NATURAL ARSENIC IN GROUNDWATER: OCCURRENCE, REMEDIATION AND MANAGEMENT - CHAPTER 20 pptx

The impact of low dissolved oxygen in recharge water on arsenic
pollution in groundwater of Bangladesh
Md. Nazrul Islam
Department of Civil Engineering, University of Toronto, Toronto, Canada
R.D. Von Bernuth
Department of Biosystems Engineering, Michigan State University,
East Lansing, Michigan,
USA
ABSTRACT: This study is based on the concept that lack of dissolved oxygen (DO) at or below
the water table and water extraction (Q) through shallow irrigation wells at a rate greater than the
aquifer recharge rate are the main causes of arsenic release in the groundwater of Bangladesh. This
study identified the hydrogeochemical processes related to shortage of DO that eventually pro-
duce high arsenic concentrations and their migration into the groundwater systems. The existing
theories of arsenic release by oxidation and reduction in the context of dissolved oxygen shortage
in recharging groundwater were studied. Both numerical and thermodynamic analyses were used
to demonstrate how oxidation theory of arsenic release is inadequate to explain the release of
arsenic into the groundwater of Bangladesh. This study quantified the amount of dissolved oxy-
gen level in deeper layers of the aquifer and their relation to the variations in redox potential
values and arsenic release processes. It also analyzed groundwater velocity and flow patterns to
establish a link between dissolved oxygen shortage and arsenic release into the groundwater. On
the basis of the findings, it was concluded that shortage of dissolved oxygen in recharging water
is the most likely the root cause of arsenic occurrence in Bangladesh groundwater.
1 INTRODUCTION
The arsenic contamination problem in Bangladesh groundwater is most likely to be associated
with shortage of dissolved oxygen in recharge water at or below the water table. The presence or
absence of dissolved oxygen (DO) in natural surface or groundwater systems has been known as
an index of Oxidation Reduction Potential (ORP) or redox potential. The ORP is a measure of ten-
dency for donating or accepting electrons during chemical reactions. As long as any measurable
amount of dissolved oxygen is present in the groundwater systems, the redox potential (p
e
) level

is controlled by the dissolved oxygen concentration and the system will remain in oxic conditions.
Under higher oxic conditions, ion activities and electron donating tendencies are less (Stumm &
Morgan 1996, Khan et al. 2000). On the contrary, under anoxic conditions, the tendency for donat-
ing electrons is high. Arsenic concentrations in the groundwater in Bangladesh are a by-product of
redox reactions where microbial derived oxidation of organic carbon plays an important role in
donating electron, and the terminal electrons are accepted by the hydrated ferric oxides and/or
hydroxides present in the groundwater systems. Upon electron acceptance, ferric oxides are
reduced from solid particulate form to dissolved ferrous ions and the associated arsenic is released
into the groundwater systems (Nickson et al. 2000). This is the most widely accepted reduction
theory of arsenic release into the groundwater of Bangladesh.
The origin of arsenic in the aquifer sediment is from natural geological processes. There are two
theories of arsenic release. One is that over time the microbial degradation of organic carbon has
caused arsenic to be released into water by reductive dissolution of iron oxides under anoxic or
173
Natural Arsenic in Groundwater: Occurrence, Remediation and Management –
Bundschuh, Bhattacharya and Chandrasekharam (eds)
© 2005, Taylor & Francis Group, London, ISBN 04 1536 700 X
Copyright © 2005 Taylor & Francis Group plc, London, UK
mildly reduced conditions. This is known as reduction theory of arsenic release. However, some
experts believe that exposure of arsenic rich pyrites to the atmospheric oxygen due to over extrac-
tion of groundwater is the main cause of arsenic release into the groundwater of Bangladesh
(Bridge & Hussain 1999). This idea is known as the oxidation theory of arsenic release. Neither of
these two theories is unequivocally accepted by all and a strong debate has been continuing on
these two current theories of arsenic release. It is not the purpose of this research to engage in this
debate. In both cases dissolved oxygen level plays an important role in arsenic contamination
problem, and that is the focus of this study.
The ORP or redox potential of the sediment water systems mainly controls the arsenic specia-
tion in groundwater systems (Bhattacharya et al. 2002a b, Ahmed et al. 2004). The Speciation or
different oxidation states of arsenic species in water is highly attributed to the physicochemical
(electrostatic force, ionic strength, pH) and molecular interactions (thermodynamic stability,

water solubility, hydrogen bonding ability etc.) (Gazsó 2001). The physicochemical and molecu-
lar interactions between arsenic species and aquifer sediments are largely influenced by the asso-
ciated biogeochemistry of the aquifer systems. The amount of dissolved oxygen present in the
systems and the rate of consumption of dissolved oxygen generally play an important role in con-
trolling the redox potential.
As long as the water table remains close to the ground surface, the dissolved oxygen content in
groundwater remains in equilibrium with atmospheric oxygen. Consumption of oxygen by micro-
organisms can shift the equilibrium from oxic to suboxic states. The amount of atmospheric oxygen
diffusion into the groundwater through the unsaturated porous medium basically depends on the
depth to the water table and the oxygen diffusion rate (ODR) (Mukhtar et al. 1996). The ODR into
174
Figure 1. Probability of exceedence of arsenic above threshold level (50 ppb) is higher in the SE (southeast)
and SW (southwest) zones of Bangladesh.
Copyright © 2005 Taylor & Francis Group plc, London, UK
the groundwater varies inversely with the depth to the water surface of the aquifer. If water table
moves downward, supply of atmospheric oxygen will be less in the groundwater.
The dissolved oxygen level in aquifer recharge water is attributed to the surface runoff and
seepage from river bed. Most recently, upstream diversion or withdrawal of river water from the
major river systems (Ganges and Jamuna rivers, Fig. 1) and large scale installation of shallow
wells in Bangladesh is believed to be responsible for rapid lowering water table. Rapid consump-
tion of dissolved oxygen in recharging groundwater due to the lowering of water table and micro-
bial use of organic carbon as their energy-supplying-electron donor might have triggered the
arsenic contamination problem in the groundwater of Bangladesh. This is the principal working
hypothesis of this study. This hypothesis is in complete contrast with the existing oxidation theory
of arsenic release but partially supports theory of arsenic release by reduction.
The main purpose of this investigation was to examine how the shortage of dissolved oxygen in
the recharging water contributed to groundwater arsenic pollution problem in Bangladesh. The
second purpose is to understand whether the concept of dissolved oxygen shortage in recharging
groundwater does or does not contradict the current theories of arsenic release by oxidation and or
reduction. The third purpose of this study is to understand how dissolved oxygen shortage might

influence the microbial activities that eventually influence the arsenic speciation processes.
Addressing these questions may help in finding a suitable bioremediation solution to groundwater
arsenic problem in Bangladesh.
2 METHODOLOGY OF THE STUDY
To accomplish the purposes of this study, five analytical approaches were adapted; (1) statistical
analyses of spatial distribution patterns of arsenic concentrations, (2) a multiple layer oxygen dif-
fusion model analysis to estimate the amount of dissolved oxygen concentration in deeper layers,
(3) analyses of hydrological factors such as hydraulic gradient, groundwater flow patterns and
their relation to arsenic contamination problem and dissolved oxygen shortage, (4) computation of
redox potential values at different depths of the aquifer and their correlation with vertical arsenic
concentration distribution, and (5) thermodynamic analyses and explanations of existing theories
of arsenic release and the validity of the hypothesis presented in this study. Thermodynamic analy-
ses were done to establish a link between dissolved oxygen shortage, lowering of water table and
their impact on thermodynamic stability of arsenic species and redox potentials values.
2.1 Statistical analyses of arsenic concentrations distribution patterns
The analyses of spatial arsenic distribution patterns were done to understand the impact of dis-
solved oxygen shortage on the arsenic release mechanism as a function of the associated hydro-
geological conditions. A question is, “Which hydrologic zone in Bangladesh did correlate best to
the arsenic concentration and how the dissolved oxygen shortage in that process does increase
arsenic concentration?”
The arsenic concentration records of about 3500 well water samples were analyzed by the
BGS/DPHE. The water quality data were analyzed at different universities in the USA and abroad.
The authors computed the probability of arsenic concentrations exceeding the threshold level
(50 ppb) by using Gumbel exponential distribution method. Based on the volume of available
water resources (rainfall, recharge, and surface water), Bangladesh was divided into six hydrolog-
ical planning zones (MPO 1987) such as, northeast (NE), north-center (NC), north-west (NW),
southeast (SE), south-center (SC) and southwest (SW) zones (Fig. 1). The coordinates (latitude
and longitude) of the arsenic contaminated well records were inserted into the map of Bangladesh.
The distribution patterns of arsenic concentration were statistically analyzed within the boundary
of each zone and results are presented in Table 1. It was found from the analyses that the prob-

ability of arsenic exceeding the threshold level (50 ppb) in the south-east (SE) zone was computed
as 71.4% by using Gumbel equation. The probability of exceeding arsenic below threshold levels
175
Copyright © 2005 Taylor & Francis Group plc, London, UK
(50 ppb) in the northwest (NW), north center (NC) zones are shown in Figure 1 and the computed
values are tabulated in Table 1.
2.2 Multi layer oxygen diffusion model to estimate the dissolved oxygen level
In order to determine the impact of shortage of dissolved oxygen in the recharge water at deeper
layers of aquifers as a function of lowering the water table, a numerical oxygen diffusion model
was built using the Finite Element Analyses technique. The Finite Element technique was used
because of its higher accuracy and adaptabilities to numerical solutions for physical processes like
convection diffusion, pollution distribution and contaminant transportation.
2.2.1 Conceptual oxygen diffusion model
For the sake of simplicity, this oxygen diffusion model considers a constant oxygen consumption
rate (␣ϭ0.0021 cm
3
/cm
3
/hr) by organic carbon in both saturated and unsaturated zone of the
aquifer. This model was built to validate the theory of arsenic release by oxidation where the
arsenic rich aquifer layers are exposed to the atmospheric oxygen. It was hypothetically assumed
that the arsenopyrite-rich sediments layers L-6 an L-10 are located respectively at 6 and 10 meter
below the ground surface (Fig. 2). A layered 10 m thick aquifer was modeled, representing the
layer numbers by L-1 to L-10 (Fig. 2) with each layer being 100 cm thick. WT-1 and WT-2 show
176
Table 1. Probability of exceedence of arsenic concentrations in different hydrologic planning zones of
Bangladesh.
Hydrologic No. of Average As
zone wells conc. (␮g/L) St. Dev Probability of exceedence (%)
10 ppb 25 ppb 50 ppb 100ppb 250 ppb

Northeastern (NE) 1039 34.0 (0.5–572) 68 58.7 48.7 34.0 14.9 Ͻ1
North center (NC) 192 28.6 (0.5–284.0) 51.4 53.9 48.0 38.9 24.4 4.9
Northwest (NW) 1072 12.3 (0.5–708) 47.1 44.9 32.7 18.1 5 0.3
Southeastern (SE) 295 174.1 (0.5–1090) 199. 80.1 76.2 71.4 59.55 29.1
Southwest (SW) 474 84.8 (0.5–1660) 145 66.2 61.4 53.3 38.74 12
South center (SC) 295 38.6 (0.5–862) 113.18 53.9 48.05 38.95 24.42 4.98
Figure 2. Conceptual oxygen diffusion model to estimate oxygen concentrations at deeper layers before and
after lowering of water table.
Copyright © 2005 Taylor & Francis Group plc, London, UK
the water table elevation before and after large-scale well installation and Case-1 and Case-2
reflect hydrologic conditions respectively before and after large-scale well installation.
This model estimated the change in amount of dissolved oxygen at the exposed deeper layers
(L-6 and L-10, Fig. 2) after lowering water table from WT-1 to WT-2. The oxygen diffusion model
was simulated up to 350 hours from beginning with time step ⌬t is equal to 1 and 4 hours.
However, results of 1-hour time step are printed every 10 hours (Fig. 3). The change in oxygen
exposure took place at layer L-6 and L-10 before and after introduction of large-scale shallow irri-
gation wells in Bangladesh was investigated. In the diffusion model the following values were
addumed: the diffusion coefficient Dx was 259.2cm
2
/hr (4.166 * 10
Ϫ4
m
2
/s) and the oxygen diffu-
sion rate was Ϫ 0.002125cm
3
/cm
3
/h.
The results of the transient oxygen diffusion model were analyzed and plotted (Fig. 3). Figure 3

shows that a 4 m lowering of the water table from layer L-2 to L-6 did not result in an increase of the
oxygen concentration at L-10 and L-6. The oxygen concentration in Case-1 at layer L-6 was predicted
0.09 atm (3.64 mg/L) after 150 hours. After lowering of the water table, using the same time interval
and the same aquifer properties, the oxygen concentration was found to be 0.06atm (2.39mg/L). This
result implies that lowering of the water table cannot increase the oxygen supply to the deeper layers
of the aquifer. These results are help to establish a link between current theory of arsenic release and
role of oxygen shortage and are further addressed in the results and discussion section.
2.3 Hydrological factors and their relation to arsenic contamination problem
The relations among arsenic distribution patterns and groundwater velocity and flow directions were
analyzed by dividing the whole country into a number of square grids where arsenic concentration
records and groundwater flow direction maps were available. The whole country was divided into 16
177
Figure 3. Change in dissolved oxygen concentration in layer L-6 and L-10 after lowering of water table 5m
from the position of WT-1 to WT-2.
Copyright © 2005 Taylor & Francis Group plc, London, UK
178
Figure 4. The schematic groundwater flow direction is shown towards the main river system and the Bay of
Bengal where the difference between the mean sea level (MSL) is about 95 m (from north to south).
Arsenic concentration and groundwater flow direction
0
6789 10
200
400
600
800
1000
1200
Location of grid numbers along groundwater flow direction (North to South)
Arsenic concentration(ppb)
Increasing trend of

arsenic contamination
Figure 5. Increased trend of arsenic concentration from north to south along the groundwater flow direction.
Copyright © 2005 Taylor & Francis Group plc, London, UK
grids where arsenic contaminated wells were available. Grids 6, 7, 8, 9 and 10 (Fig. 4) have the same
general hydraulic gradient and direction of flow. The arsenic concentration records in those five grids
showed a strong correlation between arsenic concentration and groundwater flow direction (Fig. 5).
2.4 Computation of redox potential values and their relation with
arsenic concentration distribution
The redox potential values at different layers of the aquifer system were computed by measuring the
activity of electrons in a solution expressed in units of volts (E
h
) or in units of electron activity
(p
e
). The aquifer sediment contains solid amorphous Fe(OH)
3
and it was assumed that the poten-
tial corresponds to an oxidation-reduction potential of the aquatic environment. By computing
equilibrium constant value of log K, and the standard state free energy value ⌬G
0
, the equation
that expressed the equilibrium of iron oxides with dissolved Fe
2ϩ
gave the value of the redox
potential in the groundwater systems (Figs. 7 and 8).
3 RESULT AND DISCUSSIONS
The results of the analyses are discussed in the light of the hypothesis of this study. The working
hypothesis was that layers at or below the water table receive less oxygenated water because of low-
ering water table which is due to large scale installation of irrigation wells and upstream withdrawal
of river water in Bangladesh. The shortage of dissolved oxygen in the recharge water may also be

due to microbial activity. The immediate question one would ask is how a shortage of dissolved
oxygen occurred in the recharging groundwater and how it contributes to arsenic mobilization.
It should be kept in mind that the cause of arsenic contamination in the groundwater of
Bangladesh is still poorly understood and a strongly debatable issue. However, the above questions
could be answered in the context of hydrogeochemical issues. It was apparent from Table 1 that the
probability of arsenic exceeding threshold level (50 ppb) was highest in the southeast (SE) and
179
Figure 6. Arsenic concentration distribution over aquifer depth.
Copyright © 2005 Taylor & Francis Group plc, London, UK
southwest (SW) zones of Bangladesh. These two regions were built by the sediment of the Meghna
and Ganges River Flood Plain. This delta usually experiences a high rate of sediment flow of about
479 million tons per annum (BWDB 1993). The least arsenic contaminated regions are located in
the northwest (NW) and north center (NC) regions where the surface geology is built up by the
Old Himalayan Fan, Tista-Jamuna Flood Plains, and Madhupur Tract. The top of the main aquifer
systems in the NW zone is closer to the surface than that of the SE and SW regions. The regional
groundwater gradient in the NW zone is 2 m/km but the gradient in the SE and zone is 0.1 m/km
(MPO, 1987). It is apparent that the groundwater gradient in NW zone is 20 times higher than
southern zones(SE & SW) where the arsenic concentration is maximum (Fig. 1).
3.1 Arsenic concentration is greater in the southeast (SE ) and southwest (SW) zones
The sedimentology of northwest (NW) zone is different from SE and SW zone. Moreover, the
hydraulic gradient in NW zone is twenty times higher than SW zone; therefore, the residence time
of groundwater in NW zone would be less than SW and SE zones. On the contrary, the SW and SE
zones were mostly built up through the sedimentary deposition, and the hydraulic gradient in the
SW and SE zones is less; therefore, the groundwater residence time is expected to be higher than
180
Redox potential, p
e
(V)
Aquifer depth (m)
Figure 7. Redox potential values (p

e
) decrease with aquifer depth from 10 to 30 m below the surface.
Redox potential, p
e
(V)
Aquifer depth (m)
Figure 8. The redox potential values (p
e
) start increasing with aquifer depth from 150 to 350 m below the
surface indicating that the recharge water is rich in dissolved oxygen.
Copyright © 2005 Taylor & Francis Group plc, London, UK
in the NW zone. Potential recharge water in the shallow aquifer is usually rich in dissolved oxy-
gen. Under the influence of the high hydraulic gradient in NW zone, the oxygenated groundwater
recharge can quickly and easily replace the old groundwater. After rejuvenating the aquifers by
recharge water, the groundwater in NW zone is expected to remain oxic if there is not too much
organic carbon leaching into the aquifer from the top of the ground surface.
Continuous percolation of organic carbon from the increased agricultural activities can also
result in rapid consumption of dissolved oxygen in the groundwater. The groundwater in max-
imum arsenic contaminated area (SW and SE zones, Fig. 1) might be continuously lacking in dis-
solved oxygen because it can not be easily replaced by the oxygenated recharge water because of
low gradient. As a result, the redox potential value always remains low in SW and SE region. Only
mixing with oxygenated water or lack of reductant minerals can maintain the desired oxic condi-
tion in SW and SE regions, but achieving oxic conditions is difficult in these zones due to rapid
consumption of dissolved oxygen. Therefore, arsenic mobilization is the higher in SW and SE
region than NW, NE and north center (NC) zone in Bangladesh.
3.1.1 Lack of correlation between arsenic concentrations and depth of water table
The investigation report (BGS, 2001) showed that arsenic contamination does not have any rela-
tionship with depth to the water table or amount of irrigation withdrawal rate in Bangladesh.
Moreover, in the most contaminated zone (SE and SW zone) the water table does not match with
the most intensive extraction rate. Because both arsenic release mechanisms (oxidation or reduc-

tion) are time dependent processes but concentrations spatially differ, it is inappropriate to explain
differences using a space–dependent relationship.
Therefore, it might be misleading to establish a direct correlation between arsenic concentra-
tions and depth of water table or extraction rate (Q). For example, if the volume of irrigation with-
drawal is considered as a function of water table depth, a consistent relationship can be obtained
only when the specific yields and effective porosity are the same for every aquifer. Otherwise, the
correlation will be influenced by those parameters and may destroy the relationship. The lack of
correlation between arsenic concentrations and lowering of water table or abstraction rates (Q)
does not necessarily lead to rejecting the idea that over-pumping of irrigation wells may cause
arsenic mobilization in the groundwater of Bangladesh. Recently, a group of researchers based in
Manchester University found evidence that influxes of organic carbon in groundwater are known
to occur when irrigation wells are drawn down. This group also found that introduction of organic
carbon by over-pumping of irrigation wells can be a factor in increasing arsenic mobility in shal-
low groundwater in Bangladesh (Roach 2004). Organic carbon in the sediment acts as a food
source of bacteria. Bacteria would consume dissolved oxygen and eventually could lead to change
the biogeochemistry of the arsenic contaminated groundwater.
3.2 Maximum arsenic concentration at the depth 30 to 50 meter below the ground surface
The analysis of vertical distribution pattern of arsenic concentrations is shown in Figure 6. That
shows that arsenic concentrations above threshold level (Ͼ50 to 1600 ppb) are mostly confined
within between 30 and 50 m below the ground level. But, arsenic concentrations were less than
50 ppb between 150 and 350m. The sequence of vertical arsenic distribution can be interpreted in
terms of redox potential values as shown in Figures 7 and 8. The increasing trend of arsenic at the
depth ranging from 9 to 50 m below the ground (Fig. 6) was found to be associated with decreas-
ing trend of redox potential as shown in Figure 7. The decreasing trend of arsenic from 150 to
350 m in Figure 6 is related to the increasing trend in redox values as shown in Figure 8. Although,
the two aquifer systems have many differences, the sequence of decreasing and increasing trend
of the redox potential values (Figs. 7 and 8) could be associated with the dissolved oxygen content in
the groundwater. The recharging groundwater at the depth of 150 to 350m is mostly attributed
to the regional thorough-flow which is rich in dissolved oxygen. The presence of electron donors at
the deeper aquifer layers is also less. On the other hand, the recharging groundwater at the upper

part of the aquifer is also rich in dissolved oxygen since it comes from the surface runoff and river
181
Copyright © 2005 Taylor & Francis Group plc, London, UK
bed seepage, but this dissolved oxygen content can be quickly consumed by the electron donors.
Therefore, the shortage of dissolved oxygen might play an important role in mobilizing arsenic in
the groundwater system because dissolution of iron oxides occurs under anoxic conditions (Lee
et al. 2003).
3.2.1 How does the dissolved oxygen level in recharging water decrease over aquifer depth?
It is a common misconception that lowering of water table would increase the unsaturated vadose
zone and eventually the oxygen concentration level in deeper layers of the aquifer will be
increased. The oxygen diffusion model analyses as shown in Figure 2 demonstrated that lowering
of water table does not increase the oxygen supply rate into the deeper layers of the aquifer. To
prove this, a one dimensional oxygen diffusion model was built, where the water table elevations
at WT-1 and WT-2 (Fig. 2) represent the levels before and after installation of irrigation wells
respectively. The oxygen diffusion model suggests that lowering the water table about four meters
(from WT-1 to WT-2) did not increase the dissolved oxygen concentration at the hypothetical
arsenic contaminated layers at L-6 and L-10 (Fig. 2). The oxygen concentration in Case-1 at layer
L-6 was estimated as 0.09atm (3.64mg/L) after 150 hours. But in Case-2, after lowering of the
water table, using the same interval of time and same aquifer properties, the oxygen concentration
at L-6 was found as 0.06atm (2.39mg/L). Therefore, lowering water table decreased the dissolved
oxygen concentrations in the same aquifer. Consumption or reduction of dissolved oxygen supply
helps develop reducing conditions.
3.2.2 Shortage of dissolved oxygen decreases the redox potential values
The vertical arsenic distribution patterns as shown in Figure 6 indicate that the wells contaminated
with arsenic in concentrations greater than 50 ppb are mostly located at the depth ranging from 10
to 35 m below the ground. The associated redox potential values (p
e
) were computed using the dis-
solved iron concentration records and are depicted in the Figures 7 and 8. It is apparent in Figure
7 that the redox potential values start decreasing (0 to Ϫ0.5) from the surface of the water table

and continue up to 35 m below the ground level where the p
e
values fall below zero (Ϫ1.5) (Fig. 7).
Certainly, the main reason for the decreasing redox value is the consumption of dissolved oxygen
by electron donors. But the redox potential values at the top most surface of the water table are
higher than the deeper part of the aquifer because of its close proximity to the ground surface. At
the top layer of the aquifer, the dissolved oxygen level is detectable and much higher than deeper
layers. The redox potential values again start increasing from 150 to 350 m below the ground. This
is also associated with the dissolved oxygen concentration in the recharging groundwater. This can
be reasoned that recharging groundwater at the depth below 150 m is known as the regional through-
flow. The regional through-flow is usually rich in dissolved oxygen level, and in the deeper layers
the oxygen consumption rate is negligible since organic carbon can not reach that point.
3.3 Validity of the existing theories of arsenic release from the context of
dissolved oxygen shortage
There have been divergences of views among the researchers and considerable difficulties in
explaining the cause of arsenic mobilization in Bangladesh groundwater. While it was not the pur-
pose of this analysis to focus on this debate, it was necessary to evaluate both oxidation and reduc-
tion theory of arsenic release in the context of shortage of dissolved oxygen in order to support the
working hypothesis. Since, the working hypothesis of this study directly opposes the oxidation
theory of arsenic release; inadequacies in the oxidation theory will be discussed.
The proponents of the reduction theory have rejected the explanations of oxidation theory of
arsenic release. They argue that if pyrite oxidation were the real cause of arsenic release, arsenic
concentrations must have been positively correlated with the amount of sulfate in the arsenic con-
taminated groundwater. Arsenic contaminated groundwater in Bangladesh does not provide any
evidence of correlation between arsenic and sulfate in the field. However, Bridge & Hussain
(2000) mentioned that the lack of correlation between arsenic and sulfate does not indicate the
182
Copyright © 2005 Taylor & Francis Group plc, London, UK
rejection or acceptance of pyrite oxidation theory. They reasoned that sulfate concentration
in groundwater depends on many factors like pyrite grain size, pyrite abundance, reaction rate,

migration time, etc.
In order to examine the oxidation theory of arsenic release, the authors argue that there must be
arsenic concentration gradients during pumping from the contaminated shallow wells. If oxygen
exposure due to lowering of water table were the valid reason of arsenic release, then maximum
arsenic concentration should be located at the topmost layer of the aquifer (0 to 15 m) because dis-
solved oxygen is the highest in the topmost layer. Since it takes time for arsenic contaminated
water to reach the screen of the pumping wells a concentration gradient is expected to occur dur-
ing pumping. But arsenic contaminated wells do not show any concentration gradient during
pumping in Bangladesh. In addition, the oxidation theory of arsenic release can not explain why
arsenic concentration is maximum at 30 m below the ground level and why contamination is not
maximum just at the top of the water table. Therefore, oxidation theory is inadequate to explain the
arsenic mobilization in Bangladesh groundwater.
The iron oxides reduction model of arsenic release is now widely accepted as valid for the
Bengal Basin (Anwar et al. 2003, McArthur et al. 2001, Ravenscroft et al. 2001), however, their
suggestion that reduction is driven by buried peat has not been accepted by all. (Harvey et al.
2002) suggest that reduction of FeOOH is driven by surface organic matter from river beds, ponds,
and soils that is drawn into the aquifer by irrigation drawdown. Another unanswered question to
support the reduction theory is that if organic carbon is the major cause of arsenic release, why
was arsenic contamination not found in the earlier days, and why is there not any prior report of
arsenic toxicity in Bangladesh (Adel 2000, Bridge & Hussein 1999).
3.4 Role of dissolved oxygen to influence the microbial activities in
arsenic immobilization process
Immobilization of toxic metals and radio nuclides are usually brought about by precipitation,
biosorption and bioaccumulation processes. Immobilization by co-precipitation of arsenic with
ferric oxides totally depends on amount of dissolved oxygen present or oxic condition. Therefore,
dissolved oxygen plays an important role in immobilizing arsenic onto the surface of iron oxides.
Biosorption of toxic metals and radionuclide is based on non-enzymatic processes such as adsorp-
tion. Adsorption is due to the non-specific binding of ionic species to cell surface-associated or
extra cellular polysaccharides and proteins (Gazso 2001).Therefore, the dissolved oxygen content
in recharging groundwater may influence the biosorption process of arsenic removal.

3.4.1 Groundwater flow direction and its relation to arsenic mobilization
The groundwater flow direction maps (Figs 4 and 5) demonstrate a general trend that arsenic con-
centrations increased as the groundwater moved from the recharging (NW zone) to the discharg-
ing southeast (SE) zone. The groundwater residence time increases with decreasing hydraulic
gradient in the southeast zone of Bangladesh. Over time, the dissolved oxygen is consumed up by
the electron donors and helps develop a mild reducing condition. Usually, the tendency to donate
electrons is high in reducing conditions. Metal-reducing bacteria “breathe” by passing electrons
onto metals such as iron to get energy from their food. Iron reducing bacteria may use ferric oxides
as energy-supplying electron donors in the groundwater of Bangladesh (Roach 2004). Therefore,
arsenic from the surface of iron oxides is released as a result of terminal electron accepting process.
Also, the groundwater of SW and SE regions were found highly anoxic ( no trace amount of DO)
and offered a strong correlation with dissolved iron content (Nickson et al. 2000). Highly anoxic
groundwater in the SE zone of Bangladesh might be attributed to the lower hydraulic gradient and
higher residence time than the NW zone. This is why arsenic concentration in the SW and SE zone
is higher than the NW hydrologic zones in Bangladesh.
A groundwater flow model analysis by the authors (Islam & Bernuth 2003) demonstrated that in
general, pumping conditions provide more flow through the deeper layers than the natural ground-
water flow conditions. When the pumping rate is greater than the shallow aquifer recharge (SAR)
183
Copyright © 2005 Taylor & Francis Group plc, London, UK
rate, it increases the flow through the deeper layers when the aquifer is connected to the river sys-
tem. On the contrary, when SAR is greater than the pumping rate, it reduces the flow through the
deeper layers. The volume of SAR water is partially attributed to the river water and the river water
is usually rich in dissolved oxygen. Therefore upstream withdrawal of river water and the higher
extraction rate of wells would greatly reduce the SAR rate. Under that situation, groundwater flow
rate through the deeper layers would be increased (Islam & Bernuth 2003) and may contribute to
arsenic mobilization in Bangladesh groundwater.
4 CONCLUSIONS
Groundwater in the aquifers close to the surface is usually rich in dissolved oxygen (DO). The
presence of dissolved oxygen in the recharging groundwater is mainly attributed to the seepage

from river beds, close proximity of water table to the surface, surface runoff and lack of organic
carbon present in the system. Microbial growth influences the DO concentration in the aquifer. It
is apparent that the arsenic contamination in Bangladesh is not directly relational with depth to the
water table or extraction rate. Analyses have shown that lowering the water table and extraction
rate may influence the arsenic mobilization but that it is further influenced by the dissolved oxy-
gen content. Recent studies have demonstrated that lowering the water table and/or the extraction
rate influence the amount of organic carbon available in the deeper layers and that is believed to
be responsible for arsenic mobilization.
In addition, the pumping rate has a significant influence on the flow patterns through the deeper
layers of the aquifer system. The groundwater in the deeper layers is expected to be anoxic since
the recharging water may come from the storage if extraction rate is higher than aquifer recharge
(SAR) rate. Continuous pumping at a rate greater than SAR may lead to lowering of water table
which will eventually reduce the dissolved oxygen supply to the groundwater and would con-
tribute to the arsenic mobilization in Bangladesh groundwater.
Since arsenic mobilization is not solely due to chemical changes; but rather it is a resultant of
complex multidimensional biogeochemical and hydro-geological processes. Although it is true
that arsenic mobilization is poorly understood, but it is most likely that arsenic contamination
problem in Bangladesh is attributed to the shortage of dissolved oxygen level in the recharging
water. The exact contribution of river water diversion to the shortage of dissolved oxygen in
recharging groundwater is to be estimated by reliable and large scale hydrologic model analyses.
REFERENCES
Ahmed, K.M., Bhattacharya, P., Hasan, M.A., Akhter, S.H., Alam, S.M.M., Bhuyian, M.A., Imam, M.B.,
Khan, A.A., & Sracek, O. 2004. Arsenic enrichment in groundwater of the alluvial aquifers in Bangladesh:
An overview. In: Bhattacharya, P., Welch, A.H., Ahmed, K.M., Jacks, G. & Naidu, R. (eds) Arsenic in
Groundwater of Sedimentary Aquifers Appl. Geochem. 19(2): 181–200.
Adel, M.M. 2000. “Arsenification: Searching for alternative theory”. Daily Star of Bangladesh. 28 April, 2000.
Anawar, H.M., Akai, J., Komaki, K., Terao, H., Yoshioka, T., Ishizuka, T., Safiullah, S. & Kato, K. 2003.
Geochemical occurrence of arsenic in groundwater of Bangladesh: source and mobilization processes. J.
Geochem. Explor. 77: 109–131.
Bhattacharya, P., Frisbie, S.H., Smith, E., Naidu, R., Jacks, G. & Sarkar, B. 2002a. Arsenic in the

Environment: A Global Perspective. In: Sarkar B (ed) Handbook of Heavy Metals in the Environment
(Marcell Dekker Inc., New York): 145–215.
Bhattacharya, P., Jacks, G., Ahmed, K.M., Khan, A.A. & Routh, J. 2002b. Arsenic in groundwater of the
Bengal Delta Plain aquifers in Bangladesh. Bull. Env. Cont. Toxicology 69: 538–545.
BGS (British Geological Survey) & DPHE 2001. Arsenic contamination of groundwater in Bangladesh.
Technical report WC/00/19, Volume.2.
BWDB. 1993. Hydrological map of Bangladesh. Scale 1:1, 000000. Bangladesh Water Development Board,
Motejheel CA, Dhaka.
184
Copyright © 2005 Taylor & Francis Group plc, London, UK
Bridge, T.E. & Hussain, M.T. 1999. The increased draw down and recharge in groundwater aquifer and their
relationships to the arsenic problem in Bangladesh. URL: ./english/
arsenic/arsarticles.htm.
Gazsó, Lajos, G. 2001. The key microbial processes in the removal of toxic metals and radionuclides from the
environment. CEJOEM 7 (3–4): 178–185.
Harvey, C.F., Swartz, C.H., Badruzzaman, A.B.M., Keon-Blute, N., Yu, W., Ali, M.A., Jay, J., Beckie, R.,
Niedan, V., Brabander, D., Oates, P.M., Ashfaque, K.N., Islam, S., Hemond, H.F. & Ahmed, M.F. 2002.
Arsenic mobility and groundwater extraction in Bangladesh. Science 298: 1602–1606.
Islam, N. & Bernuth, R.D. 2003. Arsenic transportation in the groundwater of Bangladesh and the policy of
irrigation development. Journal de Physique IV France 107: 661–665.
Khan, A.H., Rasul, S.B., Munir, A.K.M., Habibuddowla, M., Alauddin, M. Newaz, S. S. & A. Hussam. 2000.
Appraisal of a simple arsenic removal method for groundwater of Bangladesh. J. Environ. Sci. Health
A35 (7): 1021–1041.
Lee,-Y., Um,-I-H & Yoon, J. 2003. Arsenic (III) oxidation by iron (VI) (ferrate) and subsequent removal of
arsenic (V) by iron (III) coagulation. Environ. Sci. Technol. 37(24): 5750–5756.
McArthur, J.M., Ravenscroft, P., Safiullah, S. & Thirlwall, M.F. 2001. Arsenic in groundwater: testing pollu-
tion mechanism for sedimentary aquifers in Bangladesh. Water Resour. Res. 37: 109–117.
Mukter, S., Baker, J.L. & Kanwar, R.S. 1996. Effect of south term flooding and drainage on soil oxygenation.
Transaction of the ASAE 39(3): 915–920.
MPO (Mastar Plan Organization). 1987. The groundwater resources and its availability for development. June

1987. Technical Report No. 5. Submitted to the Ministry of Irrigation, Water Development and Flood
Control, Government of the Peoples Republic of Bangladesh assisted by the United Nations Development
Program (UNDP) and the World Bank.
Megonigal, P., Emerson, D. & Weiss, J. 1997. Iron-oxidizing bacteria occur in the rhizosphere of wetland
plants at circumneutral pH. BIOGEOMON ’97. Journal of Conference Abstracts 2(2): 245.
Nickson, R.T., McArthur, J.M., Ravenscorft, Burgess, W.G. & Ahmed K.M. 2000. Mechanism of arsenic
release to groundwater of Bangladesh and West Bengal. Appl. Geochem. 15: 403–413.
Roach, J 2004. Arsenic in Asian Drinking Water Linked to Microbes. Published in National Geographic
News. URL: />Ravenscroft, P., McArthur, J.M. & Hoque, B.A. 2001. Geochemical and palaeohydrological controls on pol-
lution of groundwater by arsenic. In: Chappell, W.R., Abernathy, C.O., Calderon, R.L. (Eds.), Arsenic
Exposure and Health Effects IV pp. 53–77. Elsevier, Oxford.
Stumm, W. & Morgan, J.J. 1996. Aquatic Chemistry, Chemical Equilibrium and rates in natural water, 3rd
Edition, Wiley Interscience.
185
Copyright © 2005 Taylor & Francis Group plc, London, UK