14
ORIGINAL ARTICLE Geochemical study of arsenic and other trace elements in groundwater and sediments of the Old Brahmaputra River Plain, Bangladesh Faruque Ahmed M. Hawa Bibi Hiroaki Ishiga Takehiko Fukushima Teruyuki Maruoka Received: 29 January 2009 / Accepted: 29 July 2009 / Published online: 13 August 2009 Ó Springer-Verlag 2009 Abstract The geochemical study of groundwaters and core sediments from the Old Brahmaputra plain of Bangladesh was conducted to investigate the distribution of arsenic and related trace elements. Groundwaters from tube wells are characterized by pH of 6.4–7.4, dissolved oxygen (DO) of 0.8–1.8 mg/l, Ca contents of 5–50 mg/l, and Fe contents of 0.2–12.9 mg/l. Arsenic concentrations ranged from 8 to 251 lg/l, with an average value of 63 lg/l. A strong positive correlation exists between As and Fe (r 2 = 0.802; p = 0.001) concentrations in groundwater. The stratigraphic sequences in the cores consist of yel- lowish silty clays at top, passing downward into grayish to yellowish clays and sands. The uppermost 3 m and lower parts (from 13 to 31 m) of the core sediments are oxidized (average oxidation reduction potential (ORP) ?170 and ?220 mV, respectively), and the ORP values gradually become negative from 3 to 13 m depths (-35 to -180 mV), indicating that anoxic conditions prevail in the shallow aquifers of the Brahmaputra plain. Age determi- nations suggest that clay horizons at *10 m depth were deposited at around 2,000 and 5,000 years BP ( 14 C ages) during the transgressive phase of sea-level change. Ele- vated concentrations of As, Pb, Zn, Cu, Ni, Cr, and V are present in the silts and clays, probably due to adsorption onto clay particles. Significant concentrations of As occur in black peat and peaty sediments at depths between 9 and 13 m. A strong positive correlation between As and Fe was found in the sediments, indicating As may be adsorbed onto Fe oxides in aquifer sediments. Keywords Arsenic Á Groundwater Á Sediment Á Peat Á Brahmaputra River Plain Á Bangladesh Introduction High concentrations of naturally occurring arsenic have been reported in groundwater from many regions, includ- ing Bangladesh, India, Nepal, Thailand, China, Taiwan, Vietnam, Chile, Hungary, and parts of the USA (Acharyya et al. 1999; Smedley and Kinniburgh 2002; Garcia-Sanchez and Alvarez-Ayuso 2003). Arsenic contamination of groundwater is now a serious hazard among the many environmental issues in Bangladesh. There has been increasing concern over the safety of groundwater from shallow aquifers in the Ganges, Brahmaputra, and Meghna (GBM) Delta, where As levels exceed the World Health Organization (WHO 2004) guideline of 10 and the 50 lg/l limit adopted in Bangladesh. Sediments act as both sources and sinks of As and other toxic trace elements (Guern et al. 2003), and the fate of trace metals is dependent on the biogeochemical transfor- mations that occur in the sediments (Peltier et al. 2003). The degradation of organic matter, present at high con- centrations in most wetland sediments, drives the formation of sediment redox gradients (Gaillard 1994). Arsenic is widely distributed as a trace constituent in rocks, soils, natural waters, and organisms, and it can be mobilized by F. Ahmed (&) Á T. Fukushima Á T. Maruoka Graduate School of Life and Environmental Sciences, University of Tsukuba, 1-1-1 Tennoudai, Tsukuba 305-8572, Japan e-mail: [email protected] M. H. Bibi Department of Environmental Science and Technology, Saitama University, 255 Shimo-okubo, Saitama 338-8570, Japan H. Ishiga Department of Geoscience, Shimane University, Matsue 690-8504, Japan 123 Environ Earth Sci (2010) 60:1303–1316 DOI 10.1007/s12665-009-0270-7

Geochemical study of arsenic and other trace elements in groundwater and sediments of the Old Brahmaputra River Plain, Bangladesh

Embed Size (px)

Citation preview

Page 1: Geochemical study of arsenic and other trace elements in groundwater and sediments of the Old Brahmaputra River Plain, Bangladesh

ORIGINAL ARTICLE

Geochemical study of arsenic and other trace elementsin groundwater and sediments of the Old BrahmaputraRiver Plain, Bangladesh

Faruque Ahmed Æ M. Hawa Bibi Æ Hiroaki Ishiga ÆTakehiko Fukushima Æ Teruyuki Maruoka

Received: 29 January 2009 / Accepted: 29 July 2009 / Published online: 13 August 2009

� Springer-Verlag 2009

Abstract The geochemical study of groundwaters and

core sediments from the Old Brahmaputra plain of

Bangladesh was conducted to investigate the distribution of

arsenic and related trace elements. Groundwaters from tube

wells are characterized by pH of 6.4–7.4, dissolved oxygen

(DO) of 0.8–1.8 mg/l, Ca contents of 5–50 mg/l, and Fe

contents of 0.2–12.9 mg/l. Arsenic concentrations ranged

from 8 to 251 lg/l, with an average value of 63 lg/l.

A strong positive correlation exists between As and Fe

(r2 = 0.802; p = 0.001) concentrations in groundwater.

The stratigraphic sequences in the cores consist of yel-

lowish silty clays at top, passing downward into grayish to

yellowish clays and sands. The uppermost 3 m and lower

parts (from 13 to 31 m) of the core sediments are oxidized

(average oxidation reduction potential (ORP) ?170 and

?220 mV, respectively), and the ORP values gradually

become negative from 3 to 13 m depths (-35 to

-180 mV), indicating that anoxic conditions prevail in the

shallow aquifers of the Brahmaputra plain. Age determi-

nations suggest that clay horizons at *10 m depth were

deposited at around 2,000 and 5,000 years BP (14C ages)

during the transgressive phase of sea-level change. Ele-

vated concentrations of As, Pb, Zn, Cu, Ni, Cr, and V are

present in the silts and clays, probably due to adsorption

onto clay particles. Significant concentrations of As occur

in black peat and peaty sediments at depths between 9 and

13 m. A strong positive correlation between As and Fe was

found in the sediments, indicating As may be adsorbed

onto Fe oxides in aquifer sediments.

Keywords Arsenic � Groundwater � Sediment �Peat � Brahmaputra River Plain � Bangladesh

Introduction

High concentrations of naturally occurring arsenic have

been reported in groundwater from many regions, includ-

ing Bangladesh, India, Nepal, Thailand, China, Taiwan,

Vietnam, Chile, Hungary, and parts of the USA (Acharyya

et al. 1999; Smedley and Kinniburgh 2002; Garcia-Sanchez

and Alvarez-Ayuso 2003). Arsenic contamination of

groundwater is now a serious hazard among the many

environmental issues in Bangladesh. There has been

increasing concern over the safety of groundwater from

shallow aquifers in the Ganges, Brahmaputra, and Meghna

(GBM) Delta, where As levels exceed the World Health

Organization (WHO 2004) guideline of 10 and the 50 lg/l

limit adopted in Bangladesh.

Sediments act as both sources and sinks of As and other

toxic trace elements (Guern et al. 2003), and the fate of

trace metals is dependent on the biogeochemical transfor-

mations that occur in the sediments (Peltier et al. 2003).

The degradation of organic matter, present at high con-

centrations in most wetland sediments, drives the formation

of sediment redox gradients (Gaillard 1994). Arsenic is

widely distributed as a trace constituent in rocks, soils,

natural waters, and organisms, and it can be mobilized by

F. Ahmed (&) � T. Fukushima � T. Maruoka

Graduate School of Life and Environmental Sciences,

University of Tsukuba, 1-1-1 Tennoudai,

Tsukuba 305-8572, Japan

e-mail: [email protected]

M. H. Bibi

Department of Environmental Science and Technology,

Saitama University, 255 Shimo-okubo, Saitama 338-8570, Japan

H. Ishiga

Department of Geoscience, Shimane University,

Matsue 690-8504, Japan

123

Environ Earth Sci (2010) 60:1303–1316

DOI 10.1007/s12665-009-0270-7

Page 2: Geochemical study of arsenic and other trace elements in groundwater and sediments of the Old Brahmaputra River Plain, Bangladesh

weathering and microbial activity (Garcia-Sanchez and

Alvarez-Ayuso 2003). Moreover, redox conditions and the

availability of possible carrier phases also account for the

mobility and toxicity of As in natural environments (Fabian

et al. 2003). The mobility of As in the environment is

strongly affected by adsorption onto mineral surfaces

(mainly oxides and hydroxides of Fe, Mn, and Al; Manning

and Goldberg 1997). Sullivan and Aller (1996) also noted

that diagenetic distribution of As in the geologic environ-

ment may be controlled by the redox behavior of Fe

(oxy)hydroxide phases in sediments.

In recent years, many geochemical studies have been

directed at sediments to determine the extent of contami-

nation caused by As (Sullivan and Aller 1996; McArthur

et al. 2001; Ahmed et al. 2004). Studies have shown that

oxides and organic matter play an important role in ele-

mental distributions in the sediments. The geochemistry of

Holocene sediments is significant due to limited alteration

during burial, diagenesis, and tectonic deformation, and it

may provide a continuous record of environmental changes

(Ishiga et al. 2000). Geochemical analyses of the Holocene

alluvial sediments have thus been performed in several parts

of Bangladesh, including the Ganges and Meghna deltaic

regions, to elucidate the behavior of As in differing sedi-

mentary environments (Dhar et al. 1997; Yamazaki et al.

2000; Nickson et al. 2000; McArthur et al. 2001; Ahmed

et al. 2004; Akai et al. 2004; many others). However, these

studies have not considered the concentrations of As in the

groundwater and sediments of the Old Brahmaputra River

Plain, and as yet very little is known about the distributions

and sources of As and other trace elements in the

Brahmaputra sediments in Bangladesh.

The previous report of BGS and DPHE (2001) showed

that the groundwater of the Mymensingh district in the Old

Brahmaputra floodplain is also contaminated with As. Such

groundwater is used for drinking and extensive irrigation

by many households and thus constitutes a human health

risk. Therefore, a detailed study of the groundwater and

sediments is required to clarify the geological background

of As contamination in the Brahmaputra plain. Ground-

water samples were collected from the Mymensingh Sadar

county in Mymensingh district. Six sediment cores were

also collected from the county to determine their geo-

chemical compositions and distribution patterns, with

emphasis on the concentrations of As and other trace

metals in the sediments and their significance. The results

show that As concentrations are greater in the silts and

clays than in the sands.

Geologic outline of the study area

Bangladesh is a low-lying country at the head of the Bay of

Bengal and occupies most of the Bengal basin. Most of the

country consists of low alluvial, coastal, and deltaic plains

and lies within an elevation of 20 m above sea level. The

sediments are derived from three major rivers, the GBM,

which form the largest delta complex in the world. The

study sites have an area of approximately 100 km2 and are

located in the Old Brahmaputra floodplain and northern

end of the Madhupur Tract within the Mymensingh district

of northern Bangladesh (Fig. 1). The Old Brahmaputra

River flows through the district from northwest to south-

east, whereas the tributaries Sutia and Barera Rivers snake

(a) (b)Fig. 1 a, b Map of Bangladesh

and location of the study sites in

Mymensingh

1304 Environ Earth Sci (2010) 60:1303–1316

123

Page 3: Geochemical study of arsenic and other trace elements in groundwater and sediments of the Old Brahmaputra River Plain, Bangladesh

through the middle part of the study area. The sampling

area is characterized by a subtropical climate. Most of the

study area is rural, and agriculture is the principal eco-

nomic activity. Deep tube wells were constructed for

extensive irrigation and drinking water supply in the

region, but shallow hand pump waters are also used for

drinking water by many households.

Madhupur tract

The Madhupur tract is a tectonically uplifted area of Pleis-

tocene clays, known as the Madhupur Clay. This tract is high

in quartz and relatively low in feldspars and micas. On the

basis of reversed magnetic polarity, Monsur (1990) has

concluded that the Madhupur Clay was formed between

730,000 and 970,000 years ago. Most of the tract has been

dissected by valleys, and some of these valleys have been

partly filled with alluvium deposits, which include organic

mud layers up to 6,700 years of age, as established by radio-

carbon dating (Brammer 1996). The soils are well drained

and characterized by dark brown to brown topsoils and red-

dish-brown to yellowish-brown or yellowish-red subsoils.

Old Brahmaputra floodplain

The evolutionary history of the youngest sedimentary

sequence in the Ganges and Brahmaputra plains was dis-

closed by Umitsu (1993). At the time of the last glacial

maximum, 18,000 years ago, world sea level was about

130 m below its present level (BGS and DPHE 2001). The

unconsolidated sediments offered little resistance to ero-

sion by the ancient course of the Brahmaputra as it tried to

cut down to a lower base level. The sea level rose rapidly

between 18,000 and 8,000 years ago to approach its present

level. In the final stage, between 6,000 and 2,000 years

ago, the sea level was about 2–3 m higher than at present,

accounting for the slightly elevated Old Brahmaputra

floodplain (BGS and DPHE 2001). The floodplain is

comprised of alluvial sediments, mainly silt and clay, with

lesser fine- to medium-grained sand. Matured dark gray

soils are rich in organic matter.

Materials and methods

Analytical procedures for water samples

Sample collection and preservation

Groundwater samples were collected from 13 tube wells in

the Mymensingh Sadar county of the Brahmaputra plain

(Fig. 1). The waters were sampled from both shallow (12–

56 m) and deep aquifers (84–137 m). The tube wells were

pumped for 10 min before sampling to stabilize the tem-

perature and dissolved oxygen (DO) content. About 200 ml

of water were taken at each site, acidified with HNO3

(Cica-Reagent, Kanto Chemicals Co., Inc., Japan) to pH 2,

and stored in acid-prewashed high-density polyethylene

bottles (rinsed with 1% HNO3 followed by thorough rins-

ing with water) without headspace. The water samples

were stored in a cooler at 4�C immediately after collection

and analyzed within 2 weeks of collection. Locations and

depths of the tube wells were recorded at each site. The

water samples were not filtered because no particulates

were observed during collection or prior to analysis. This

strategy was also adopted because an accurate measure-

ment of the total amount of As present was sought and

because the groundwater from the tube wells is consumed

directly by the local villagers without any filtration

(McArthur et al. 2004; Nickson et al. 2005).

Field measurement of water quality

Water parameters including pH, electrical conductivity

(EC), and DO concentration were measured in the field

during sampling using a portable Horiba U-23 combined

instrument (Rikagaku kenkujo, Japan). Chemical oxygen

demand (COD; alkaline KMnO4 method), ammonia

(NH4?; Nessler Test Kit), phosphorus (PO4

3-; ammonium

molybdate tetrahydrate method), nitrate (NO3-; Nessler

Test Kit), and calcium contents (Ca2?; EDTA titrimetric

method) were examined in the field.

Geochemical analysis

Concentrations of total As and Fe in the water samples were

determined using an atomic absorption spectrophotometer

(AAS; Shimadzu AA-660G) with a graphite furnace atom-

izer (GFA-4B) according to standard procedures and using

commercial laboratory standards. The detection limit of the

AAS was 1 lg/l. The instrument was linearly calibrated

with custom single element standards (Kanto Chemicals

Co., Inc., Japan) for As and Fe, yielding r2 of 0.9986 and r2

of 0.9988, respectively. Calibration was made using a blank

solution (0.01 mol/l HNO3) and standard stock arsenic and

iron solutions (100 and 1000 mg/l, respectively). For

arsenic and iron analyses, groundwater samples were diluted

in 1.3 mol/1 HNO3 to bring the concentrations within the

range of the calibration. Samples with higher concentrations

were reanalyzed after further dilution.

Each sample was injected five to six times for each

element. Acceptance criteria for each run were \10%

variation in the consistency standard. Precision was B5%,

based on repeated measurements of standard solutions with

known concentrations. All analyses were made at Shimane

University, Japan.

Environ Earth Sci (2010) 60:1303–1316 1305

123

Page 4: Geochemical study of arsenic and other trace elements in groundwater and sediments of the Old Brahmaputra River Plain, Bangladesh

Analytical procedures for sediment samples

Sampling

Sediment cores were taken at six locations (My1, Parail

Village; My2, Najirabad Village; My3, Ghagra Village;

My4, Youth Training Center (YTC); My5, Bangladesh

Agricultural University (BAU) and My6, Mymensingh

City; Fig. 1), using the rotary drill percussion reverse cir-

culation (PRC) method. Depths of the cores ranged

between 21 and 31 m. About 200 g of wet sediments were

collected for each sample and packed in Ziploc� bags.

A total of 153 sediment samples were collected from the

cores, and ORP was measured with a Horiba U-23 ORP

meter. The samples were stored at 4�C in a cooler box for

transport to the laboratory. These cores were logged

and photographed, and the layers were visually classified

during sampling.

A total of 75 sediment samples were selected from

Cores My1, My2, and My6 for analysis. Approximately

50 g of each sediment sample were oven-dried at 110�C for

24 h. The dried samples were then ground for 20 min in an

automatic agate mortar and pestle grinder.

X-ray fluorescence analysis

Selected major oxide [Fe2O3* (total iron is expressed as

Fe2O3*), TiO2, CaO, and P2O5], total sulfur (TS), and trace

element (As, Pb, Zn, Cu, Ni, Cr, V, and Sr) concentrations

were determined by X-ray fluorescence (XRF) at Shimane

University using a RIX-2000 spectrometer (Rigaku Denki

Co. Ltd., Japan) equipped with a Rh-anode X-ray tube. All

analyses were made on pressed powder briquettes

(prepared using a force of 200 kN for 60 s) following the

method of Ogasawara (1987). Average errors for all ele-

ments are less than ±10% relative. Analytical results for

USGS standard SCo-1 (Cody Shale) were acceptable

compared to the proposed values of Potts et al. (1992).

Carbon dating

Conventional radiocarbon dating of two selected samples

from Cores My2 and My4 was carried out at the Geosci-

ence Laboratory of Nagoya University, Japan, using

accelerator mass spectrometry (AMS). Reference materials

of known ages were analyzed to verify the accuracy of the

results. The ages were calculated as 14C year BP, corrected

for 13C, and expressed at the ±1r level of analytical

confidence.

Results

Physico-chemical properties of the water samples

The pH values of groundwater were approximately neutral,

within a range of 6.4–7.4 (Table 1). Electrical conductivity

(EC) ranged from 303 to 550 lS/cm with an average of

370 lS/cm. COD values ranged between 1 and 20 mg/l,

but COD in two samples were under the detection limit.

DO values varied from 0.8 to 1.2 and 1.1 to 1.8 mg/l in the

shallow and deep aquifers, respectively. High DO values

may be related to pump aeration during sampling, as

reported by Nickson et al. (2000) for groundwater in

Bangladesh. The concentrations of DO suggest that the

waters from the deep tube wells were more oxic than those

Table 1 Groundwater quality of tube wells in Mymensingh, Bangladesh

Well

no.

Depth

code

Depth

(m)

pH EC

(lS/cm)

DO

(mg/l)

NH4?

(mg/l)

PO43-

(mg/l)

COD

(mg/l)

NO3-

(mg/l)

Ca2?

(mg/l)

FeT

(mg/l)

AsT

(lg/l)

Tw1 S 12 7.4 303 1.1 0.3 0.5 3 nd 50 5.8 57

Tw2 S 15 6.9 325 0.9 0.1 0.7 1 nd 50 4.9 9

Tw3 S 56 7.4 361 1.2 nd 0.2 2 nd 5 0.7 8

Tw4 S 55 6.4 319 1.2 nd 0.2 3 nd 30 1.7 12

Tw5 D 84 7.1 315 1.2 nd 0.3 nd nd nd 0.3 nd

Tw6 S 12 6.8 491 1.0 3.0 5.0 5 nd 50 10.8 159

Tw7 S 15 6.6 550 1.2 3.0 0.7 20 nd 50 12.9 251

Tw8 S 53 7.1 335 1.0 0.2 0.3 1 nd 50 4.3 18

Tw9 D 137 7.3 400 1.8 nd 0.6 nd nd 30 0.2 nd

Tw10 S 52 7.2 372 1.0 0.3 0.5 5 nd 50 7.0 26

Tw11 S 34 7.2 341 0.8 0.5 1.0 8 nd 50 5.6 27

Tw12 D 84 7.2 353 1.1 0.3 0.5 1 nd 15 0.7 nd

Tw13 D 137 7.4 340 1.1 nd 0.2 2 nd 25 0.3 nd

S shallow; D deep; nd not detected

1306 Environ Earth Sci (2010) 60:1303–1316

123

Page 5: Geochemical study of arsenic and other trace elements in groundwater and sediments of the Old Brahmaputra River Plain, Bangladesh

from the shallow wells. Ammonia concentrations in some

water samples from the shallow aquifers exceeded 3 mg/l.

Nitrate concentrations were not detected in any samples

from the deep and shallow wells. Phosphorus concentra-

tions varied from 0.2 to 5 mg/l and calcium ranged from 5

to 50 mg/l.

Arsenic and iron concentrations in groundwater

Dissolved As concentrations (as determined by AAS) of

shallow tube well waters in Mymensingh show significant

variation, ranging from 8 to 251 lg/l (12–56 m depths)

with an average 63 lg/l (Table 1). However, As con-

centrations were under the detection limit (1 lg/l) in all

deep aquifer samples Tw5, Tw9, Tw12, and Tw13 (84,

137, 84, and 137 m deep, respectively). Elevated As

contents of 251 lg/l (Tw7; 15 m) and 159 lg/l (Tw6;

12 m) were observed at Ghagra Village and 57 lg/l

(Tw1; 12 m) at Parail Village. Total As concentrations

in the remainder of the samples were moderately low

(8–27 lg/l).

Dissolved Fe concentrations show greater contrast

between the wells, ranging from 0.2 to 12.9 mg/l (Table 1).

The two highest Fe concentrations (12.9 and 10.8 mg/l)

were recorded in water samples from shallow tube wells

(Tw7 and Tw6, respectively). Relatively high Fe concen-

trations (7.0, 5.8, and 5.6 mg/l) were also found in samples

from depths of 52 m (Tw10), 12 m (Tw1), and 34 m

(Tw11) in the upper aquifer. The highest Fe concentration

(12.9 mg/l) was found in sample Tw7 from a 15-m tube

well, which also had the highest As content (251 lg/l).

Total Fe concentrations in all deep aquifer wells (84–

137 m depths) were relatively low, ranging from 0.2 to

0.7 mg/l (avg. 0.4 mg/l).

Sediment characteristics

Arsenic-affected areas in Bangladesh are mainly confined

to the Holocene alluvial aquifers at shallow and interme-

diate depths (BGS and DPHE 2001; Mukherjee and

Bhattacharya 2001), as reported by Acharyya et al. (2000)

for West Bengal, India. The Brahmaputra sediments in

Bangladesh are rich in quartz, mica (both muscovite and

biotite), feldspar, calcite, and dolomite (Brammer 1996;

BGS and DPHE 2001). Characteristics of the core sedi-

ments (My1, My2, My3, My4, My5, and My6) from sur-

face to depths of 21, 25, and 31 m are shown in Fig. 2. The

occurrence of peat and peaty clay layers, which are spread

widely in the Mymensingh region, are a possible boundary

between the Holocene and Pleistocene beds (Umitsu 1993;

Brammer 1996). Comparing the characteristics of the core

samples to Umitsu (1993) and Brammer (1996), the upper

parts of the cores (about 0–13 m) are supposed to be

Holocene sediments, and the lower parts of the cores (from

13 m depths) are Pleistocene sediments. The stratigraphic

sequence consists of silty clays and clays at top, passing

downward into grayish to yellowish clays and fine- to

medium-grained sands. The uppermost layers commonly

consist of yellowish silty clays. Color typically changes

from yellowish to gray and dark gray within the shallow

depths (0–13 m).

The fine-grained sediments prevalent in all six cores

within the shallow depths (0–13 m) are relatively rich in

organic matter, representing overbank facies. Clay layers

occur from 1 to 5 and 11 to 30 m in Core My1; from 2 to 5,

6 to 9, and 10.5 to 17 m in My2; from 2 to 6 and 10.5 to

21 m in My3; from 1 to 6, 7 to 10.5, 12 to 20, and 22 to

25 m in My4; from 2 to 7, 8 to 10, and 11.5 to 27 m in

My5; and 1 to 2, 5 to 8, and 13.5 to 20 m in My6. In Core

ORP, oxidation reduction potential

grayish clay

grayish fine sand

grayish silt

yellowish clay

bluish clay

yellowish silty clay

peat

grayish coarse sand

peaty clay

14C age 4315BP

reddish iron oxide

yellowish fine sand

black carbonate iron

yellowish silt

wood fragment

My4

4315BP

My5

+250

ORP+106(mV)

0

5

10

15

20

25

30

Depth(m) My1 My2

-152

+205

-150

My3

+138

-88

+210

Holocene

Madhupur Fm(Pleistocene)

My6Fig. 2 Columnar sections of

core sediments in Mymensingh

(My1, My2, My3, My4, My5,

and My6)

Environ Earth Sci (2010) 60:1303–1316 1307

123

Page 6: Geochemical study of arsenic and other trace elements in groundwater and sediments of the Old Brahmaputra River Plain, Bangladesh

My6, thick grayish silt sections were present at 2 and 10 m

below surface in the anoxic zone. The uppermost 3 m of

the sediments in all cores are oxidized (ORP from ?30 to

?220 mV), and ORP values gradually become negative

with increasing depth up to 13 m of the sediments (ORP

from –35 to –180 mV). These low ORP values indicate

anoxic conditions in the Holocene sediments. However,

ORP values become positive from 13 m depths (ORP from

?40 to ?260 mV), indicating the lower parts of the core

sediments were under oxic conditions.

Yellowish fine sands are more abundant in the aqui-

fers at depths of about 17–31 m in My2 and 20–31 m in

My4. Yellowish and bluish clays are abundant in other

core sediments of the Pleistocene bed. Black peat and

peaty clay layers occur at a depth of about 9–10.5 m in

My2 and My3, from 10.5 to 12 m in My4, from 11 to

11.5 m in My5, and at 13 to 13.5 m in My6. The sec-

tions immediately above and below the peat layers were

mainly grayish and yellowish clays, respectively. Oxide

and carbonate concretions rich in ferruginous minerals

(e.g., Fe-coated quartz and Fe oxide) were observed

mainly in the clay and silty layers. Exceptionally, wood

fragments were found in the gray silt sections in Core

My6.

Major and trace elements

Elemental compositions of the core sediments analyzed by

XRF are summarized in Table 2. Average upper conti-

nental crust (UCC) from Taylor and McLennan (1985) and

all lithotypes from Cores My1, My2, and My6 are included

in Tables 2 and 3, respectively, for comparison.

Table 2 The range, the mean value, and the standard deviation of the major and trace elements of Cores My1, My2, and My6 in Mymensingh

Core/area Trace elements (ppm) Major oxides and TS (wt%)

As Pb Zn Cu Ni Cr V Sr TiO2 Fe2O3 CaO P2O5 TS

My1 (n = 27)

Range 3–14 17–37 37–165 2–53 22–69 71–117 77–190 51–193 0.50–0.84 4.04–9.24 0.82–2.47 0.04–0.30 0.03–0.07

Mean 7 22 75 29 48 95 137 129 0.70 6.61 1.47 0.12 0.04

SD 3.1 4.0 28.2 12.2 13.7 13.2 34.2 40.7 0.08 1.69 0.46 0.06 0.01

My2 (n = 23)

Range 4–65 17–48 32–124 6–118 23–103 65–130 45–246 42–180 0.30–0.88 2.45–10.30 0.70–2.46 0.04–0.22 0.03–2.98

Mean 12 25 75 36 54 99 145 123 0.65 6.35 1.48 0.10 0.18

SD 12.7 7.9 30.5 27.2 23.9 15.5 58.9 47.1 0.18 2.47 0.50 0.05 0.61

My6 (n = 25)

Range 3–733 16–40 25–129 7–91 28–206 60–143 55–594 14–182 0.10–1.02 3.38–15.52 0.64–2.38 0.05–0.16 0.03–6.51

Mean 42 25 82 38 70 105 185 102 0.76 7.98 1.31 0.10 0.38

SD 144.9 5.7 28.0 18.1 34.1 17.1 95.5 46.0 0.21 2.36 0.48 0.03 1.32

UCC

Mean 2 20 71 25 20 35 60 350 0.50 5.00 4.20 0.16 na

UCC Upper continental crust (Taylor and McLennan 1985), na not available

Table 3 Concentrations of trace elements (mean ± SD) in all lithotypes from Cores My1, My2, and My6, Mymensingh

Core Lithology Trace elements (ppm)

As Pb Zn Cu Ni Cr V Sr

My1 Sand (n = 6) 4 ± 0.9 18 ± 0.8 48 ± 6.4 12 ± 5.4 29 ± 3.8 94 ± 12.6 96 ± 11.6 184 ± 8.6

S, Clay (n = 21) 8 ± 2.8 24 ± 3.7 83 ± 27.2 33 ± 8.6 53 ± 10.0 96 ± 13.6 149 ± 28.6 114 ± 31.2

My2 Sand (n = 7) 4 ± 0.7 19 ± 1.0 39 ± 7.1 9 ± 2.0 26 ± 2.5 87 ± 16.9 71 ± 27.4 164 ± 11.0

S, Clay (n = 14) 11 ± 3.2 27 ± 6.3 90 ± 22.7 43 ± 15.5 63 ± 13.9 104 ± 10.8 170 ± 29.2 110 ± 47.1

P, PC (n = 2) 46 ± 26.9 37 ± 15.1 96 ± 4.7 89 ± 41.0 94 ± 13.3 104 ± 24.9 224 ± 31.0 76 ± 20.9

My6 Sand (n = 5) 5 ± 1.3 21 ± 1.7 57 ± 20.9 17 ± 10.9 42 ± 11.8 88 ± 19.4 115 ± 49.8 110 ± 29.9

S, Clay (n = 18) 12 ± 4.8 26 ± 4.7 91 ± 22.4 42 ± 9.5 69 ± 14.2 108 ± 12.1 181 ± 31.1 104 ± 47.8

P, WF (n = 2) 411 ± 455.7 28 ± 17.0 66 ± 58.0 53 ± 53.8 157 ± 69.3 124 ± 26.6 397 ± 278.8 54 ± 56.6

S silt; P peat; PC peaty clay; WF wood fragment

1308 Environ Earth Sci (2010) 60:1303–1316

123

Page 7: Geochemical study of arsenic and other trace elements in groundwater and sediments of the Old Brahmaputra River Plain, Bangladesh

The sediments in Core My1 contain less As (avg. 7 ppm;

range 3–14 ppm), Pb (avg. 22 ppm; range 17–37 ppm), Cu

(avg. 29 ppm; range 2–53 ppm), Ni (avg. 48 ppm; range 22–

69 ppm), Cr (avg. 95 ppm; range 71–117 ppm), and V (avg.

137 ppm; range 77–190 ppm) than in either My2 or My6

(Table 2). The highest concentrations of Zn and Sr (165 and

193 ppm) and the highest average value of Sr (129 ppm)

occur in My1. CaO and P2O5 contents are elevated, in the

range of 0.82–2.47 and 0.04–0.30 wt% (avg. 1.47 and

0.12 wt%), respectively. Ca and Sr have similar behaviors in

sediments, and therefore, both CaO and Sr show surface

enrichment in the core samples. Sediments in this core contain

the least TS (ranging from 0.03 to 0.07 wt%; avg. 0.04 wt%)

among the three cores.

On average the sediments in Core My2 contain 12 ppm As

(range 4–65 ppm), 75 ppm Zn (range 32–124 ppm), 54 ppm

Ni (range 23–103 ppm), and 99 ppm Cr (range 65–130 ppm).

Average concentrations of Pb and Cu are similar to those in

My6. However, the highest concentrations of Pb (48 ppm) and

Cu (118 ppm) were found in this core sample. Black peat

sample My2-9 had the highest As content (65 ppm), along

with the highest CaO (2.46 wt%) and TS (2.98 wt%) at 9–

9.5 m depth. The concentrations of TiO2 and P2O5 in My2

range from 0.30 to 0.88 and 0.04 to 0.22 wt%, respectively.

CaO contents vary between 0.70 and 2.46 wt% in My2 and

yield the highest average value of 1.48 wt%. Abundances of

Fe2O3 and TS show significant variation in My2, with ranges

of 2.45–10.30 and 0.03–2.98 wt%, respectively.

Abundances of trace metals in the sediments of Core My6

are higher than in those from My1 and My2. The highest

average values of As (42 ppm), Pb (25 ppm), Zn (82 ppm),

Cu (38 ppm), Ni (70 ppm), and Cr (105 ppm) were all

recorded in sediments from this core. Vanadium abundances

range between 55–594 ppm, and the average (185 ppm) is

also highest among the three sample sets. Anomalously high

As (733 ppm), Ni (206 ppm), and V (594 ppm) concentra-

tions were found in a wood fragment sample (My6-11;

Table 2). Black peat sample My6-15 also had a higher con-

centration of As (89 ppm) at 13–13.5 m depth. Sediments in

this core contain the lowest Sr (avg. 102 ppm; range 14–

182 ppm) among the three cores. Overall concentrations of

TiO2, CaO, and P2O5 in the sediments range from 0.10 to 1.02,

0.64 to 2.38, and 0.05 to 0.16 wt%, respectively. Average

concentrations of P2O5 are almost identical in all cores. Fe2O3

and TS contents vary considerably, with ranges of 3.38–15.52

and 0.03–6.51 wt%, respectively. The highest Fe2O3 and TS

values (15.52 and 6.51 wt%) were observed in the same wood

fragment sample (My6-11) at a depth of 10.8 m.

Sediment dating

Organic-matter-rich black clays were used for dating. The

conventional radiocarbon ages were 1985 ± 30 years BP

at 9–9.5 m in Core My2 and 4315 ± 35 years BP at 10.5–

11 m in My4. As Umitsu (1985, 1993) described, the

radiocarbon ages of the peat layers in the Ganges plain

(5 m depth) and Sylhet basin (9 m depth) of Bangladesh

were dated 3230 ± 110 and 4180 ± 120 years BP,

respectively, showing similar ages to those of this study.

Discussion

Characteristics of As release in groundwater

Relationships between (a) As and Fe, (b) As and COD, and (c)

Fe and COD in groundwater are shown in Fig. 3. A strong

positive relationship exists between As and Fe (r2 = 0.802;

p = 0.001; Fig. 3a) in the water samples. This correlation is

consistent with many reports from the Ganges and Meghna

regions (e.g., Anawar et al. 2003; Tareq et al. 2003). Arsenic is

also positively correlated with COD (r2 = 0.704; p = 0.004;

Fig. 3b). A moderate positive relationship was found between

Fe and COD (r2 = 0.567; p = 0.007; Fig. 3c) in groundwa-

ters from the Mymensingh Sadar.

In this study, the highest As value observed (251 lg/l)

was associated with the highest COD (20 mg/l), indicating

decomposition of organic matter in the shallow aquifer.

The combination of these conditions with the highest

Fe contents (12.9 mg/l) supports reduction of Fe

(oxy)hydroxides (Nickson et al. 1998, 2000; Zheng et al.

2004), because organic matter plays an indispensable role

in the release of As (McArthur et al. 2001, 2004).

As and Fe distributions with well depth

Concentrations of both As and Fe in the shallow wells are

above the WHO (10 lg/l and 0.3 mg/l, respectively) and

Bangladesh limits (50 lg/l and 1 mg/l, respectively) for

drinking water. No linear relationship exists between well

depth and As and Fe contents in the waters, but concen-

trations of both elements clearly tend to be lower at greater

depth (Fig. 4). Additional As and Fe data from the

respective areas (Mymensingh Sadar; BGS and DPHE

2001; n = 11) are also shown in Fig. 4 for comparison.

Concentrations are broadly similar to those determined in

this study. Maximum As concentrations occur at depths

between 12 and 40 m, whereas samples deeper than 60 m

are arsenic-poor (\10 lg/l) and are below the WHO limit

for potable water. Moreover, the overall decrease in As

concentration with increasing aquifer depth is consistent

with other studies (Acharyya et al. 1999; Chowdhury et al.

1999). Arsenic and Fe abundances in the waters of the

upper aquifer vary considerably (8–251 lg/l and 0.7–

12.9 mg/l, respectively). These variations can be attributed

to varying redox states at shallow depths within the aquifer.

Environ Earth Sci (2010) 60:1303–1316 1309

123

Page 8: Geochemical study of arsenic and other trace elements in groundwater and sediments of the Old Brahmaputra River Plain, Bangladesh

All shallow tube wells of the aquifer contain high-As

concentrations, but waters from deep tube wells are As-free or

As-poor. Bhattacharyya et al. (2003) also reported that As

concentrations tend to be greatest in the intensively exploited

shallow aquifer (10–50 m), whereas waters from deeper

aquifers up to 150 m tend to have much lower As contents.

This contrast in As concentrations between groundwater from

shallow and deep aquifers can be explained by resorption of

As onto residual FeOOH (McArthur et al. 2004) and by the

nature of the aquifer sediments (Zheng et al. 2004).

Inter-element relationship in core samples

Table 4 shows the correlation matrix for elements in the

sediments. Strong positive relationships were observed

between the concentrations of Fe2O3 and As, Pb, Zn, Cu,

Ni, V, and TiO2, and between V and As, Pb, Zn, Cu, Ni,

and TiO2 in Core My1 (Table 4). Non-significant correla-

tions for Sr, CaO, and P2O5 suggest a different source for

these elements in the sediments. In Core My2, Fe2O3

concentrations are positively correlated with As, Pb, Zn,

Cu, Ni, Cr, V, and TiO2 (Table 4). Significant associations

of V with Pb, Zn, Cu, Ni, and Cr were observed in the

sediments. Negative or poor relationships for Sr, CaO, and

P2O5 reflect a differing control for these elements. Strong

positive relationships of Fe2O3 with As, Ni, Cr, and V, of

TS with As, Ni, V, and Fe2O3, and of V with As, Ni, and Cr

were observed in the sediments of Core My6. Conversely,

concentrations of Sr, CaO, and P2O5 again show no rela-

tionships, indicating different behavior of these elements.

(a) (b)

Fig. 4 a, b Relationship

between well depth and As and

Fe contents in Mymensingh

groundwater. Filled symbols are

the data from this study; open

symbols are from BGS and

DPHE (2001). The WHO and

Bangladesh As (10 and 50 lg/l,

respectively) and Fe (0.3 and

1 mg/l, respectively) for potable

water are indicated by the

dashed lines

(a)

(c)

(b)

Fig. 3 a–c Relationships

between Fe, COD, and As, and

COD and Fe in groundwater,

Mymensingh

1310 Environ Earth Sci (2010) 60:1303–1316

123

Page 9: Geochemical study of arsenic and other trace elements in groundwater and sediments of the Old Brahmaputra River Plain, Bangladesh

Relationships among the geochemical data show that

metallic elements are strongly or positively correlated with

total Fe (Table 4), suggesting that Fe2O3 may exert a major

role in controlling the metal concentrations in the core

sediments, as discussed by Singh et al. (2005) for river

sediments in India. The strong positive correlation matrices

of a suite of metals (As, Pb, Zn, Cu, and Ni in My1; Pb, Zn,

Cu, Ni, and Cr in My2; As, Ni, and Cr in My6) with V

Table 4 Correlations between the elements in sediments of three cores

As Pb Zn Cu Ni Cr V Sr TiO2 Fe2O3 CaO P2O5 TS

Core My1 (n = 27)

As 1.00 0.47 0.28 0.48 0.69 0.34 0.65 -0.82 0.55 0.69 -0.74 -0.28 0.04

Pb 1.00 0.85 0.79 0.74 0.37 0.70 -0.51 0.59 0.67 -0.62 -0.32 0.35

Zn 1.00 0.84 0.72 0.38 0.68 -0.23 0.45 0.70 -0.36 0.03 0.26

Cu 1.00 0.86 0.43 0.89 -0.54 0.78 0.84 -0.62 -0.19 0.09

Ni 1.00 0.48 0.91 -0. 70 0.75 0.90 -0.75 -0.31 0.11

Cr 1.00 0.63 -0.18 0.64 0.50 -0.12 -0.18 0.21

V 1.00 -0.69 0.86 0.95 -0.69 -0.30 0.03

Sr 1.00 -0.66 -0.69 0.96 0.51 0.17

TiO2 1.00 0.72 -0.62 -0.38 0.10

Fe2O3 1.00 -0.71 -0.22 -0.06

CaO 1.00 0.58 0.16

P2O5 1.00 -0.02

TS 1.00

Core My2 (n = 23)

As 1.00 0.42 0.39 0.56 0.61 0.11 0.55 -0.50 -0.08 0.42 0.21 0.08 0.93

Pb 1.00 0.68 0.95 0.90 0.65 0.77 -0.44 0.41 0.31 -0.36 -0.23 0.11

Zn 1.00 0.75 0.80 0.53 0.83 -0.27 0.57 0.79 -0.08 0.15 0.15

Cu 1.00 0.95 0.63 0.85 -0.46 0.43 0.43 -0.22 -0.09 0.26

Ni 1.00 0.64 0.92 -0.60 0.54 0.58 -0.29 -0.17 0.33

Cr 1.00 0.71 -0.32 0.64 0.41 -0.28 -0.09 -0.14

V 1.00 -0.66 0.71 0.79 -0.32 -0.10 0.26

Sr 1.00 -0.55 -0.50 0.69 0.50 -0.31

TiO2 1.00 0.63 -0.58 -0.14 -0.32

Fe2O3 1.00 -0.18 -0.01 0.25

CaO 1.00 0.70 0.41

P2O5 1.00 0.22

TS 1.00

Core My6 (n = 25)

As 1.00 -0.27 -0.40 -0.18 0.86 0.48 0.90 -0.42 -0.67 0.69 0.04 -0.11 0.99

Pb 1.00 0.48 0.80 0.10 0.28 -0.05 -0.19 0.37 0.12 -0.26 -0.55 -0.21

Zn 1.00 0.78 -0.01 0.26 -0.17 0.43 0.43 0.09 0.26 -0.08 -0.36

Cu 1.00 0.28 0.41 0.07 0.03 0.35 0.25 0.06 -0.40 -0.09

Ni 1.00 0.75 0.92 -0.40 -0.40 0.78 0.01 -0.30 0.87

Cr 1.00 0.75 -0.39 0.14 0.76 -0.22 -0.31 0.46

V 1.00 -0.55 -0.31 0.90 -0.14 -0.27 0.88

Sr 1.00 -0.07 -0.51 0.83 0.59 -0.39

TiO2 1.00 -0.04 -0.37 -0.15 -0.70

Fe2O3 1.00 -0.15 -0.27 0.67

CaO 1.00 0.61 0.09

P2O5 1.00 -0.12

TS 1.00

Bold text highlights strong correlations

Environ Earth Sci (2010) 60:1303–1316 1311

123

Page 10: Geochemical study of arsenic and other trace elements in groundwater and sediments of the Old Brahmaputra River Plain, Bangladesh

suggest the possibility of formation of complexes with

organic matter. This is consistent with the study conducted

by Tribovillard et al. (1994) of sediments in the United

Kingdom. The negative or poor correlations of Sr, CaO,

and P2O5 in all cores reflect a different control for these

elements. Strong positive correlations between Sr and CaO

were found in all cores and are most likely related to their

similar geochemical behavior. With some exceptions,

inter-relationships between As, Pb, Zn, Cu, Ni, and Cr

show that all are significantly correlated with each other,

indicating a common source or a similar enrichment

mechanism in the sediments.

Vertical distributions of major and trace elements

The vertical distribution of As in Core My1 is similar

to that of Fe2O3, although from 15 m downward the

distributions vary (Fig. 5). Abundances of As, Pb, Zn, Cu,

Ni, Cr, and Fe2O3 in the silts (0–0.5 m) and clays (0.5–5

and 11–30 m) are greater than those in the sands (5–11 and

30–30.5 m; Table 3). In contrast, CaO and Sr contents tend

to have greater concentrations in sands than in the clays.

A suite of trace metals (Pb, Zn, Cu, Ni, and Cr) are enri-

ched in the bluish clays at a depth of 15 m, possibly due to

the migration of soluble metals to this depth and adsorption

onto clay particles. These particles have a high capacity for

adsorbing or retaining trace metals, as found by Yan et al.

(2000) in a clay-rich aquitard sequence in Saskatchewan,

Canada, and by Peltier et al. (2003) in Dead Stick pond

sediments in Chicago, IL, USA.

The vertical profile of As in Core My2 is similar to that

of iron, though from 16 m upward it shows a less consis-

tent trend (Fig. 6). Concentrations of As, Pb, Zn, Cu, and

Ni show peaks at 4–5 m, 9–10 m, and 13–16 m (Fig. 6).

Fig. 5 Vertical distribution of

As, Fe2O3, Pb, Zn, Cu, Ni, Sr,

and CaO in Core My1,

Mymensingh

Fig. 6 Vertical distribution of

As, Fe2O3, Pb, Zn, Cu, Ni, Sr,

and CaO in Core My2,

Mymensingh

1312 Environ Earth Sci (2010) 60:1303–1316

123

Page 11: Geochemical study of arsenic and other trace elements in groundwater and sediments of the Old Brahmaputra River Plain, Bangladesh

The trends of these elements are similar to Fe, which shows

peak concentrations at the same depths. The contents of

these trace metals are thus probably controlled by the

presence of Fe oxides (Preda and Cox 2002), as Fe

(oxy)hydroxides have high affinity for trace metals (Tessier

et al. 1994). Trace metals (As, Pb, Zn, Cu, Ni, and Cr) are

enriched in the silts, clays, peat, and peaty clay at 0–0.5

and 2–17 m relative to the sands at 0.5–2 and 17–31 m

(Fig. 6). Padmalal et al. (1997) and Singh et al. (2005)

reported trace metal-enrichment in fine-grained clays, as

they possess higher surface areas than coarser grains. CaO

and Sr show similar trends in Core My2.

Moderate trends in the depth-wise variation of trace

elements are seen in Core My6 (Fig. 7). Arsenic and Fe

exhibit almost similar vertical distributions, with strong

spikes in abundances at 10.8 and 13 m (Fig. 7). The

highest concentrations of TS were also found at the same

depths. This association suggests co-precipitation of these

elements with Fe sulfide minerals or adsorption on to Fe

sulfides in an anoxic environment, as proposed by Huerta-

Dıaz et al. (1998) for lake sediments in Canada and by

Ayyamperumal et al. (2006) for river sediments on the

southeast coast of India. Arsenic is present in the crystal

structure of many sulfide minerals as a substitute for S. As

the geochemistry of As follows closely that of S, the

greatest concentrations of the element tend to occur in

sulfide minerals, of which pyrite is the most abundant

(Smedley and Kinniburgh 2002). The vertical distributions

of Ni, Cr, and V also show distinct peaks at 10.8 m and Pb,

Cu, and Ni at 13 m (Fig. 7). These elements have a strong

affinity for organic matter, Fe oxides and clay minerals

(Tessier et al. 1994; Tribovillard et al. 1994; Singh et al.

2005). Therefore, the vertical profiles of Pb, Cu, Ni, Cr,

and V show similarities throughout the core. The concen-

tration of Sr decreases sharply at 10.8 m depth. From the

bottom of the core to the surface, CaO and Sr increase

irregularly upward and reach their highest values

(2.38 wt% and 182 ppm) in the uppermost sample. The

sandy character of the sediments reduces their adsorptive

properties, leading to a reduction of metal levels in sedi-

ments from the My6 core (Fig. 7).

In this study, coincident peaks for some trace metals

(Pb, Zn, Cu, Ni, Cr) at approximately the same depth (My1,

15 m; My2, 9–16 m; My6, 10–13 m) suggest the influence

of post-depositional effects, such as reduction of sulfides,

precipitation of metallic sulfides under anoxic conditions,

or re-precipitation of trace metals on Fe (oxy)hydroxide

coatings (Millward and Moore 1982). Moreover, because

there are no major industries in the study areas and no

known natural point sources for metal enrichment in the

sediments, post-depositional contamination of the cores

from anthropogenic pollution is improbable. Therefore, the

trace metals present are likely incorporated in the clay

matrix (e.g., from weathering of bedrock) and in the fine-

grained sediments and have been influenced by post-

depositional diagenetic remobilization.

Characteristics of As in sediments

In this study, the clays in Cores My1, My2, and My6

contain 5–14, 6–15, and 7–20 ppm As, respectively, while

contents in the sands (3–6, 4–6, and 3–6 ppm, respectively)

are one-half to one-third lower. These results are compa-

rable to data for samples from the Ganges Delta in Jessore

region (4–18 ppm in clays; 3–7 ppm in sands; Yamazaki

et al. 2003). The Jessore region is one of the most

As-contaminated areas of Bangladesh. This contrasting

distribution of As between clays and sands in the

Ganges–Meghna–Brahmaputra (GMB) River system is

also consistent with other reports (e.g., Yamazaki et al.

Fig. 7 Vertical distribution of

As, Fe2O3, Cu, Ni, Cr, V, Sr,

and CaO in Core My6,

Mymensingh

Environ Earth Sci (2010) 60:1303–1316 1313

123

Page 12: Geochemical study of arsenic and other trace elements in groundwater and sediments of the Old Brahmaputra River Plain, Bangladesh

2000; Anawar et al. 2002). Arsenic may be scavenged by

clay minerals (Yan et al. 2000), as they have high surface

areas and negative surface charge (Padmalal et al. 1997;

Singh et al. 2005).

The core sediments in the Mymensingh region are sig-

nificantly enriched in As (Tables 2, 3) compared with the

UCC (avg. 2 ppm As from Taylor and McLennan 1985;

avg. 5 ppm As from Rudnick 2005). Excluding wood

fragment, the maximum As abundances observed in Cores

My2 and My6 (Table 2) are 10–20 times higher than the

baseline concentrations of As in sediments (3–10 ppm;

Smedley and Kinniburgh 2002). These elevated As con-

centrations in the core sediments are most probably related

to diagenetic redistribution controlled by the redox

behavior of Fe (oxy)hydroxide phases (Sullivan and Aller

1996). McArthur et al. (2001) and Ravenscroft et al. (2001)

reported that the presence of arsenic-bearer phases, i.e.,

Fe (oxy)hydroxides, is one of the major hosts of As in

the Ganges sediments. Moreover, re-adsorption onto or

co-precipitation with Fe oxides and hydroxides and sub-

stitution in Fe sulfide minerals could also contribute to As

levels in these sediments, as suggested by the strong cor-

relation between As and Fe2O3 in Cores My1 (r = 0.69)

and My6 (r = 0.69) and between As and TS in My2

(r = 0.93) and My6 (r = 0.99; Table 4).

The existence of peat bands at around 9–13 m depth

(ORP -60 to -170 mV) in the study areas represents a

reducing and marshy wetland in an anoxic environment. In

the tropics, peat evolves in situ by accumulation of fallen

trees, leaves, and branches (Markgraf 1989). Ishiga et al.

(2000) reported that the fluvial sediments supplied by the

Ganges are rich in organic matter, derived from higher

plants that grew in a humid and warm climate. Such natural

organic matter (peat and peaty sediments, which spread

widely in the Bengal lowland after ca. 5,000 years BP;

Umitsu 1993) is one of the potential sources of As con-

tamination of groundwater in Bangladesh (Ishiga et al.

2000; Yamazaki et al. 2000).

In this study, the black peat in My2 and peat and wood

fragment in My6 have the highest abundances of As

(65 ppm in My2; 89 and 733 ppm in My6, respectively),

TS (2.98 wt%; 1.76 and 6.51 wt%, respectively) and Fe2O3

(9.29 wt%; 8.80 and 15.52 wt%, respectively). Decom-

posed wood fragments (ORP -110 mV; at 10.8 m in My6)

as a source of carbon and organic matter in silts with

detectable odor reflect simultaneous fermentation. Arsenic

might have been deposited onto fine-grained organic-rich

sediments that were preferentially deposited under low

energy conditions in the Ganges and Meghna floodplain

during the Mid-Holocene sea-level rise (Acharyya et al.

2000). Appearance of these peat bands in the Holocene

sediments accelerated the supply of organic matter, as

evidenced by higher COD value (about 20 mg/l; Table 1)

in Mymensingh groundwater. High concentrations of

NH4? (about 3 mg/l; Table 1) in shallow tube wells are

likely related to the burial of peat. This result is consistent

with the study conducted by McArthur et al. (2001) of

sediments in the Ganges regions. Peat occurs extensively

beneath the As-affected areas of Samta and Deuli villages

in southwestern Bangladesh, and peat contains high con-

centrations of As (50–260 ppm), within which microbial

degradation is high (Ishiga et al. 2000; Yamazaki et al.

2000). Akai et al. (2003) also reported that As was dis-

solved into water when reducing conditions were achieved

by microbial activity. Thus, the peat is the driver for

reduction of Fe (oxy)hydroxides in the Old Brahmaputra

plain, as observed by McArthur et al. (2004) in the Ganges

plain. Widespread occurrence of peat and peaty sediments,

normally rich in As, has possibly enriched the transgressing

seawater (Umitsu 1993) and is inferred to be one of the

main sources of As in the Bengal basin.

Conclusions

Our present results show that groundwaters from the

shallow aquifers in the Mymensingh region are character-

ized by approximately neutral pH, moderate COD, and

high concentrations of As and Fe. Arsenic concentrations

in groundwaters are highly variable (8–251 lg/l), and

many exceed the maximum permissible limits of the WHO

(10 lg/l) and Bangladesh (50 lg/l) for drinking water.

Strong positive correlation between As and Fe in the water

samples suggests that dissolved As may be adsorbed to Fe

(oxy)hydroxides.

The cores in the study areas pass downward from yel-

lowish silty clays into grayish to yellowish clays and sands.

The shallow aquifer sediments of the Brahmaputra plain

are in reducing conditions, as indicated by negative ORP

values in the upper part core samples. Abundances of As

and other trace elements (Pb, Zn, Cu, Ni, Cr, and V) and

Fe2O3 are greater in silts and clays compared to those in the

sands. Anomalously high concentrations of As, V, Fe2O3,

and TS were found in a blackish wood fragment sample at

11 m depth. Concentrations of these elements are also

higher in black peat and peaty sediments at depths between

9 and 13 m, possibly because these consist of blackish very

fine clays rich in organic matter. Arsenic and other trace

metal contents are strongly correlated with Fe2O3,

suggesting that Fe oxides play a role in controlling

abundances.

This study demonstrates that the shallow aquifers of the

Holocene sediments in the Old Brahmaputra plain are

affected by As-enrichment, whereas the aquifers in the

Pleistocene sediments are low in As. High-As and low-As

groundwaters are produced from the Holocene (anoxic) and

1314 Environ Earth Sci (2010) 60:1303–1316

123

Page 13: Geochemical study of arsenic and other trace elements in groundwater and sediments of the Old Brahmaputra River Plain, Bangladesh

Pleistocene (oxic) aquifers, respectively. Since groundwa-

ter is used as a source of drinking water in Bangladesh,

routine monitoring of As levels in the waters from wells

must be performed to avoid high-As concentrations in

water.

Acknowledgments The authors thank Professor Yoshikazu Sampei

and Professor Yoshihiro Sawada of Shimane University for access to

the AAS and XRF facilities, and Professor M. A. Sattar of Bangladesh

Agricultural University, Mymensingh, Bangladesh, for his coopera-

tion in sampling. We thank the Mymensingh City Authority for their

valuable support during sample collection. The radiocarbon dating

was carried out at Nagoya University, Japan.

References

Acharyya SK, Chakraborty P, Lahiri S, Raymahashay BC, Guha S,

Bhowmik A (1999) Arsenic poisoning in the Ganges delta.

Nature 401:545. doi:10.1038/44052

Acharyya SK, Lahiri S, Raymahashay BC, Bhowmik A (2000)

Arsenic toxicity of groundwater in parts of the Bengal basin in

India and Bangladesh: the role of Quaternary stratigraphy and

Holocene sea-level fluctuation. Environ Geol 39:1127–1137.

doi:10.1007/s002540000107

Ahmed KM, Bhattacharya P, Hasan MA, Akhter SH, Alam SMM,

Bhuyian MAH, Imam MB, Khan AA, Sracek O (2004) Arsenic

enrichment in groundwater of the alluvial aquifers in Bangla-

desh: an overview. Appl Geochem 19:181–200. doi:10.1016/

j.apgeochem.2003.09.006

Akai J, Izumi K, Fukuhara H (2003) Bacterial activities as a key role

in arsenic contaminations of groundwater suggested by geo-

microbial experiments using lake sediments. In: International

symposium on bio-mineralization, Japan, pp 12–17

Akai J, Izumi K, Fukuhara H, Masuda H, Nakano S, Yoshimura T,

Ohfuji H, Anawar HM, Akai K (2004) Mineralogical and

geomicrobiological investigations on groundwater arsenic

enrichment in Bangladesh. Appl Geochem 19:215–230. doi:

10.1016/j.apgeochem.2003.09.008

Anawar HM, Komaki K, Akai J, Takada J, Ishizuka T, Takahashi T,

Yoshioka T, Kato K (2002) Diagenetic control on arsenic

partitioning in sediments of the Meghna River delta, Bangladesh.

Environ Geol 41:816–825. doi:10.1007/s00254-001-0459-x

Anawar HM, Akai J, Komaki K, Terao H, Yoshioka T, Shizuka T,

Safiullah S, Kato K (2003) Geochemical occurrence of arsenic in

groundwater of Bangladesh: sources and mobilization processes.

J Geochem Explor 77:109–131. doi:10.1016/S0375-6742(02)

00273-X

Ayyamperumal T, Jonathan MP, Srinivasalu S, Armstrong-Altrin JS,

Ram-Mohan V (2006) Assessment of acid leachable trace metals

in sediment cores from River Uppanar, Cuddalore, Southeast

coast of India. Environ Pollut 143:34–45. doi:10.1016/

j.envpol.2005.11.019

BGS and DPHE (2001) Arsenic contamination of groundwater in

Bangladesh, vol 2. Final report, BGS Technical Report WC/00/

19, 267 pp

Bhattacharyya R, Jana J, Nath B, Sahu S, Chatterjee D, Jacks G

(2003) Groundwater arsenic mobilization in the Bengal Delta

Plain, the use of ferralite as a possible remedial measure—a case

study. Appl Geochem 18:1435–1451

Brammer H (1996) The geography of the soils of Bangladesh.

University Press Ltd, Dhaka, p 287

Chowdhury TR, Basu GK, Mandal BK, Biswas BK, Samanta G,

Chowdhury UK, Chanda CR, Lodh D, Roy SL, Saha KC, Roy S,

Kabir S, Quamruzzaman Q, Chakraborti D (1999) Arsenic

poisoning in the Ganges delta. Nature 401:545–546

Dhar RK, Biswas BK, Samanta G, Mandal BK, Chakraborty D, Roy

S, Jafar A, Islam A, Ara G, Kabir S, Khan AW, Ahmed SA, Hadi

SA (1997) Groundwater arsenic calamity in Bangladesh. Curr

Sci 73(1):48–59

Fabian D, Zhou Z, Wehrli B, Friedl G (2003) Diagenetic cycling of

arsenic in the sediments of eutrophic Baldeggersee, Switzerland.

Appl Geochem 18:1497–1506

Gaillard J-F (1994) Early diagenetic modeling: a critical need for

process studies, kinetic rates, and numerical methods. Trends

Chem Geol 1:239–252

Garcia-Sanchez A, Alvarez-Ayuso E (2003) Arsenic in soils and

waters and its relation to geology and mining activities

(Salamanca Province, Spain). J Geochem Explor 80:69–79

Guern CL, Baranger P, Crouzet C, Bodenan F, Conil P (2003) Arsenic

trapping by iron oxyhydroxides and carbonates at hydrothermal

spring outlets. Appl Geochem 18:1313–1323

Huerta-Dıaz MA, Tessier A, Carignant R (1998) Geochemistry of

trace metals associated with reduced sulfur in freshwater

sediments. Appl Geochem 13:213–233

Ishiga H, Dozen K, Yamazaki C, Ahmed F, Islam MB, Rahman MH,

Sattar MA, Yamamoto H, Itoh K (2000) Geological constraints

on arsenic contamination of groundwater in Bangladesh. In:

Proceedings of the 5th international forum on arsenic contam-

ination of groundwater in Asia, Asia Arsenic Network, Yoko-

hama, Japan, pp 53–62

Manning BA, Goldberg S (1997) Adsorption and stability of As (III)

at the clay mineral-water interface. Environ Sci Technol

31:2005–2011

Markgraf V (1989) Palaeoclimates in central and South America

since 18,000 BP based on pollen and lake level-records. Quatern

Sci Rev 8(1):1–24

McArthur JM, Ravenscroft P, Safiullah S, Thirlwall MF (2001)

Arsenic in groundwater: testing pollution mechanisms for

sedimentary aquifers in Bangladesh. Water Resour Res

37:109–117

McArthur JM, Banerjee DM, Hudson-Edwards KA, Mishra R,

Purohit R, Ravenscroft P, Cronin A, Howarth RJ, Chatterjee

A, Talukder T, Lowry D, Houghton S, Chadha DK (2004)

Natural organic matter in sedimentary basins and its relation to

arsenic in anoxic groundwater: the example of West Bengal and

its worldwide implications. Appl Geochem 19:1255–1293

Millward GE, Moore RM (1982) The adsorption of Cu, Mn and Zn by

iron oxyhydroxides in model estuarine solutions. Water Res

16:981–985

Monsur MH (1990) Stratigraphical and palaeomagnetical studies of some

quaternary deposits of the Bengal Basin, Bangladesh. Doctoral

dissertation, Free University of Brussels, Belgium, 241 pp

Mukherjee AB, Bhattacharya P (2001) Arsenic in groundwater in the

Bengal Delta Plain: slow poisoning in Bangladesh. Environ Rev

9:189–220

Nickson RT, McArthur JM, Burgess WG, Ahmed KM, Ravenscroft P,

Rahman M (1998) Arsenic poisoning of Bangladesh groundwa-

ter. Nature 395:338

Nickson RT, McArthur JM, Ravenscroft P, Burgess WG, Ahmed KM

(2000) Mechanism of arsenic release to groundwater, Bangla-

desh and West Bengal. Appl Geochem 15:403–413

Nickson RT, McArthur JM, Shrestha B, Kyaw-Myint TO, Lowry D

(2005) Arsenic and other drinking water quality issues, Muzaf-

fargarh district, Pakistan. Appl Geochem 20:55–68

Ogasawara M (1987) Trace element analysis of rock samples by

X-ray fluorescence spectrometry, using Rh anode tube. Bull Geol

Surv Jpn 38(2):57–68

Padmalal D, Maya K, Seralathan P (1997) Geochemistry of Cu, Co,

Ni, Zn, Cd and Cr in the surficial sediments of a tropical estuary,

Environ Earth Sci (2010) 60:1303–1316 1315

123

Page 14: Geochemical study of arsenic and other trace elements in groundwater and sediments of the Old Brahmaputra River Plain, Bangladesh

southwest coast of India: a granulometric approach. Environ

Geol 31(1–2):85–93

Peltier EF, Webb SM, Gaillard JF (2003) Zinc and lead sequestration

in an impacted wetland system. Adv Environ Res 8:103–112

Potts PJ, Tindle AG, Webb PC (1992) Geochemical reference

material compositions. Whittles Publishing, Caithness, p 313

Preda M, Cox ME (2002) Trace metal occurrence and distribution in

sediments and mangroves, Pumicestone region, southeast

Queensland, Australia. Environ Int 28:433–449

Ravenscroft P, McArthur JM, Hoque BA (2001) Geochemical and

palaeohydrological controls on pollution of groundwater by arsenic.

In: Chappell WR, Abernathy CO, Calderon R (eds) Arsenic

exposure and health effects IV. Elsevier, Oxford, pp 53–77

Rudnick RL (2005) The crust. In: Holland HD, Turekian KK (eds)

Treatise on geochemistry, vol 3. Elsevier, Oxford, p 537

Singh M, Sharma M, Tobschall HJ (2005) Weathering of the Ganga

alluvial plain, northern India: implications from fluvial geo-

chemistry of the Gomati River. Appl Geochem 20:1–21

Smedley PL, Kinniburgh DG (2002) A review of the source, behavior

and distribution of arsenic in natural waters. Appl Geochem

17:517–568

Sullivan KA, Aller RC (1996) Diagenetic cycling of arsenic in

Amazon shelf sediments. Geochim Cosmochim Acta 60(9):

1465–1477

Tareq SM, Safiullah S, Anawar HM, Rahman MM, Ishizuka T (2003)

Arsenic pollution in groundwater: a self-organizing complex

geochemical process in the deltaic sedimentary environment,

Bangladesh. Sci Total Environ 313:213–226

Taylor SR, McLennan SM (1985) The continental crust: its compo-

sition and evolution. Blackwell, Oxford, p 312

Tessier A, Carignan R, Belzile N (1994) Processes occurring at the

sediment-water interface: emphasis on trace elements. In: Buffle

J, DeVitre RR (eds) Chemical and biological regulation of

aquatic system. Lewis Publishers, Boca Raton, pp 137–173

Tribovillard NP, Desprairies A, Verges EL, Bertrand P, Moureau N,

Ramdani A, Ramanampisoa L (1994) Geochemical study of

organic-matter rich cycles from the Kimmeridge Clay Formation

of Yorkshire (UK): productivity versus anoxia. Palaeogeogr

Palaeoclimatol Palaeoecol 108:165–181

Umitsu M (1985) Natural levees and landform evolutions in the

Bengal Lowland. Geogr Rev Jpn Ser B 58:149–164

Umitsu M (1993) Late quaternary sedimentary environments and

landforms in the Ganges delta. Sed Geol 83:177–186

WHO (World Health Organization) (2004) Guidelines for drinking-

water quality recommendations, vol 1. WHO, Geneva, p 515

Yamazaki C, Ishiga H, Dozen K, Higashi N, Ahmed F, Sampei Y,

Rahman MH, Islam MB (2000) Geochemical composition of

sediments of the Ganges delta of Bangladesh-arsenic release

from the peat? Earth Sci (Chikyu Kagaku) 54:81–93

Yamazaki C, Ishiga H, Ahmed F, Itoh K, Suyama K, Yamamoto H

(2003) Vertical distribution of arsenic in Ganges delta sediments

in Deuli village, Bangladesh. Soil Sci Plant Nutr 49(4):567–574

Yan X-P, Kerrich R, Hendry MJ (2000) Distribution of arsenic (III),

arsenic (V) and total inorganic arsenic in porewaters from a thick

till and clay-rich aquitard sequence, Saskatchewan, Canada.

Geochim Cosmochim Acta 62(15):2637–2648

Zheng Y, Stute M, van Geen A, Gavrieli I, Dhar R, Simpson HJ,

Schlosser P, Ahmed KM (2004) Redox control of arsenic mobili-

zation in Bangladesh groundwater. Appl Geochem 19:201–214

1316 Environ Earth Sci (2010) 60:1303–1316

123