 ± 7 66 ± 7
of brown color of the aquifer sediment were noticed to decrease with increasing depth of the aquifer (Fig. 3), as also noticed by Hoque et al. (2012) at Barasat, North 24-Parganas, West Bengal (45 km south of the present study area). Because of the color of the aquifer materials, hereafter this aquifer is designated as brown sand aquifer (BSA). Though, throughout the area, this aquifer is laterally connected to the aquifer-1, in all the boreholes it is markedly separated from the overlying overbank deposits by the successive sequences of olive to pale blue and brown colored hard clay layers (Figs. 2 and 3). The thickness of the olive to pale blue clay layer spatially varies between 0.9 and 6.1 m (average: 3.3 m) within the depth range of 23–35 m bgl. While, the thickness of the brown clay layer varies between 1.5 and 7.5 m (average: 3.89 m) in the depth interval of 24–36 m bgl, except in two boreholes (BH-11 and 12) in the southwestern part of the study area. In these two boreholes the brown clay layer continued from 21 m bgl to the depth of 50 m bgl, where drilling was stopped (Fig. 2). However, the driller confirmed the presence of BSA at the base of brown clay layer around the depth of 70–80 m bgl. Brown to black patches of Fe and Mn oxyhydroxides were also common in the brown colored clay layers (Fig. 3). Goodbred and Kuehl (2000) have designated this clay as a paleosol of lateritic uplands and McArthur et al. (2008) have further classified it as the Last Glacial Maximum Paleosol (LGMP). A similar shallow hydrostratigraphic framework has also been reported as close as at Barasat (Hoque et al., 2012; McArthur et al., 2008, 2011) and
3.3. Ages of the aquifer sediment and assessment of the Holocene sedimentation rate The results of OSL dating of the two sand samples and 14C ages of the four peaty sediments, collected from the different drilling depths of the study area have been given in Table 1. The results indicate that the sand samples, collected from the aquifer-3 at the depth of 37.3 m bgl and 43.6 m bgl of BH-13 correspond to the age of 72 ± 7 ka and 66 ± 7 ka respectively. Since the luminescence characteristics of sample BH-13_OSL_2 were not as good as the first sample, the age of this sample (66 ± 7 ka) should be considered as minimum age of the sample. However, the close agreements of ages of the two samples confirm that these sediments were deposited around 70 ka. Out of the four samples dated for 14 C, three represent the overbank deposits and the fourth one was retrieved from the inter-bedded clay layer in channel fill sands. For the overbank deposits, the oldest age of 9.51 ± 0.02 ka is observed for the sediment at the depth of 25.8 m bgl in BH-13 (Table 1). The other peaty sediments retrieved from the depth of 15.2 m bgl in the same borehole and 10.6 m bgl in BH-2 correspond to the age of 7.21 ± 0.03 ka and 2.99 ± 0.04 ka, respectively (Table 1). For the channel fill deposit, the 14C dating of peaty sediment collected from the depth of 34.8 m bgl in BH-6 has yielded the age of 5.63 ± 0.02 ka. It is interesting to note that despite the deeper depth of sample collection for the channel deposits (34.8 m bgl), the age is ~ 4 ka younger compared to the deepest sample collected from the overbank deposits (25.8 m bgl) (Table 1). The three 14C dates available for the sediment of overbank deposits have been used to assess the Holocene sedimentation rate. The linear interpolation of the age-depth relation indicates that within the period of early Holocene to mid Holocene (9.50–7.00 ka) the rate of sedimentation was 4.6 m/ka, which was decreased to 1.1 m/ka in the mid Holocene to late Holocene period (7.00–3.00 ka). It is worthwhile to mention that these estimations very closely agree to the estimations by Sarkar et al. (2009) (4.4 m/ka for early Holocene to mid Holocene
40
Concentration of Fe (mg/L)
500 400 300 200 100
32 24 16 8 0
0 Aqu-1 Aqu-2 Aqu-3 (n = 128)(n = 35) (n = 114)
Aqu-1 Aqu-2 Aqu-3 (n = 128)(n = 35)(n = 114)
4000
Concentration of Mn (µg/L)
Concentration of As (µg/L)
17
3200
Legend Max. 75 percentile Median 25 percentile Min.
2400 1600 800 0 Aqu-1 Aqu-2 Aqu-3 (n = 128)(n = 35)(n = 114)
Fig. 4. Distribution of As, Fe and Mn in groundwater of different aquifers in the study area.
18
A. Biswas et al. / Science of the Total Environment 485–486 (2014) 12–22
and 0.5–1.2 m/ka for mid Holocene to late Holocene period) for the southern part of the western Bengal Basin. The close agreements of our estimations to that of Sarkar et al. (2009) probably indicate that these Holocene sedimentation rates are regionally valid for the whole western Bengal Basin. 3.4. Distribution of As, Fe and Mn in groundwater of different aquifers The aquifer type-wise classification indicates that out of the 277 TWs, for which the screen is placed within sand units according to the proposed lithologic model, 128, 35 and 114 TWs are installed in the aquifer-1, aquifer-2 and aquifer-3 respectively. The results of the determination of As, Fe and Mn in groundwater collected from these TWs have been provided in Appendix Table A.1 and their aquifer type-wise distributions have been presented in Fig. 4. The results indicate that the concentrations of dissolved As and Fe are very high in aquifer-1 and 2, compared to aquifer-3. Only 12% (n = 15) and 17% (n = 6) of the TWs installed in aquifer-1 and 2 respectively are safe for As compared to the WHO provisional drinking water guideline and the Indian national drinking water standard for acceptable limit of 10 μg/L. In contrast, in 75% (n = 85) of the TWs of aquifer-3 the concentration of As is within the safe limit and more specifically the concentration is b2 μg/L in 55% (n = 63) of the wells. Similarly, in aquifer-1 and 2 the concentration of Fe is b1 mg/L in only 6 (5%) and 2 (6%) TWs respectively and the concentration is N 5 mg/L in 66 (52%) and 11 (31%) TWs respectively.
However, in aquifer-3 the concentration of Fe in 82 TWs (72%) is b1 mg/L and only in 9 (8%) TWs the concentration exceeds 5 mg/L (Fig. 4). The distribution of dissolved Mn in aquifers shows opposite trend to that of As and Fe, the concentration being high in aquifer-3 and low in aquifer-1 and 2 (Fig. 4). In 80% of the TWs of aquifer-1 (n = 102) and aquifer-2 (n = 28) the concentration of Mn is below the value of the previous health-based WHO drinking water guideline (400 μg/L), while the concentration exceeds this cut-off value for 85% (n = 97) of the TWs installed in aquifer-3. If the national drinking water standard of India for the permissible limit of Mn (300 μg/L) is considered as the cut-off value, the extent of contamination in the TWs of aquifer-3 becomes 89%. More specifically, in 16% (n = 18) of these TWs the concentration of Mn is N1500 μg/L with a maximum concentration of 3396 μg/L. The results indicate that the aquifer-1 and 2 are very much similar in terms of water quality and thus collectively considered as GSA because of gray color of the aquifer sand as mentioned before. Henceforth, the aquifer-wise hydrogeochemical comparison is limited in terms of color of the aquifer sand such as BSA and GSA only. The spatial distributions of As, Fe and Mn in BSA and GSA have been shown in Fig. 5, which indicates that the TWs in BSA with exceptionally high concentration of As, Fe and Mn are mostly located along the margin, which supports the findings of McArthur et al. (2012) at Barasat, West Bengal. However, no such pattern in the spatial distributions is observed for the TWs installed in GSA (Fig. 5). Similar aquifer type-wise
Fe
As
Mn
Latitude
GSA
BSA
Longitude Fe
As
