Content uploaded by Gary Owens
Author content
All content in this area was uploaded by Gary Owens on Feb 19, 2015
Content may be subject to copyright.
REVIEW PAPER
Human arsenic exposure and risk assessment
at the landscape level: a review
Nasreen Islam Khan ÆGary Owens ÆDavid Bruce Æ
Ravi Naidu
Received: 7 March 2008 / Accepted: 17 September 2008 / Published online: 27 January 2009
ÓSpringer Science+Business Media B.V. 2009
Abstract Groundwater contaminated with arsenic
(As), when extensively used for irrigation, causes
potentially long term detrimental effects to the land-
scape. Such contamination can also directly affect
human health when irrigated crops are primarily used
for human consumption. Therefore, a large number of
humans are potentially at risk worldwide due to daily
As exposure. Numerous previous studies have been
severely limited by small sample sizes which are not
reliably extrapolated to large populations or land-
scapes. Human As exposure and risk assessment are no
longer simple assessments limited to a few food
samples from a small area. The focus of more recent
studies has been to perform risk assessment at the
landscape level involving the use of biomarkers to
identify and quantify appropriate health problems and
large surveys of human dietary patterns, supported by
analytical testing of food, to quantify exposure. This
approach generates large amounts of data from a wide
variety of sources and geographic information system
(GIS) techniques have been used widely to integrate
the various spatial, demographic, social, field, and
laboratory measured datasets. With the current world-
wide shift in emphasis from qualitative to quantitative
risk assessment, it is likely that future research efforts
will be directed towards the integration of GIS,
statistics, chemistry, and other dynamic models within
a common platform to quantify human health risk at
the landscape level. In this paper we review the present
and likely future trends of human As exposure and GIS
application in risk assessment at the landscape level.
Keywords Arsenic Exposure Landscape
Risk assessment Human health GIS
Introduction
Arsenic contamination of groundwater in Bangladesh
continues to be a widespread problem despite the best
international efforts of many countries including
Australia. Recent research has been directed to
studies of the transfer of As from groundwater to
soil to crops and the subsequent risk posed to human
health from ingestion. Extensive groundwater con-
tamination can have long term effects on the
environment and the landscape when extensively
N. I. Khan (&)G. Owens R. Naidu
Centre for Risk Assessment and Remediation (CERAR),
University of South Australia, Mawson Lakes Campus,
Mawson Lakes, SA 5095, Australia
e-mail: nasreen.khan@unisa.edu.au;
nasreen_ikhan@yahoo.com
N. I. Khan
Department of Geography and Environment, Dhaka
University, Dhaka 1000, Bangladesh
D. Bruce
The Barbara Hardy Centre for Sustainable Urban
Environments, School of Natural and Built Environments,
University of South Australia, City East Campus,
Adelaide, SA 5095, Australia
123
Environ Geochem Health (2009) 31:143–166
DOI 10.1007/s10653-008-9240-3
used for irrigation. Consequently this can have a
direct effect on human health when the primary use
of the landscape is to produce crops for human
consumption. Thus water contamination, soil con-
tamination, food contamination, and ultimately
human health are all interdependent. This interde-
pendence together with the heterogeneity in the soil
and groundwater environments has limited the
extrapolation of current research findings to the
landscape level because current research findings
are based on site specific risk investigations of As
exposure. A further constraint with the application of
published data to the landscape level is that limited
spatial/geographic information system (GIS) based
models currently exist which are capable of
incorporating site specific data on As in various
environmental media with the predictive capacity for
As contamination at the landscape level. This paper
reviews the current trends in risk assessment of As
with particular emphasis on the potential application
of GIS to human health and landscape level risk
assessment.
Arsenic is a widespread contaminant in the
groundwater aquifers of Bangladesh due to natural
geological formations (Islam and Nehaluddin 2002).
Arsenic occurs naturally in two main forms, arsenite
As(III) and arsenate As(V), where As(III) is consid-
erably more toxic than As(V) (Smith et al. 1998).
Under certain environmental conditions As can be
both bioavailable and mobile and can accumulate in
edible crops and fodder. Consequently ingestion of
affected crops and vegetables, meat from animals
ingesting contaminated fodder, and drinking of
contaminated water can potentially pose a serious
threat to human health. The recent large scale
incidents of human As poisoning in South East Asian
countries has led to a major socioeconomic, health,
and management crisis (Hassan et al. 2005). These
incidents are of international importance because
exposure to As is not a localized phenomena and is
not occurring in isolated areas or from point sources,
but is rather occurring at the landscape level.
Landscape has typically been defined in the
ecological framework as assemblages of habitat,
community, and its dynamic structure, function, and
spatial pattern (Forman and Gordon 1986, Hsu and
Cheng 1999). In this review paper landscape is
defined as a spatial extent and structure of a region
encompassing landuse and physiographic (i.e.,
terrain) as well as administrative boundaries. It is
therefore anticipated that the landscape will encom-
pass an area larger than a single village and that, for
chemical contaminants, landscape level risk can be
defined as the probability that an exposure event
occurs and causes potential impairment of human
health. Moreover, landscape level risk correlates the
environmental chemistry of contaminants with human
activities, management practices, and the subsequent
toxicological consequences. Previous authors have
stressed the need for a landscape level As risk model
due to the large spatial and temporal variation of As
concentrations in the environmental media (soil,
groundwater, and plants) with geographic region and
numerous different exposure pathways.
In the field of contaminated landscape research,
risk assessment has so far been used only in a
comparative and qualitative framework. Within this
framework, risk is used as an indicator, not as a
quantitative tool for assessing environmental and
human health impacts of soil and groundwater
contamination. However, implementation of quanti-
tative risk assessment and management at the
landscape level is difficult because risk is based on
both spatially and temporally distributed phenomena
as well as dynamic, physical, social, and cultural
environmental processes. Parameters related to the
level of As contamination risk are its extent, dynam-
ics, and temporal variation.
A GIS can provide a powerful and highly flexible
tool to quantify environmental processes that can
increase the sophistication of any risk assessment
methodology (Fedra 1993; Goodchild et al. 1993;
Maidment 1996; Mauro et al. 2000). Valuable quan-
titative information can be incorporated into risk
assessment procedures with the help of GIS and,
through spatial representation, the estimated risk
becomes more apparent, thus facilitating the decision
making process (Korre et al. 2002). GIS has been
widely used to visualize, integrate, and analyze
spatial data pertinent to evaluating changes in envi-
ronmental ecological systems (Fedra 1996; Frysinger
1996; Zandbergen 1998; Ortigosa et al. 2000). The
use of GIS software as a model integration frame-
work is often preferred because of the important role
of spatial and temporal dynamics in evaluating
complex ecosystem processes (Sydelko et al. 2000).
The risk assessment process is based on exposure
pathways that describe the means by which a receptor
144 Environ Geochem Health (2009) 31:143–166
123
is exposed to a contaminant or chemical of concern
(Mauro et al. 2000). Previous risk assessment models
have mainly focused on identification of exposure
and modeling of exposure pathways. Therefore, risk
assessment is a process of estimating the potential
harmful effects of chemical exposure to the environ-
ment where chemical related risk is a function of both
exposure and toxicity.
Risk ¼fExposure;ToxicityðÞ:
Exposure identifies the existence of a chemical in the
environment and the potential receptors that would be
exposed to that chemical, while toxicity is a function
of response of the receptor to a chemical and the
existence of a toxicity mechanism. Therefore, devel-
oping a risk assessment model is an iterative process
that links exposure pathways with toxicity and dose
response of receptors to identify contaminant hot-
spots and determine the best approach for protecting
humans and the environment. This paper reviews the
current methodologies available for integrating
human As exposure and risk assessment at the
landscape level.
Arsenic in the food chain
Humans may be exposed to As via a variety of
different exposure pathways. Typically the direct
ingestion of water or the water–soil–plant–human
transfer pathways are considered to be the most
important. Arsenic present in groundwater used for
crop irrigation or naturally occurring in the soil can
be accessed by plants through their root systems and
can be bioaccumulated in various parts of the plant.
Subsequently, humans can be exposed to As by
consumption of these plant products (fruits, vegeta-
bles, and grains). Other As transfer pathways include
water–soil–plant–animal–human transfer where the
exposure can be via consumption of either animal
products containing As, such as eggs or milk, or by
direct consumption of meat or fish. In the latter case,
fish is known to be high in As, but usually in nontoxic
organic forms, and is therefore not considered to be a
significant threat to human health. Scant research has
currently been conducted to extensively analyze As
in the food chain by collecting various foods to
identify and evaluate human exposure to As (Sapun-
ar-Postruznik et al. 1996; Bae et al. 2002;
Roychowdhury et al. 2002; Alam et al. 2003; Dux-
bury et al. 2003; Roychowdhury et al. 2003; Das
et al. 2004; Islam et al. 2004a,b; Sekhar et al. 2003;
Watanabe et al. 2004; Al Rmalli et al. 2005). In this
section the As content of various food chain compo-
nents including water, rice, and vegetables will be
reviewed.
Arsenic in foods and water
Food has been identified as one of the major sources
of As exposure where exposure can occur not only
when food is cooked with As-contaminated water but
also, in many cases, when the food itself is found to
contain significant concentrations of As (Naidu and
Skinner 1999; Del Razo et al. 2002; Abernathy et al.
2003; Watanabe et al. 2004; Patel et al. 2005). For
this reason As distribution and dynamics in the food
chain has been identified as one of the major research
areas for quantifying As exposure and consequential
risk.
A number of studies have reported total As
concentration in various foods (fruit, grain, vegetable,
fish, and milk) from several different countries
(Table 1). It is evident from Table 1that there is
considerable variation in the As content in various
foods between countries and within countries as well
as large variations of As content both between and
within specific food types. Compared with the
majority of the other countries represented in
Table 1, Bangladesh, West Bengal, and Mexico show
significantly higher concentrations of food As. The
elevated As content of food in Bangladesh and West
Bengal may be attributed to irrigation using As-
contaminated groundwater in the dry season and this
can be a significant route for both short and long term
As exposure to humans (Naidu and Skinner 1999;
Das et al. 2004). Significantly higher As concentra-
tions reported in Mexican food compared with other
countries may be related to the amount of water used
in food preparation and the cooking time, and
Del Razo et al. (2002) indicated that prolonged
cooking with contaminated water elevated the As
concentration in cooked food.
In comparison, some countries have relatively low
overall total dietary As intake. In a study of total As
intake from Croatian foods, the overall dietary intake
was very low (11.7 lg/person/day) and the highest mean
concentration of As was recorded in fish (498 lg/kg),
Environ Geochem Health (2009) 31:143–166 145
123
Table 1 Total As concentration (lg/kg) in food from different studies
Type of food Origin of food As concentration (lg/kg) References
Mean Range
Amaranthus Hyderabad, India 30 Sekhar et al. (2003)
Aram leaf Bangladesh 557 71–7850 Owens et al. (2004)
Aram stem Bangladesh 533 32–25,100 Owens et al. (2004)
Beef USA 51.5
a
Schoof et al. (1999)
Bottle gourd leaf
b
Bangladesh 306 Alam et al. (2003)
Brinjal Hyderabad, India 37 Sekhar et al. (2003)
Carrot USA 7.3
a
Schoof et al. (1999)
Chicken USA 86.4
a
Schoof et al. (1999)
Eggs
c
Mexico 320 10–1078 Del Razo et al. (2002)
Eggs USA 20
a
Schoof et al. (1999)
Freshwater finfish USA 160
a
Schoof et al. (1999)
Fish Bangladesh 350 9–1318 Al Rmalli et al. (2005)
Green papaya
b
Bangladesh 389 Alam et al. (2003)
Green amaranthus Bangladesh 264 52–1200 Owens et al. (2004)
Indian Spinach Bangladesh 185 31–646 Owens et al. (2004)
Ladies finger Hyderabad, India 31 Sekhar et al. (2003)
Milk Hyderabad, India 16 Sekhar et al. (2003)
Pasta Soup
c
Mexico 1003 450–1032 Del Razo et al. (2002)
Pinto Bean
c
Mexico 2070 20–5300 Del Razo et al. (2002)
Potato Bangladesh 12 \10–92 Owens et al. (2004)
Potato
c
Mexico 480 220–1043 Del Razo et al. (2002)
Potato USA 2.8
a
Schoof et al. (1999)
Radish Bangladesh 239 99–3690 Owens et al. (2004)
Red amaranthus Bangladesh 399 39–3070 Owens et al. (2004)
Rice Bangladesh 450 Naidu and Skinner (1999)
Rice (Boro)
b
Bangladesh 183 10–420 Duxbury et al. (2003)
Rice (Aman)
b
Bangladesh 117 10–420 Duxbury et al. (2003)
Rice
b
Bangladesh 228–377 Bae et al. (2002)
Rice
b
Bangladesh 136 40–270 Das et al. (2004)
Rice
b
Bangladesh 496 58–1830 Meharg and Rahman (2003)
Rice Central India 274 18 –446 Patel et al. (2005)
Rice USA 303
a
Schoof et al. (1999)
Rice grain Bangladesh 139 \10–557 Owens et al. (2004)
Rice grain Bangladesh 130 30–300 Williams et al. (2005)
Rice grain Europe 180 130–220 Williams et al. (2005)
Rice grain Spain 290–410
a
Laparra et al. (2005)
Rice grain USA 260 240–300 Williams et al. (2005)
Rice grain West Bengal, India 40 40–430 Williams et al. (2005)
Rice, cooked Bangladesh 145 15–1640 Owens et al. (2004)
Rice, cooked Spain 1410–2820
a
Laparra et al. (2005)
Sauce
c
Mexico 1016 160–3070 Del Razo et al. (2002)
Saltwater finfish USA 2356
a
Schoof et al. (1999)
Snake gourd
b
Bangladesh 489 Alam et al. (2003)
146 Environ Geochem Health (2009) 31:143–166
123
followed by fishery products (270 lg/kg) and dairy
products (39 lg/kg), while the lowest mean concen-
trations of As were detected in fruit, chocolate,
and fruit products, being 0.2, 0.2, and 0.3 lg/kg
respectively (Sapunar-Postruznik et al. 1996). Inor-
ganic As concentration in milk ranged from 1 to
160 lg/kg (Sekhar et al. 2003).
In general, dietary intake of As was related to the
geographic location of where the food was sourced,
reflecting the level of contamination at the foods
geographic source. Al Rmalli et al. (2005) compared
the As concentrations in foods (vegetables and fish)
imported from Bangladesh with those grown in the
UK and found that As concentrations in imported
vegetables from Bangladesh were 2–3 fold higher than
UK grown vegetables. The highest As concentration in
the imported vegetables were found in the skin of arum
tuber (540 lg/kg), the arum stem (168 lg/kg), ama-
ranthus (160 lg/kg), and in freshwater fish, total As
levels ranged between 97 and 1318 lg/kg. Watanabe
et al. (2004) identified that the people living in As-
contaminated areas ingested As from food at a level
approximately 30 times higher than people living in
non-contaminated areas due to direct exposure to As-
contaminated groundwater and food. A significant
proportion of indirect water As intake also occurred in
Bangladesh through frequent consumption of rice and
curries, which were generally cooked with a sub-
stantial amount of As-contaminated water (Watanabe
et al. 2004). The total intake of As must include
studies from all possible source media and not only
focus on groundwater exposure alone. Indeed, if As
intake from other media (food, soil, and air) were
ignored then the dose–response relationship between
As exposure and any health effect might be overes-
timated (Bae et al. 2002). Since As toxicity varied
with the As species, Watanabe et al. (2004) and Al
Rmalli et al. (2005) both indicated that it was
Table 1 continued
Type of food Origin of food As concentration (lg/kg) References
Mean Range
Spinach Hyderabad, India 38 Sekhar et al. (2003)
Spinach USA 5.1
a
Schoof et al. (1999)
Taro
b
Bangladesh 440 Alam et al. (2003)
Tomato Bangladesh 31 \10–212 Owens et al. (2004)
Tomato Hyderabad, India 39 Sekhar et al. (2003)
Tomato USA 9.9
a
Schoof et al. (1999)
Tortillas
c
Mexico 540 160–1013 Del Razo et al. (2002)
Total diet
b
Bangladesh 214/120 Watanabe et al. (2004)
Total diet Croatia 0.4 0–1.5 Sapunar-Postruznik et al. (1996)
Total diet
b
West Bengal, India 123 \0.04–690 Roychowdhury et al. (2002)
Total diet
b
West Bengal, India 81.1 \0.04–605 Roychowdhury et al. (2003)
Meat (beef and pork)
c
Mexico 1028 270–4014 Del Razo et al. (2002)
Oyster Taiwan 91 71–145 Liu et al. (2006)
Vegetables Bangladesh 54.5
d
\5–540 Al Rmalli et al. (2005)
Vegetables UK 24.2 \5–87 Al Rmalli et al. (2005)
Vegetables
b
Bangladesh 70–3990 Das et al. (2004)
Vegetables
b
Bangladesh 76.52–92.44 Islam et al. (2004)
All values are on a dry weight basis unless otherwise noted
a
As concentrations expressed as wet or fresh weight (FW) basis
b
The foods analyzed in these studies were collected from As-affected regions
c
As concentrations were derived from analysis of typical Mexican cooked food
d
The mean without As concentration in Arum tuber skin (40.5 lg/kg)
Environ Geochem Health (2009) 31:143–166 147
123
important for accurate risk assessment and toxicolog-
ical investigations to identify the nature of As species
in food and water.
Roychowdhury et al. (2002) investigated the con-
centration of As and other heavy metals in foods and
drinking water from two As-affected villages of West
Bengal and showed that mean As concentrations in
vegetables were 20.9 and 21.2 lg/kg, in cereals and
bakery products 130 and 179 lg/kg, and in spices 133
and 202 lg/kg for the Jalangi and Domkal villages,
respectively. However, the estimation of As concen-
tration in vegetables, cereals, and bakery products
were approximately three and seven times higher
than concentrations reported in a survey of similar
Canadian foods (Dabeka et al. 1993). High concen-
trations of As in the West Bengal study areas were
attributed to extensive use of As-contaminated
groundwater for drinking, cooking, other household
purposes (such as washing dishes, vegetables, rice,
and bathing), and irrigation. Moreover, most of the
vegetables and other foods for local consumption
were grown and available within these areas. How-
ever, mean As concentrations in tubewell and shallow
tubewell water were 10 and 0.018 lg/L, respectively,
which may contribute towards contaminations of
foods in these areas. However, since these concen-
trations are well below the Bangladeshi ‘‘safe’’ water
limit of 50 lg/L, we would anticipate that water was
not the only source of As exposure in this case and
that other foods such as rice must be contributing to
As intake.
Arsenic intake via rice consumption
Besides drinking water, cooked rice could be one of
the most important sources of As human exposure in
Bangladesh and West Bengal, especially if the rice is
cooked with As-contaminated water (Bae et al. 2002;
Duxbury et al. 2003; Meharg and Rahman 2003; Das
et al. 2004; Watanabe et al. 2004; Williams et al.
2005). This is because rice, being a staple food in
Bangladesh, comprises about 70% of total diet (Bae
et al. 2002), and is consumed in large amounts on a
regular basis. Williams et al. (2005) mentioned that,
even at background levels, ingestion of As in rice
contributes significantly in subsistence rice diets.
Meharg (2006) mentioned that, if rice containing only
0.1 and 0.2 lg of As were consumed, then intake of
As via rice consumption would be 17% and 30% of
the total As intake, respectively. An adult Banglade-
shi man consumes an average of 1,500 g/day fresh
weight of cooked rice, which contains approximately
1 L of drinking water (Bae et al. 2002; Watanabe
et al. 2004). Therefore, indirect water intake through
rice consumption would add substantial amounts of
ingested As to the total dietary intake of the
Bangladeshi population. The amount of As measured
in cooked rice was 10–35% higher than that predicted
from the concentration of As in uncooked rice,
suggesting that either As in water was chelated or
sorbed by the rice grains or that during cooking
evaporation resulted in As becoming more concen-
trated (Bae et al. 2002). The major limitation
identified in the Bae et al. (2002) study was that
they had analyzed only a very small number of rice
samples from Bangladesh. Williams et al. (2005)
analyzed 15 different rice samples comprising 13
from Aman rice (wet season) and 2 from Boro rice
(dry season) from the local markets of Dhaka city to
identify the As concentration and As species. The
mean As concentration in Bangladeshi rice grain was
120 lg/kg and the As concentrations ranged from 70
to 170 lg/kg. Arsenic speciation showed that the
main species detected in the rice grains were As(III),
DMA(V) and As(V).
Inorganic As contents in Bangladeshi, European,
US, and Indian rice samples were 81 ±4% (n=
15), 64 ±1% (n=7), 42 ±5% (n=12), and
80 ±3% (n=11). While Williams et al. (2005)
found highly significant differences (P\0.001) in
the percentage of inorganic As between Bangladeshi,
European, North American, and Indian market rice
samples, they found no significant difference
between white and brown rice (P\0.05) between
those countries. One of the major limitations of this
study was that it used only a small number of
samples, from the Dhaka city market, which were
supplied to the market from different parts of
Bangladesh. Moreover, these rice samples were not
locally grown in any of the highly As-contaminated
areas of Bangladesh (Williams et al. 2005). There-
fore, neither the As concentrations nor the percentage
of inorganic As reported is necessarily representative
of the whole country.
Duxbury et al. (2003) analyzed 150 rice samples
from various parts of Bangladesh and concluded that
‘‘human exposure to As through rice would be
148 Environ Geochem Health (2009) 31:143–166
123
equivalent to half of that in water containing 50 lg/L
for 14% of the paddy rice samples at rice and water
intake levels of 400 g and 4 L/person/day, respec-
tively,’’ estimating that a person would be exposed to
200 lg As per day simply by drinking 4 L of water per
day containing 50 lg/L As (the acceptable drinking
water limit in Bangladesh). Duxbury et al. (2003)
found that the As concentration in different varieties of
rice varied from 10 to 420 lg/kg at 14% moisture
content and that rice yield and grain As concentration
were 1.5 times higher in Boro rice, the winter (dry
season) variety, than Aman rice, the summer (wet
season) variety. The mean values for As concentration
in Boro and Aman rice ranged from 108 to 331 lg/kg
and 72 to 170 lg/kg, respectively, suggesting that
groundwater irrigation with As-contaminated water in
the dry season contributed towards the higher concen-
trations of As in the rice. Processing of rice (parboiling
and milling) has significant effect (P\0.001) on As
content in rice grain that resulted in an average 19%
reduction of As concentration in some (21) samples
collected from Bangladeshi households (Duxbury
et al. 2003).
Laparra et al. (2005) reported a 5–117 fold
increase of inorganic As content in raw rice when
cooking with contaminated water and concluded that
bioavailability, rather than total As content, provided
a better estimate of risk from As-contaminated
rice. In their study they used 12 rice samples from
Valencia, Spain, and found that at the lowest (3.9%)
and highest (17.8%) uptake of total As, consumption
of 5.7 and 1.2 kg of cooked rice/day would be
required to reach the tolerable daily intake (TDI) set
by the Food and Agriculture Organization (FAO)/
World Health Organization (WHO) for inorganic As
(150 lg inorganic As per day for a person weighing
70 kg) (Laparra et al. 2005). Comparing European
with Asian TDI, it can be concluded that the
populations from Asian As-contaminated areas might
reach the TDI with rice consumption alone, since
calorie intake mainly depends on cooked rice.
Therefore people in countries where there is a
reliance on rice as a dietary staple, and the rice
contains elevated levels of inorganic As, are most at
risk (Williams et al. 2005). Such countries include
Bangladesh where rice is the major food staple and is
a significant contributor to total daily As intake and is
therefore of major concern for researchers and policy
makers.
Arsenic total daily intake
A number of studies have estimated total dietary intake
of As from various geographic locations (Sapunar-
Postruznik et al. 1996; Roychowdhury et al. 2002;
Watanabe et al. 2004). Mean daily total As intake
from different countries is shown in Table 2. It was
observed that Croatia had the lowest dietary intake
compared with any other country. On the basis of
the level of As consumption of the sample food, the
estimated mean weekly dietary intake of As for the
Croatian population was 81.9 lg/person/week and a
corresponding daily intake of 11.7 lg/person was
calculated on the basis of per capita consumption,
amount, and level of As (Sapunar-Postruznik et al.
1996). However, this study did not consider the
frequency with which foods were consumed by each
person in the calculation of total dietary intake of As,
and their total dietary intake value was constructed
based on weekly consumption rather than daily
consumption. In Croatia the highest As dietary intake
was received through the consumption of fish and
dairy products (cheese), being 55.3 and 9.6 lg/person/
week, respectively. The contribution to As exposure
through fish was lower than that from dairy products
(cheese) as fish consumption was lower (25 g/person/
week) than dairy products (cheese) in Croatia, even
though As concentration in fish was higher than that in
cheese and dairy products. The global dietary intake of
fish is 231 g/person/week while in Croatia it was
136 g/person/week. The intake of As via consumption
of vegetables, fruits, and rice in Croatia was of little
concern, being only 0.3 lg/person/week when the rate
of consumption was 62 g/person/week.
In contrast to Croatia, studies in Bangladesh
indicated significantly higher dietary intake of As
from food compared with most other countries, with
the exception of Japan and The Netherlands
(Table 2). In Bangladesh As exposure from food
contributed 214 and 120 lg/day to the total diet for
males and females, respectively, (Watanabe et al.
2004), while the total daily intake of As for West
Bengal was 189 and 171 lg/day for males and
females, respectively (Roychowdhury et al. 2002).
In countries such as Bangladesh and West Bengal,
India, the consumption of rice and vegetables is
considerably higher and consequently the total daily
intake of As received through rice and vegetables
would be significant relative to fish, meat, and dairy
Environ Geochem Health (2009) 31:143–166 149
123
products. Roychowdhury et al. (2002) estimated that
the provisional tolerable weekly intake (PTWI)
values for inorganic As (lg/kg body weight/day)
were 11.8 and 9.4, 13.9 and 11, and 15.3 and 12 for
adult male, adult female, and children of Jalangi and
Domkal villages of West Bengal, respectively.
Most of the countries listed in Table 2had total
daily As intakes \220 lg/person/day; the U.S.
Environmental Protection Agency (USEPA) guide-
line value. One exception was Japan, which had one
of the highest daily As dietary intake of any of the
countries considered. Total daily As intake depends
on the type and amounts of food ingested, and the
consequential potential effects of As exposure depend
on the speciation of As in the food products. Thus
while the amount of As ingested via fish could be
high, its potential impact and risk would be much
lower than that from rice and vegetables due to the
form of As present in fish. In Japan, seafood is
prevalent throughout the diet and this higher total
value reflects the generally higher concentrations of
nontoxic organic As compounds prevalent in fish and
seafood rather than detrimental inorganic As expo-
sure. So Japanese ingesting more As via fish may not
have the same implications as that from ingesting
vegetables and rice containing inorganic As.
The maximum tolerable daily intake (MTDI) of
inorganic As through rice for Bangladeshi, North
American, European, and Indian populations was
calculated by Williams et al. (2005). MTDI of North
American, European, and Indian populations were
based on a 70 kg body weight and a rice consumption
rate of 0.5 kg/day and were found to be 8–15%,
21–50%, and 7–18%, respectively. In comparison to
these relative small values, the MTDI varied 4–88% for
the Bangladeshi population, which was based on 60 kg
body weight and a rice consumption rate of 0.5 kg/day.
This indicated that rice will contribute 1.7 lg As/kg
body weight/day. Furthermore, the total dietary intake
of As will increase to twice that of the WHO’s
maximum tolerable As intake when water consump-
tion is also considered with rice consumption.
Calculation of the MTDI for the Bangladeshi popula-
tions using a 70 kg body weight, which was the typical
body weight of the American and European popula-
tions, would overestimate the weight of an average
Bangladeshi (60 kg) and would therefore underesti-
mate the health effects resulting from As exposure
from rice, which will be significantly higher for the
lower body weight Bangladeshi population.
Human exposure assessment and GIS
There has been increased interest among environmen-
tal scientists, health practitioners, and policy makers in
Table 2 Worldwide
variation of mean daily total
As intake through food
a
Children from 1 to
6 years age
Country Mean daily total
As intake lg/person/day
References
Bangladesh 214 (males) Watanabe et al. (2004)
120 (females)
Belgium 12 Buchet et al. (1983)
Canada 59.2 Dabeka et al. (1993)
Canada 16.7 Dabeka et al. (1987)
Croatia 11.7 Sapunar-Postruznik et al. (1996)
Japan 182 Mohri et al. (1990)
Japan 160–280 Tsuda et al. (1995)
Mexico 394 Del Razo et al. (2002)
The Netherlands 38 Dokkum et al. (2007)
Spain 223.6 Llobet et al. (2003)
UK 65–67 Ministry of Agriculture, Fisheries,
and Food (MAFF) (1999)
USA 61.5 Gartrell et al. (1985)
USA 88 Gunderson (1995)
USA 3.2 (children
a
) Yost et al. (2004)
West Bengal, India 60.3–102 Roychowdhury et al. (2003)
150 Environ Geochem Health (2009) 31:143–166
123
continuing efforts to develop a spatially based model
that integrates GIS modeling techniques for risk
assessment. Contaminant exposure pathway identifi-
cation and its impact on human health are at the
epicenter of interest in environmental risk assessment
and human health studies. The heterogeneous spatial
distribution of contaminants requires an effective tool
to blend information from different disciplines for the
characterization of risk at various levels (landscape,
administrative, and geophysical). Environmental risk
assessment provides a qualitative and quantitative
description of the exposure pathway by which a
chemical of concern travels from source to receptor,
and the interactions associated with the chemical and
the transport media (Wilson 1997). Determining
exposure pathways is often difficult since contami-
nants are dynamic, complex, and numerous (Harris
1997). Exposure of chemicals in the environment is
spatial in nature, and exposure can therefore be treated
as a spatial object because it has certain characteristics
that can be linked with real world locations using
georeferencing in a GIS framework. GIS also provides
the architecture for developing predictive models
based on available data and is able to predict the
future consequences of chemical contamination in the
environment. Thus, GIS provides a useful framework
for exposure and risk assessment.
Human exposure pathways are complex systems
which are spatially and temporally variable (Harris
1997). While the origin of contamination may be a
single point source it may have multiple exposure
pathways. For examining the human exposure path-
ways there is a need to combine various sources of
information in a single framework to predict the
possible pathways and its impact on public health.
For effective estimation of risk there is a need to
integrate a huge amount of demographic, survey,
epidemiological, environmental, and physiochemical
data. GIS can provide the appropriate platform for
this integration. Risk assessment based on a GIS
approach has the potential to evaluate human expo-
sure in a quick and accurate manner. However, the
visualization capabilities of GIS provide a better
understanding of the contamination problem and its
impact on the human and physical environment.
According to Harris (1997) the accuracy of the
exposure pathway is highly dependent upon the
available amount of data. Chen et al. (1998) sug-
gested that the combination of risk assessment results
with spatial information would be meaningful for
identifying and assessing pollution impacts on spe-
cific receptors. However, the accuracy of the
identification of exposure pathway depends on the
amount of site specific data.
A number of approaches are currently being
developed to integrate human exposure assessment
with GIS for a variety of applications. Korre et al.
(2002) attempted to quantify and visualize lead (Pb)
exposure in soils and potential risk to human health
using a probabilistic approach within a GIS frame-
work. This risk assessment model was structured to
predict and assess the risk associated with high
concentrations of Pb in the soil. This probabilistic
approach to risk assessment has inherent uncertainty
which was overcome by determining the confidence
of the mean of the exposure distributions established
for each estimation grid point in a GIS (Korre et al.
2002). Duker (2005) established a relationship
between As exposure and the incidence of Buruli
ulcer using an integrated spatial statistical model.
Morra et al. (2006) developed a tool called HHRA-
GIS that provides the facilities to do the analysis in a
georeferenced structure to model multimedia, multi-
exposure pathways, and multireceptor risk caused by
an industrial source of contaminant and used this tool
to model cancer risk for children caused by dioxin.
Arsenic in water and urine correlation
Sekhar et al. (2003) identified in human exposure
pathway studies that the concentration of As in urine
ranged between 60 and 160 lg/L (control 6–10 lg/L)
at 14 different sites in Hyderabad city. They showed
that the major portion of As ingested through
drinking water and food was excreted in urine
(approximately 50%) with a small portion excreted
via the skin, hair, and nail. However, they did not
make any attempt to establish a relationship between
As excreted in urine and total As ingested through
drinking water and food with either different age
groups and/or gender. In a later study, Caceres et al.
(2005) used GIS for georeferencing sample popula-
tions and evaluated the relationship between
inorganic As exposure through drinking water and
total urinary As excretion and its variability with age
and geographic space. They found that the As
concentration in drinking water was the key contrib-
uting factor to exposure to inorganic As in the
Environ Geochem Health (2009) 31:143–166 151
123
Chilean population. They conducted exposure assess-
ment integration of demographic information,
exposure sources, time spent indoors and outdoors,
contact with soil, and direct and indirect smoking
with diet and water consumption. The sample pop-
ulation was divided into three groups: the elderly,
students, and workers. They observed a positive
correlation between As excreted in urine and total As
exposure through drinking water for workers and
students (R
2
=0.048 and 0.057) and a negative
correlation in the elderly population (R
2
=0.039)
living in Antofagasta city in Chile. It was evident
from their study that As intake and age were
significantly associated with individual level vari-
ables but there was no observed association with
gender, body mass index (BMI) or smoking habit.
Their results indicated that drinking contaminated
water increased the As concentration in urine by
0.53 lg/g/day, and therefore urinary As was a useful
indicator for As exposure. However, Caceres et al.
(2005) did not provide any detailed information
regarding their use of GIS, and As speciation was not
performed, so that the relative contributions from
nontoxic organic As and toxic inorganic As could not
be quantified. Also As intake was measured based on
an average concentration of As for a neighborhood
rather than a household, and no other drinking water
source was considered. Uchino et al. (2006) deter-
mined the concentration of As in tubewell water and
food composites (mainly vegetables and cereals) to
establish whether there was any correlation between
As dietary intake and As concentrations in hair and
urine. They observed a good correlation between As
concentration in hair and dietary intake (R
2
=0.452,
P\0.001) and a less significant correlation between
As concentration in urine and dietary intake
(R
2
=0.134, P\0.001).
Bioavailability of As
For assessment of exposure pathways, in addition to
chemical characteristics and environmental distribu-
tion of metal(loid)s, it is important to understand the
mechanism of distribution, excretion, storage, and
mobilization of metal(loid)s in the human body. In
the exposure study it is important to understand the
absorption mechanism and the factors that influence
this mechanism, i.e., cations in the diet. Moreover,
absorptive capacity varies among individuals and
with species and these control the amount of
metal(loid)s released from the contaminant and taken
into the blood stream.
Caussy et al. (2003) stated that bioavailability was
the focal point of the assessment of toxicity of
metal(loid)s. In general, bioavailability is the ability
of metal(loid)s to access different organs of humans
and animals and cause toxicity. Some scientist divide
bioavailability of metal(loid)s into: (1) external
(in vitro) bioavailability, sometimes referred to as
bioaccessibility, which refers to the ability of
metal(loid)s to be solubilized and released from
environmental media, i.e., soil, water, plant, and
food; and (2) internal (in vivo) bioavailability, which
refers to the ability of metal(loid)s to reach target
organs (human, animal, plant), be absorbed, and
cause toxicity.
Caussy et al. (2003) tried to understand the
diversity of exposure pathways and different factors
governing bioavailability. The mechanisms of envi-
ronmental exposure of humans and ecosystems
include the natural processes of metal(loid) release
via volcanic activity, erosion, and bioaccumulation,
anthropogenic processes such as mining, smelting,
industrial uses, and cultural practices. Generally,
environmental exposure to metal(loid)s may occur
via agricultural, industrial, and medical practice.
Authors typically defined exposure pathways after
Gochfeld (1998) as the combination of medium and
route of exposure. However, there are several
exposure pathways for each metal(loid) which
depend on the environmental media (air, water, soil,
food, and target population). Regional and local
consumption of food creates an ecoregion of
metal(loid) contamination where food plays an
important role in the exposure pathway. Metal(loid)
contamination pathways can also include occupa-
tional exposure routes, such as airborne As from
smelters and lead from vehicle exhausts. Bioavail-
ability of metal(loid)s results from the release of
metal(loid)s into the gastrointestinal (GI) tract,
where only a certain fraction of metal, with the
potential to cause toxicity, is absorbed and the rest is
eliminated by the GI tract (Caussy et al. 2003;
Robson 2003; Sekhar et al. 2003). This absorption
capacity depends on the intrinsic capacity of various
organs to adsorb metal(loid)s, as well as host factors
and varies with metal(loid), and metal(loid) species.
Arsenic specifically targets the skin, nervous system,
152 Environ Geochem Health (2009) 31:143–166
123
and blood, and long term exposure can result in
cancer.
While bioavailability can be measured directly in
the laboratory, in the field it can be measured using
concentration factors (CFs). Concentration factors are
the ratio of the concentration in an organism to the
concentration in its food or aquatic environment
(Caussy et al. 2003). Many factors can affect the
bioavailability of a metal(loid), including pH and
redox potential (Robson 2003). Robson (2003) iden-
tifies these factors as host factors and defined a host
factor as ‘‘any attribute of an individual that can
influence the amount and degree of metal(loid)
exposure, uptake, absorption, biokinetics and suscep-
tibility.’’ In the case of As exposure to a human
receptor the attributes of these factors are age,
gender, size and weight, nutritional status, genetics,
and some behavior which potentially results in both
acute and chronic human health risks (Robson 2003).
According to Robson (2003), biomonitoring was a
traditional method for assessing human exposure to
metal(loid)s. Biomonitoring is a useful supplement
for estimating the level of metal(loid) exposure and
for assessing various sources of metal(loid) uptake
through different exposure pathways (Christensen
1995).
Identification of metal(loid) toxicity using risk
assessment is usually based on the form of the
metal(loid) and implicitly assumes that the bioavail-
ability of that particular metal(loid) is equivalent for
all forms of metal(loid) and all kinds of exposure
pathways (Robson 2003). However, bioavailability of
metal(loid)s differs with their form and the exposure
medium that carries it to the receptor. So that the
bioavailability of As can vary from 100% for soluble
salts in water to less than 10% in mine tailings. The
bioavailability of As from soil is always considerably
lower than the bioavailability of As from waters.
Typically between 80% and 90% of a single dose of
uncomplex inorganic As(III) or As(V) is absorbed
from the human GI tract (Caussy et al. 2003; Pomroy
et al. 1980). As(V) generally absorbs more strongly
to mineral surfaces than does As(III) and is therefore
less mobile and significantly less bioavailable (Korte
and Fernando 1991). Abiotic and biotic factors can
both play an important role in the amount of
metal(loid) uptake, and the cation exchange capacity
can also influence bioavailability in soils and
sediments.
Laparra et al. (2005) suggested that, where rice
was the staple food, the contribution of inorganic As
from cooked rice should be considered in human
health risk assessment and that the determination of
inorganic As bioavailability in cooked rice provided a
better evaluation of human health risk than total As.
They found that after cooking rice with As-contam-
inated water there was a significant increase in
inorganic As content of rice. Using a Caco-2 cell
based intestinal epithelial model, gastrointestinal
digestion bioaccessibility of inorganic As reached
63–99%, and As retention, transport, and total uptake
(retention ?transport) in the Caco-2 cells varied
between 0.6% and 6.4%, 3.3% and 11.4%, and 3.9%
and 17.8%, respectively (Laparra et al. 2005).
Roberts et al. (2002) attempted to measure relative
bioavailability of As in soil using a primate model.
Five different soil samples from five different sites
(electric substation, cattle dip, two pesticide sites, and
wood treatment sites) were assessed by administra-
tion of an oral dose of soil to monkeys. Relative
bioavailability was measured based on urinary excre-
tion of As from the soil dose compared with excretion
following an oral dose of As in solution. Estimated
relative bioavailabilities ranged from 10.7 ±4.9% to
24.7 ±3.2% for the five different soil samples. The
result of the study indicated that some of the
experimental animals had a tendency to absorb more
As from soils than others, but the tendency was
within experimental error. They identified the pre-
dominant route of soil As exposure as being the
incidental ingestion of soil, and the relative bioavail-
ability of As was assumed to be 100%. Naturally, any
adjustment in the oral relative bioavailability from
the default 100% assumption would create a propor-
tional effect on the overall risk estimates which
indicates bioavailability of As in soil is equivalent to
the bioavailability of soluble As in water (Roberts
et al. 2002).
These previous bioavailability studies have high-
lighted an important point. When assessing risk it is
important to rely on species specific data because of
the variation of both bioavailability and toxicity with
species. Consequently analytical speciation is impor-
tant when evaluating bioavailability of As from the
food chain. It is important to quantify the contribution
of food to As exposure, and As from all sources need
to be speciated to accurately evaluate the dose
toxicity response. To date, assessment for risk of
Environ Geochem Health (2009) 31:143–166 153
123
metal(loid) exposure in various environmental media
(soil, water, air, and plant) has usually been based on
individual contaminant doses received from various
environmental media and compared with the safe
dose of a metal(loid). Since the primary source of As
exposure to humans is drinking water rather than food
ingestion (ATSDR 1989; WHO 2001; Robson 2003),
the exposure of As through drinking water is likely to
drive risk especially considering the high bioavail-
ability (*98%) of As in ingested water.
As exposure and human health effects
The severity of an adverse health effect due to As
exposure is related to the chemical form of As, innate
toxicity, exposure time, and dose (Tchounwou et al.
2004). Long term As exposure has led to a significant
number of health effects, i.e., hyperkeratosis, jaun-
dice, vascular diseases, and cancer of various organs
or tissues such as skin, liver, lung, and bladder (Guo
and Tseng 2000; Hassan et al. 2003; Yu et al. 2003;
Tchounwou et al. 2004). Guo and Tseng (2000)
detected a statistically significant association between
high concentrations of As in drinking water
([640 lg/L) and occurrence of bladder cancer, but
did not detect any association of As exposure with
lower concentrations of As in drinking water. The
temporal variability of As exposure via drinking
water is of major concern because the toxicological
effects (cancer) are considered to result from long
term chronic exposure, rather than short term events
(Abernathy et al. 1999; Ryan et al. 2000). Mazumder
et al. (1998) predicted that long term exposure to As
from water at 10 lg/L and 50 lg/L would cause
health affects to 46 million and 28 million people,
respectively. Among the affected people they esti-
mated 1,200,000 cases of hyperpigmentation,
600,000 with keratosis, 125,000 cancers, and 6,000
fatalities due to internal cancer per year. These
estimates are at present uncertain because of limited
data, but there is an option to input better data as they
become available into their developed methodologi-
cal framework.
In another study, Liu et al. (2006) estimated
human health risk associated with ingestion of
inorganic As through oyster consumption in Taiwan
where oyster consumption was between 18.6 and
56.0 g/day and observed that increased human expo-
sure to inorganic As through ingesting oysters
resulted in increased cancer risk. The resulting target
cancer risks (TR) ranged from 1.26 910
-5
(for
18.6 g/day) to 3.82 910
-5
(for 56 g/day) at the 95%
confidence level.
However, while there is strong evidence for human
carcinogenicity due to As, the exact mechanism for
producing tumors in the human body is not known.
Methylation is the major detoxification metabolic
pathway for inorganic As in the human body. Arsenic
toxicity and its effects on the human body can be
assessed using traditional risk assessment processes
such as risk identification, dose–response assessment,
exposure assessment, risk characterization, and risk
management (Caussy et al. 2003; Tchounwou et al.
2004). During dose–response analysis Tchounwou
et al. (2004) identified that the risk of As intoxication
increases as a function of level and duration of
exposure. Toxicity of As also depends on available
exposure routes, frequency of exposure, biological
species, age, gender, individual susceptibilities,
genetics, and nutritional sources. There was a positive
correlation between As levels in drinking water and
age adjusted mortality rates for various kinds of
cancer (Caussy et al. 2003; Tchounwou et al. 2004),
but the authors stressed the need for new epidemi-
ological studies to provide better information on the
dose–response relationship. They proposed that the
new study should consider a number of elements such
as population sample size, histories of As exposure
ranges, presence and absence of adverse individual
health effects with knowledge of their long term
exposure to As.
Arsenic exposure to humans occurs via ingestion,
inhalation, dermal contact, and the parental route. In
most cases diet is the main source of As exposure
with an average intake of about 50 lg/day, whereas
intake from air, soil, and water is generally much
smaller. However, in As-contaminated areas expo-
sure via air, soil or water can be significant. Since
metabolic processes differ from person to person,
even within a family, in order to determine an
individual’s correct dose–response relationship long
term exposure data of individuals over decades and
individual health histories would need to be consid-
ered (Tchounwou et al. 2004). It has not been
identified unequivocally that the consumption of As
in drinking water at the current maximum contam-
ination level (approximately 1 lg/kg/day) would
cause cancer, although WHO suggested that a total
154 Environ Geochem Health (2009) 31:143–166
123
daily intake of 2 lg/kg/BW of inorganic As by
humans may cause skin lesions within a few years.
Therefore, existing characterization of risk at the
maximum contamination level relies on extrapola-
tion, which is based on observed epidemiological
results, experimental data, and information regarding
human susceptibility. Therefore, studies need to be
conducted to identify the real value for maximum
contamination level in order to characterize the real
risk posed by As.
The investigation of environmental exposure
requires an interdisciplinary approach (Caussy et al.
2003). In order to quantify risk assessment and to
identify risk management protocols of metal(loid)
exposure it is important to include a number of
elements such as size and characteristics of the
exposed population, the natural distribution of
metal(loid), its various pathways of exposure,
metal(loid) species, and bioavailability. Therefore,
information regarding metal(loid) concentration,
chemical speciation, and chemical characteristics of
metal(loid) are the main contributing factors that
need to be identified. For assessing exposure they
proposed that five general systematic steps should be
followed, consisting of
Identification of metal(loid)s
Adequate data on metal(loid) concentrations
Speciation of metal(loid)s
Direct estimate of bioavailability
Incorporation in exposure assessment
Sekhar et al. (2003) identified the following ele-
ments as being essential for the assessment of As
exposure on human health:
Evaluation of As sources
Evaluation of As concentrations in environmental
media (soil, surface, and groundwater) and other
media such as vegetables, fodder, and milk
Identification of exposure pathways and determi-
nation of resultant ambient levels from source
Identification of population exposed to As
Sekhar et al. (2003) tried to achieve risk and
exposure assessment based on speciation studies on
soil, water (ground and surface), plant, and milk.
Human pathway studies were conducted using a
questionnaire survey in conjunction with collecting
and analyzing blood, urine, nail, and hair samples.
Results of the soil speciation showed that soils with
low metal(loid) retention capability resulted in
increased As mobility. The two processes controlling
As mobility in groundwater aquifers were (i) adsorp-
tion and desorption reactions and (ii) solid phase
precipitation and dissolution reactions. Plant uptake
studies showed that plant As uptake was passive, and
that As was translocated to roots and leaves, where
natural As concentrations rarely exceeded 11 mg/g.
The leaf-to-soil concentration ratio showed that As
translocation to leaves was about 0.2. The main
contributing factor for concentration of As in vegeta-
bles was the soil As concentration. Arsenic concen-
tration in milk ranged from 0.01 to 0.16 mg/kg due to
repeatedly feeding forage grass, high in As, to cattle,
which indirectly increased As exposure to humans via
food chain transfer.
In the exposure assessment framework Serre et al.
(2003) have emphasized the knowledge-based (KB)
approach to the adverse health effect of As contam-
ination across space and time and demographic
criteria. Human health exposure due to bladder
cancer caused by As ingestion was investigated in
their research. They indicated that the developed
approach is flexible and could be used for identifying
other health effects. The population expected to
suffer a health effect due to As ingestion through
contaminated drinking water was expressed in terms
of a probability. The resultant map showed that 45
lifetime cancer incidents were expected to occur per
km
2
, which roughly corresponding to three incidents
of bladder cancer every 4 years for a life span of
58 years. However, quantification of human health
exposure is not clearly shown in their map. They also
expressed human health exposure in terms of prob-
ability. The disadvantages of this research were that
only As intake from water was considered and total
dietary intake from other sources such as foods was
not included.
Arsenic toxicity and nutrition
The nutritional status of a landscape plays an
important role in the rate and magnitude of observed
human As toxicity. A number of studies have found
that people in Bangladesh with poor nutritional status
are more susceptible to arsenicosis compared with
people with better nutritional status (Hadi and
Parveen 2004; Hasnat 2005; Ahmad 2007). The
interaction between clinical symptoms of As toxicity
Environ Geochem Health (2009) 31:143–166 155
123
and BMI was inversely related, and poor nutritional
status in combination with high level of As in water
and large amounts of daily water intake can signif-
icantly increase the risk of adverse health effects
from As exposure (Islam et al. 2004a,b). In one of
the As-affected areas of Bangladesh the chronic
energy level was estimated at 47%, and 69% of the
total arsenicosis patients were malnourished
(BMI \20 kg/m
2
) (Ali et al. 2002). In this study
the mean concentrations of As in hair and nail were
20.28 ±15.05 and 5.82 ±4.97 mg/kg, respectively,
where females showed a significantly higher concen-
tration of As in nails than males. Other surveys
(Sekhar et al. 2003) indicated that the numbers of
arsenicosis patients were larger in the age group
[40 years and this may be attributed to prolonged
exposure to As being necessary before symptoms of
disease actually become evident. Vitamin C and
methionine reduced As toxicity (Roychowdhury
et al. 2002; Sekhar et al. 2003), but deficiency of
vitamin A increased sensitivity to As (Roychowdh-
ury et al. 2002). Another study indicated that there
was no significant relationship between nutritional
intakes and the percentage of dimethylarsenic acid
(DMA) measured in urine, but significant association
was observed between dietary intakes of methionine,
vitamin B-12, calcium, protein, riboflavin, and vita-
min A and the percentage of monomethylarsonous
acid (MMA) in urine (Heck et al. 2007). Moreover,
high carbohydrate, high protein or high fat diets
along with calcium, folate, and dietary fiber can
enhance the effects of As toxicity (Roychowdhury
et al. 2002; Sekhar et al. 2003; Mitra et al. 2004). In
addition Sekhar et al. (2003) found a positive corre-
lation between the consumption of low nutritious
food and increases in As human toxicity. As
discussed in the next section, nutritional status and
associated toxicity are most often manifested using
human biomarkers.
Biomarkers for As exposure assessment
Biomarkers have a great potential in epidemiological
and toxicological research for assessing risk posed by
exposure to chemicals (Harrison and Holmes 2006).
Therefore the use of biomarkers would be appropriate
for the measurement of exposure or the effects of
exposure, since biomarkers can help to understand
disease processes and provide a good basis for risk
assessment (Caussy et al. 2003; Robson 2003; Chen
et al. 2005). Chen et al. (2005) have defined bio-
markers as early biological effects of ingested
inorganic As, which include blood levels of reactive
oxidants and antioxidant capacity, genetic expression
of inflammatory molecules as well as cytogenetic
changes. Chen et al. (2005) found that the As levels
in urine, hair, and nail were biomarkers for short term
(\1 year) internal dose and the effects of long term
internal doses were skin hyperpigmentation and
palmoplantar hyperkeratosis, while the percentage
of monomethylarsonic acid in total metabolites of
inorganic As in urine were considered good exposure
markers for biologically effective dose. Ingested and
inhaled As caused systematic As toxicity and directly
affected methylated metabolites. There was a signif-
icant association between As in drinking water with
As in urine, hair and toenail. Inorganic As metabo-
lites in urine, hair and toenail increase by 0.18 lg/L,
0.9 ng/g, and 2.7 ng/g, respectively for every 1 lg/L
increase in well water As concentration. Adair et al.
(2006) identified that the total arsenic (TA) in blood,
urine, and toenails ranged from below detection to
0.03, 0.76, and 12 lg/L, respectively. They suggested
that toenails are better biomarkers of As exposure
than urine because the As contents in toenails have
been consistent among samples collected years apart.
Therefore toenails provide long term integrators of
As exposure which does not occur in urine. In another
study Anawar et al. (2002) estimated that 95% of
nails, 96% of hair, and 94% of urine samples
contained As above the normal level in some As-
affected districts in Bangladesh. Exposure to
0.103 mg/L As in drinking water produced dermato-
logical disease in the As-contaminated districts of
Bangladesh (Anawar et al. 2002), and therefore As-
induced skin lesions may be considered a long term
biomarker of cumulative As exposure. Moreover long
term ingestion of As can induce atherosclerotic
diseases including peripheral vascular disease, ische-
mic heart disease, and cerebral infraction in a dose–
response relationship. Further studies on associations
between biomarkers including biologically effective
dose, early biological effects, altered structures and
functions of other organs as well as genetic suscep-
tibility will provide better understanding of As-
induced health hazards. Therefore there is a need
for long term follow up of As-exposed people and
repeated collection of biospecimens.
156 Environ Geochem Health (2009) 31:143–166
123
Roychowdhury et al. (2002) and Sekhar et al.
(2003) concluded that it was very difficult to draw
any correlation between As concentrations in hair/
nail/blood/urine with age and the number of person
exposed to As, as other factors contributed to human
exposure to As such as duration, quantity and quality
of water, amount and frequency of As-contaminated
food/vegetables consumption, and individual health.
The question is: are blood/urine/hair analyses reli-
able indicators of As exposure at all? Hair samples
unless properly washed and prepared can be con-
taminated by ambient dusts or soils, giving
overestimates of As exposure. Likewise blood and
urine As levels may not accurately represent toxic
As exposure because As is generally readily excreted
from the body and such samples may give false high
As levels if seafood forms a substantial part of the
dietary intake during the survey period. While these
biological indicators/samples can be taken and
analyzed effectively under laboratory conditions,
such as in animal feeding trials, their utility and
accuracy for determining As exposure in the field
will be less reliable.
Present trend in As contamination risk assessment
Recently a number of studies have used GIS to map
As concentration levels in groundwater, location of
tubewells based on safe and unsafe sources of
drinking water, distribution of As-affected popula-
tions, and spatial relationships between As exposure
and human health hazard (Albertson et al. 1993;Aral
and Maslia 1996; Zhang et al. 2001; Aslibekian and
Moles 2003; Hassan et al. 2003; Serre et al. 2003;Yu
et al. 2003; Basu and Sil 2004; Duker 2005). It is
evident from recent and ongoing studies that the
present trend in As risk assessment of human
exposure is to integrate physical (physiography,
geology, soil, land use, elevation, hydrology), socio-
economic (demographic, economic condition), in situ
(soil, and water samples collected from the field and
analyzed in the laboratory), and biospecimen (blood,
urine, hair, and nail) samples into a GIS based
database. Current research by the authors aims to
develop a GIS based model to quantify the As
contamination risk to human health by integrating
physical, socioeconomic, in situ, and biospecimen
datasets within a landscape.
In a number of recent studies, researchers have used
GIS as a framework for As risk assessment. A
geostatistical technique, such as Kriging, was per-
formed to evaluate and characterize risk, and the
resultant risk was presented as a function of probability
(Hassan et al. 2003; Serre et al. 2003; Zhang et al.
2000; Zhang et al. 2001; Yu et al. 2003; Serre et al.
2003; Gay and Korre 2006; Goovaerts et al. 2005). A
few attempts were made to develop and incorporate
human health risk assessment as a tool in GIS to assess
and quantify risk (Hargrove et al. 2001; Zhang et al.
2001; Hellweger et al. 2002; Hassan et al. 2003; Bien
et al. 2004; Duker 2005; Morra et al. 2006; Gay and
Korre 2006). The raster GIS data structure has the
ability to perform cumulative risk calculations (Hell-
weger et al. 2002). The Map Spreadsheet presented by
Hargrove et al. (2001) is an ideal format for the spatial
presentation of risk assessments. Within the Map
Spreadsheet a perspective of the combined risk from
the exposure to multiple contaminants could be
obtained and heterogeneity of contamination across
the space could be mapped. A GIS based model
Human Health Risk Evaluation Tool (HIRET) was
developed by Bien et al. (2004) in conjunction with
risk assessment and spatial planning. The HIRET
model identifies potential human health risk with
respect to land use. The output of the model identified
the areas of potential human health risk using various
datasets within a GIS framework.
Present trends in As risk assessment stress speci-
ation of As in food together with determination of As
bioavailability from foods using animal models.
GIS and risk assessment
GIS can be used to represent the landscape by means
of spatially and temporally referenced data describing
the character and shape of geographic features and
their interaction with various environmental media.
GIS has been widely used in environmental modeling
and risk assessment (Albertson et al. 1993; Hiscock
et al. 1995; Lovertt et al. 1997). GIS and environ-
mental modeling developed separately, so their
computer programs have very different data struc-
tures, functions, and methods for inputting and
outputting spatial information. This separate histor-
ical development of GIS and environmental modeling
makes it difficult to link existing GIS systems with
Environ Geochem Health (2009) 31:143–166 157
123
existing environmental models, or even to link one
model with another. Some of these difficulties can be
overcome by embedding environmental models into a
GIS and thus make use of data from GIS structures
(Maidment 1996).
However, most of the environmental models are
limited to interactive visualization of the representa-
tion of the model result (Stein et al. 1995; Bober
et al. 1996). Although some research has been
conducted towards the integration of GIS with an
environmental model (Fedra 1993; Emmi and Horton
1996), very few studies have been conducted for
developing a standalone environmental risk model at
the landscape level within a GIS framework.
Risk assessment model data requires easy transla-
tion into simpler forms for use by political decision
makers. The spatial risk model output can be easily
translated and interpreted with the help of maps
(Bartels and Van Beurden 1998). Thus maps of risk,
metal(loid) concentrations, and exposure pathways
can optimize the subsequent contaminant site inves-
tigation. GIS and risk are discussed as cultural
problems in the policy arena, with the different
paradigms and views of scientists, policy makers, and
the public coming together (Rejeski 1993). GIS has
successfully been linked with risk assessment on a
number of occasions. Wadge et al. (1993) used GIS
for natural hazard assessment, and similar applica-
tions and approaches have been used by Parrish et al.
(1993) to identify indices for ecological risk and
overlay analysis of groundwater vulnerability studies
(Barrocu and Biallo 1993; Engel et al. 1996; Al-
Adamat et al. 2003). Risk mapping of potential
pollution was also accomplished by combining GIS,
remote sensing (RS), and the DRASTIC model (Al-
Adamat et al. 2003), while others have applied
similar approaches to investigate groundwater vul-
nerability to human exposure and geological hazard
assessment (Luzi 1995; Emmi and Horton 1996).
Kelly and Lunn (1999) developed a prototype
contaminated land assessment system (CLASS)
based on Arc/Info GIS software to simultaneously
map contaminated land and to classify the site in
terms of its future pollution potential.
These examples all used static approaches to
characterize risks, which are mainly based on the
overlay analysis available in GIS. The static risk
modeling approach is used due to limited information
being available on the dynamic and temporal
variables of contaminant datasets because their
collection is time consuming and costly. Mauro et al.
(2000) developed a GIS based risk model to represent
continuous evolution of the system over a series of
time steps. They generated model input using a fate
and transport model for groundwater and integrated
their data with land use with the help of a cell based
GIS spatial analysis technique. Their model was used
to assess the risk in various landuse units caused by
chlorinated solvents and petroleum hydrocarbons.
Another application of GIS in the field of exposure
and risk was conducted by Sabel et al. (2000)in
which they modeled space time patterns of motor
neuron risk using the kernel estimation calculation
technique in the GRID component of the Arc/Info
software (ESRI, USA). Foster and McDonald (2000)
have conducted research on GIS based risk assess-
ment of public water supply intake. In their research
they evaluated the applicability of various GIS
techniques to identify and evaluate risk, i.e., GIS
overlay technique is suitable for hazard identification,
probabilistic techniques in GIS are capable of mod-
eling risk, and overall GIS is suitable for derivation of
monitoring strategies.
Zhang et al. (2000) assessed the risk of airborne
hazardous materials (chlorine, ammonia) transported
through networks using GIS and modeled the
contaminants using a Gaussian plume model. They
modeled the probability of unexpected consequences
(i.e., injury, illness, and death) as a function of
contaminant concentration and treated risk as the
product of population affected and the probability of
unexpected consequences. Map algebra techniques
available in raster based GIS were widely used in
their study to estimate risk at any point of the
network. Aslibekian and Moles (2003) identified the
extent of soil contamination by Cd, Pb, and Zn near a
silver mine in Ireland to assess environmental risk.
Their risk assessment approach related to individual
metal(loid) contamination, spatial extent of contam-
ination, and contamination targets and they used GIS
based mapping and site investigation techniques
widely to assess risk from metal(loid) contamination.
Various geostatistical techniques have been used to
investigate and map soil contamination by heavy
metals (Lin 2002; Romic and Romic 2003; McGrath
et al. 2004). The geostatistical technique of Kriging
was used by Lin (2002) to estimate the spatial
distribution of Pb in soil. Distribution and risk zone
158 Environ Geochem Health (2009) 31:143–166
123
were calculated based on the level of Pb present in
soil and the relationship between Pb levels and soil
depth. The study indicated that Kriging was a useful
tool which can delineate risk zones in heavy
metal(loid) contaminated sites.
To assess nitrate leaching in agricultural areas
integration of PC ArcCad GIS software and ground-
water loading effects of agricultural management
system (GLEAMS) model version 2.0 was used
(de Paz and Ramos, 2002). A graphical user interface
(GUI) was developed for easy operation by nonex-
perts to produce a thematic map (pollution risk map)
of nitrogen leaching. The pollution risk maps showed
that the irrigated areas had the highest potential of
groundwater nitrate contamination. In a recent study
Gay and Korre (2006) presented a methodology to
assess human heath risk with the combination of
quantitative probabilistic and spatial statistical meth-
ods from exposure to contaminated land. They used
GIS to assess the spatial distribution and variability of
soil contamination and spatial distribution of popu-
lation at risk. To test their methodology they used
various steps. Firstly, they used geostatistics for
assessing the level of soil contamination to produce
spatial distribution maps. Secondly, they mapped age
stratified populations across the contaminated area.
Thirdly, they calculated the probabilistically person
specific intake of soil contaminant following the
procedures used in the contaminated land exposure
assessment (CLEA) model to identify where popula-
tions were at risk. Another example is presented by
Fytianos and Christophoridis (2004) to map the level
of concentration of drinking water pollution caused
by nitrate, arsenic, and chloride. In their mapping
exercise they first mapped the locations of various
contaminants and in the second step mapped the
spatial variation of concentration of the pollutant.
However, due to the limited number of samples, the
GIS based maps failed to provide an accurate picture
of concentration of metal(loid) in drinking water.
In summary, GIS has been identified by Foster and
McDonald (2000) as a useful computing technique for
pollution risk based on spatial and attribute data.
Moreover, it can contribute to enhanced pollution risk
assessment through storage, analysis, and management
of highly variable environmental data. Therefore, GIS
is expected to continually shape the research and
practice of risk assessment and management of chem-
ical contamination and its effect on human health.
GIS and arsenic risk assessment
A number of international studies where GIS has been
specifically applied to As risk assessment are dis-
cussed in this section. Yu et al. (2003) applied
geostatistical methods to create As concentration
maps of Bangladesh by dividing the country into
regions based on concentration and estimated the
vertical As concentration trends in these regions.
Census data were used to evaluate exposure distribu-
tion of the whole country and epidemiological data
from West Bengal (Mazumder et al. 1998) and
Taiwan (Tseng 1977) and were used to estimate the
dose response for As-exposed people in Bangladesh.
Integration was made between regional exposure
distribution and the dose–response model to estimate
the health effects caused by groundwater As. They
suggested that replacement of 31% of the existing
shallow tubewells with deep tubewells would reduce
approximately 70% of the As concentration in drink-
ing water. While this may seem a simple option it does
not address the cost associated with replacing 31% of
wells in a country where the average income is
approximately 120 USD per annum (CIA 2006). The
number of shallow tubewells in Bangladesh is
approximately 924,023 (BADC 2003), so 31% of
the tubewells would correspond to 286,447 tubewells.
Replacing deep tubewells at a cost of around USD 100
per well would equate to an outlay cost of USD
28,644,700 and is well beyond the government’s
ability to fund. Therefore, this cannot be accepted as a
reasonable option to combat As contamination in
Bangladesh.
A GIS database was constructed by Zhang et al.
(2001) and was used to identify the As distribution
profile of Thailand in various environmental media
(deep and shallow groundwater and near surface soil).
GIS interpolation estimates showed that 72.3% of the
study area had As level higher than 50 lg/L in shallow
groundwater in Thailand. With the aid of the innate
GIS based distance function they identified a correla-
tion between As contamination level in soil and
groundwater associated with mining and waste dump-
ing activities. In this study, drinking water was the
major pathway for As exposure to humans and they
found a positive correlation between As concentration
in drinking water and cancer (bladder, lung, liver,
kidney, and prostate). The total average exposure to
inorganic As from food, water, and other beverages
Environ Geochem Health (2009) 31:143–166 159
123
was 17 lg/day where drinking water contributed 5 lg/
day (the EPA standard reference), but the exposure
estimated from one of the study sites was 79.1 lg/day,
which far exceeded the EPA estimate. Estimated
values for cancer risk also exceeded the EPA standard
reference values in the study area.
Serre et al. (2003) suggested that a GIS based
holistochastic framework could be used to identify
human health exposure due to As contamination. The
holistochastic framework is based on Bayesian max-
imum entropy theory, which offers powerful tools to
assimilate a variety of knowledge bases (physical,
epidemiologic, toxicokinetic, demographic, etc.) and
uncertainty sources (soft data, measurement errors,
etc.). Using this holistochastic approach Serre et al.
(2003) created a number of As distribution maps for
Bangladesh drinking water. Reis et al. (2005) applied
geostatistical methods to a soil dataset to predict As
concentration at unsampled locations and used geo-
statistical methods to produce probability maps for
As to assess the contamination level of the soil due to
mining activities.
Hassan et al. (2003) have used GIS based data
processing to analyze and map groundwater As con-
centrations and to identify risk zones in selected areas of
Bangladesh. They used multiple methods combining
spatial and attribute information along with a question-
naire surveyrelated to As concentration in groundwater.
The pattern of As concentration and its spatial variabil-
ity was identified using spatial interpolation methods
such as Kriging. They estimated As concentration on a
sample population to be 12 lg/day based on water
intake and water usage. In this estimation they assumed
that (i) As concentration in drinking water was 50 lg/L,
(ii) body weight was 60 kg, (iii) water intake was
2 L/day, and (iv) water used in cooking was 2 L/day.
The relationship between As concentration and distance
between sample tubewells was established, where 46%
of sample tubewells which were located within 25 m of
each other showed significant variation in As concen-
tration. To assess the human health effects they
identified a few individuals having long term exposure
to As (18 and 26 years). Exposure to As in water at
concentrations exceeding 446 and 353 lg/L caused
hyperkeratosis and gangrene, respectively. Their study
showed that the long term exposure to As at the rate of
50–100 lg/L As was four times more likely to cause
arsenicosis symptoms in people than when exposed at a
‘‘safe’’ level\0.10 lg/L.
Duker (2005) has applied integrated spatial prox-
imity analysis and its statistical techniques to model
As exposure response and relationship with Buluri
ulcers. It was evident from the model result that there
was a significant relationship between As concentra-
tion in surface water and Buluri ulcers occurrences
(R
2
=0.82, p\0.05) but no such relationship was
observed between As concentration in groundwater
(R
2
=0.10, p=0.60). Close proximity between
settlement and artisanal mine sites showed higher
Buluri ulcers incidence (R
2
=0.48, p\0.01) and, as
the proximity increased between settlement and
artisanal mine sites, the incidence of Buluri ulcers
decreased (R
2
=0.00, p=0.90). Contributory fac-
tors for Buluri ulcers identified in this research were
ingestion of As-contaminated water and food crops
grown on As-contaminated soil.
Arsenic risk in the spatio-temporal context
Arsenic contamination is spatially variable through
natural processes (sedimentation, bed rock formation,
and weathering) and human activities (irrigation,
pesticide, and groundwater extraction), and As
exposure to humans is time dependent. Ingestion of
As through water and food has both long and short
term human health effects. Arsenic exposure through
drinking water is associated with gastroenteritis,
neurological manifestations, vascular changes, dia-
betes, and cancers (bladder, lung, liver, skin, kidney,
and prostate) (Abernathy et al. 2003). Environmental
media responsible for As contamination have spatial
dimensions, that is As contamination occurs in a
horizontal and vertical space and the extent of As
contamination is spread over an area rather than a
single point. The extent and severity of As exposure
within a landscape also varies with time and conse-
quently the type of human health risk caused by As is
also spatially (Duker 2005) and temporally variable.
Spatial variability of As exposure to human health
is due to the variability of socioeconomic conditions
of the space, nutrient level, and BMI. Temporal
variability of As exposure to human health risk is
attributed to the duration and frequency of ingestion
of As-contaminated food and water; for instance, As
concentrations in tubewells can fluctuate with season,
being diluted in the rainy season and become more
concentrated in the drier months when water levels in
160 Environ Geochem Health (2009) 31:143–166
123
wells are lower. In addition irrigation practices for
food cultivation may change over time or with
season. Therefore, exposure risk varies temporally
as well as spatially.
Arsenic exposure is a function of the amount and
duration of ingestions from more than one source/
media (water, food, soil, and air). The As content of
the source/media can vary depending on groundwater
As concentration as well as geographic location. So,
the potential risk from As exposure can vary based
on a space time composite. Therefore, in risk
assessment processes, preference should be given to
a spatio-temporal context to quantify accurate risk
values.
Conclusion and future direction of human As
exposure risk assessment
The issue of human health risk assessment is of major
concern worldwide and there is a need to embrace
new technologies that can assist in mapping, quan-
tifying, and predicting the associated risk to a given
contaminant. Integration of existing information into
GIS and the associated statistical modeling available
in such packages is one option that has the potential
to significantly enhance the ability of health practi-
tioners and governments to appropriately assess and
manage risk. A number of efforts have been made by
researchers to apply GIS based geostatistical
techniques for investigating and mapping As con-
tamination in soil and water and assessment of human
health risk.
However, many previous studies have been limited
in the scope of their analysis, tending to be of small
sample size, and making broad extrapolations on the
basis of a few selected vegetables and other foods
from often a single isolated village. Often this is
justified by considering the village studies as being
the worst case scenario with the majority of other
villages being considered to be far less at risk of As
contamination and subsequent exposure to the human
population. Since all of these studies were carried out
on a relatively small scale and none of the research
has been completed at the landscape scale, small
scale research has failed to provide the spatial pattern
and variability of As contamination in various
exposure sources and pathways over the landscape.
However current progress in this field to date is not
without some limitations that will need to be
addressed as the field grows. Some of the general
areas that need to be addressed are:
a. Sample size
b. Study scale
c. Appropriate data
d. Spatio-temporal modeling of As contamination
e. GIS and geostatistical/statistical analysis
techniques
f. Human health risk assessment based on multi-
exposure pathways
g. Future trends
The major conclusions from this review indicate
that several specific points need to be addressed in the
future, as discussed briefly below.
1. The majority of the total dietary intake studies
reported in the literature to date are based on
extremely small sample sizes, at times as few as
ten samples. Therefore, in the assessment of
human exposure to As concentration large
extrapolations are required to account for As
exposure for any significant population with
inherit error. In the future, studies involving
much larger sample sizes are required to
quantify As exposure from various food and
As contamination in the food chain without
extensive extrapolations.
2. For As risk assessment there is a need for
synergy between GIS and laboratory analysis
data. GIS can provide visual representation
(maps) of risk which can be easily understood
by the public and government policy makers.
3. Future risk assessment based on an integrated
GIS approach would allow quantifying risk
rather than simple concentration mapping.
4. There is also a need for integrating spatial and
temporal dimensions in the As contamination
risk assessment process.
5. Future research should concentrate on assessing
risk at the landscape level (area) rather than as
discrete points to provide better options for As
remediation policy measures.
6. Future As risk assessment research should use an
interdisciplinary approach, where medical prac-
titioners, statisticians, hydrogeologists, environ-
mental scientists, geographers, sociologists,
Environ Geochem Health (2009) 31:143–166 161
123
economists, and modelers are all involved to
address the broad scope of As contamination,
which is a multidisciplinary issue.
7. More detailed information needs to be gathered
on information that is important in quantifying
As exposure. This information includes As
interactions within the terrestrial food chain and
how As transfers to humans and causes the
various human health problems.
8. While the focus of much research has been
inventing and offering techniques for remedia-
tion of As-contaminated water, limited effort
has been directed towards assessing the associ-
ated risk that groundwater and crops grown in
contaminated soils pose to human health at the
landscape level.
9. A large number of models already exist and are
used in various disciplines to model different
aspects of an environmental contamination
issue. What is lacking is a system of incorporat-
ing pre-existing information in a cohesive unit.
Therefore there is a need to couple GIS with
other existing groundwater and soil models for
better assessment and quantification of As risk.
10. It is also important to include socio-economic
datasets in conjunction with chemical and
spatial datasets in the process of human health
risk assessment because nutritional condition
and food habits may have some influence on As
exposure.
11. Many studies have indicated that As speciation
in foods and rice is an important issue. There-
fore, future studies should be directed towards
speciation of As in foodstuffs with particular
attention to rice when total As concentrations
are elevated.
12. In the case of human health risk assessment
from As contamination for Asian countries,
especially for Bangladesh, there is need to
assess the contribution of inorganic As from
cooked rice and cooked dal (pulses, i.e., lentil,
split peas, etc.) in order to calculate accurate
estimates of TDI. This is important because rice
and dal are cooked with a large amount of
water. Cooked rice absorbs significant amount
of water after cooking and both cooked dal and
rice retain large amounts of water.
13. Arsenic content of vegetable and crops does not
generally exceed the crop As guideline; rather
the human health risk is from the amount of
vegetables and rice ingested, which leads to
exceedance of allowable total daily intake.
Therefore it would be wise to calculate TDI
based on different countries’ dietary pattern to
assess the human health risk caused by As
contamination.
References
Abernathy, C. O., Liu, Y.-P., Longfellow, D., Aposhian, H. V.,
Beck, B., Fowler, B., et al. (1999). Arsenic: Health
effects, mechanisms of actions, and research issues.
Environmental Health Perspectives, 107, 593–597.
Abernathy, C. O., Thomas, D. J., & Calderon, R. L. (2003).
Health effects and risk assessment of arsenic. American
Society for Nutritional Sciences, 133, 1536S–1538S.
Adair, B. M., Hudgens, E. E., Schmitt, M. T., Calderon, R. L.,
& Thomas, D. J. (2006). Total arsenic concentrations in
toenails quantified by two techniques provide a useful
biomarker of chronic arsenic exposure in drinking water.
Environmental Research, 101(2), 213–220.
Ahmad, S. A. (2007). Sociocultural aspects of arsenicosis in
Bangladesh: Community perspective. Journal of Envi-
ronmental Science and Health Part A, 42(12), 1945–1958.
Al-Adamat, R. A. N., Foster, I. D. L., & Baban, S. M. J. (2003).
Groundwater vulnerability and risk mapping for Basaltic
aquifer of the Azraq basin of Jordan using GIS, Remote
sensing and DRASTIC. Applied Geography, 23, 303–324.
Alam, M. G. M., Snow, E. T., & Tanaka, A. (2003). Arsenic
and heavy metal contamination of vegetables grown in
Samta village, Bangladesh. Science of the Total Envi-
ronment, 308, 83–96.
Albertson, P. E., Girstmair, G., & Otero Valle, M. N., (1993).
Development and application of a groundwater model
integrated in the GIS GRASS. IAHS Publication, 211,
333 pp.
Ali, S. M. K., Edib, K., Pramanik, M. M. A., Alam, A. M. S.,
Rabbani, G. H., Anwar, K. S., Hossain, A., & Nasir, M.
(2002). Nutritional status of patients with chronic arseni-
cosis in rural Bangladesh. Second International conference
on Bangladesh Enviorment (ICBEN-2002), pp. 167–172.
Al Rmalli, S. W., Haris, P. I., Harrington, C. F., & Ayub, M.
(2005). A survey of arsenic in foodstuffs on sale in the
United Kingdom and imported from Bangladesh. Science
of the Total Environment, 337(1–3), 23–30.
Anawar, H. M., Akai, J., Mostofa, K. M. G., Safiullah, S., &
Tareq, S. M. (2002). Arsenic poisoning in groundwater:
Health risk and geochemical sources in Bangladesh.
Environment International, 27(7), 597–604.
Aral, M. M., & Maslia, M. L. (1996). Evaluation of human
exposure to contaminated water supplies using GIS and
modeling. In K. Kover & H. P. Nachtnebel (Eds.), Appli-
cation of geographic information systems in hydrology and
water resources management (pp. 243–252), IAHS
Publication No. 235. Wallingford: IAHS Press.
162 Environ Geochem Health (2009) 31:143–166
123
Aslibekian, O., & Moles, R. (2003). Environmental risk
assessment of metals contaminated soils at silvermines
abandoned mine site, CO Tipperary, Ireland. Environ-
mental Geochemistry and Health, 25(2), 247–266.
ATSDR. (1989). Toxicological profile of arsenic. In Agency for
toxic substances and disease registry (ASTDR). Atlanta:
US Public Health Services.
BADC. (2003). Survey Report on Irrigation Equipment and
Irrigated Area in Boro 2003 Season, Bangladesh Agri-
cultural Development Corporation (BADC).
Bae, M., Wantanabe, C., Inaoka, T., Sekiyama, M., Sudo, N.,
Bokul, M. H., et al. (2002). Arsenic in cooked rice in
Bangladesh. The LANCET, 360(9348), 1839–1840.
Barrocu, G., & Biallo, G. (1993). Application of GIS for
aquifer vulnerability evaluation. In K. Kover & H. P.
Nachtnebel (Eds.), Application of geographic information
systems in hydrology, water resources management
(pp. 571–578). Wallingford: IAHS.
Bartels, C. J., & Van Beurden, A. U. C. J. (1998). Using
geographic and cartographic principles for environmental
assessment and risk mapping. Journal of Hazardous
Materials, 61, 115–124.
Basu, B., & Sil, S. (2004). Arsenic mapping for North 24—
Pargana District of West Bengal using GIS and Remote
Sensing technology. http://www.gisdevelopment.net/appli
cation/environment/water/mi03204pf.htm.
Bien, J. D., ter Meer, J., Rulkens, W. H., & Rijnaarts, H. H. M.
(2004). A GIS based approach for the long-term predic-
tion of human health risks at contaminated sites.
Environmental Modelling and Assessment, 9, 221–226.
Bober, M. L., Wood, D., & McBridge, R. A. (1996). Use of
digital analysis and GIS to assess regional soil compaction
risk. PE&RS: Photogrammetric Engineering & Remote
Sensing, 62, 1397.
Buchet, J. P., Lauwerys, R., Vandevoorde, A., & Pycke, J. M.
(1983). Oral daily intake of cadmium, lead, manganese,
copper, chromium, mercury, calcium, zinc and arsenic in
Belgium: a duplicate meal study. Food and Chemical
Toxicology, 21(1), 19–24.
Caceres, D. D., Pino, P., Montesinos, N., Atalah, E., Amigo,
H., & Loomis, D. (2005). Exposure to inorganic arsenic in
drinking water and total urinary arsenic concentration in
Chilean Population. Environmental Research, 98(2), 151–
159.
Caussy, D., Gochfeld, M., Gurzau, E., Neagu, C., & Ruedel, H.
(2003). Lessons from case studies of metals: Investigating
exposure, bioavailability, and risk. Ecotoxicology and
Environmental Safety, 56, 45–51.
Chen, C.-J., Hsu, L.-I., Wang, C.-H., Shih, W.-L., Hsu, Y.-H.,
Tseng, M.-P., et al. (2005). Biomakers of exposure, effect,
and susceptibility of arsenic-induced health hazards in
Taiwan. Toxicology and Applied Pharmacology, 206(2),
198–206.
Chen, Z., Huang, G. H., Chakma, A. (1998). Integrated envi-
ronmental risk assessment through a GIS based decision-
support system. Paper presented at the ESRI International
User Conference, San Diego, USA.
Christensen, J. M. (1995). Human exposure to toxic metal:
Factors influencing interpretation of biomonitoring results.
The Science of the Total Environment, 166, 89–135.
CIA. (2006). World Fact Book,www.cia.gov/cia/publications/
factbook.
Dabeka, R. W., McKenzie, A. D., Lacroix, G. M., Cleroux, C.,
Bowe, S., Graham, R. A., et al. (1993). Survey of arsenic
in total diet food composites and estimation of the dietary
intake of arsenic by Canadian adults and children. Journal
of AOAC International, 76(1), 14–25.
Dabeka, R. W., McKenzie, A. D., & Lacroix, G. M. (1987).
Dietary intakes of lead, cadmium, arsenic and fluoride by
Canadian adults: A 24-hour duplicate diet study. Food
Additives and Contaminants, 4(1), 89–101.
Das, P. K., Mitra, A. K., Sengupta, P. K., Hossain, A., Islam,
F., & Rabbani, G. H. (2004). Arsenic concentration in
rice, vegetables, and fish in Bangladesh: A preliminary
study. Environment International, 30(3), 383–387.
Del Razo, L. M., Garcia-Vargas, G. G., Garcia-Salcedo, J.,
Sanmiguel, M. F., Rivera, M., Hernandez, M. C., et al.
(2002). Arsenic levels in cooked food and assessment of
adult dietray intake of arsenic in the Region Lagunera,
Mexico. Food and Chemical Toxicology, 40(10), 1423–
1431.
de Paz, J. M., & Ramos, C. (2002). Linkage of a geographical
information system with the gleams model to assess
nitrate leaching in agricultural areas. Environmental Pol-
lution, 118, 249–258.
Dokkum, W. V., De Vos, R. H., Muys, T. H., & Wesstra, J. A.
(2007). Minerals and trace elements in total diets in the
Netherlands. British Journal of Nutrition, 61(1), 7–15.
Duker, A. A. (2005). Spatial analysis of factors implicated in
mycobacterium ulcerans infection in Ghana (pp. 1–155).
Unpublished Ph.D. Thesis Enschede: ITC.
Duxbury, J. M., Mayer, A. B., Lauren, J. G., & Hassan, N.
(2003). Food chain aspects of arsenic contamination in
Bangladesh: Effects on quality and productivity of rice.
Journal of Environmental Science and Health, A38(1),
61–69.
Emmi, P. C., & Horton, C. A. (1996). Seismic risk assessment,
accuracy requirements, and GIS based sensitivity analysis.
In M. F. Goodchild, L. T. Steyaert, B. O. Parks,
C. Johnston, D. Maidment, D. Crane, & S. Glendinning
(Eds.), GIS and environmental modeling: Progress and
research issues (pp. 191–195). Fort Collins: GIS World
Books.
Engel, B., Navulur, K., Cooper, B., & Hahn, L. (1996). Esti-
mating groundwater vulnerability to nonpoint source
pollution from nitrates and pesticides on a regional scale.
In K. Kover & H. P. Nachtnebel (Eds.), Application of
geographic information systems in hydrology, water
resources management (pp. 521–526). Wallingford:
IAHS.
Fedra, K. (1993). GIS and environmental modeling. In M. F.
Goodchild, B. O. Parks, & L. T. Steyaert (Eds.), Envi-
ronmental modeling with GIS (pp. 35–50). New York:
Oxford University Press.
Fedra, K. (1996). Distributed models and embedded GIS:
Integration strategies and case studies. In M. F. Good-
child, L. T. Steyaert, B. O. Parks, C. Jhonston,
D. Maidment, M. Crane, & S. Glendinning (Eds.), GIS
and environmental modelling: Progress and research
issues (pp. 413–418). Fort Collins: GIS World.
Environ Geochem Health (2009) 31:143–166 163
123
Forman, R. T. T., & Gordon, M. (1986). Landscape ecology.
New York: Wiley.
Foster, J. A., & McDonald, A. T. (2000). Assessing pollution
risks to water supply intakes using geographical infor-
mation systems (GIS). Environmental Modelling &
Software, 15(3), 225–234.
Frysinger, S. P. (1996). Environmental decision support sys-
tems (EDSS): An open architecture integrating modeling
and GIS. In M. F. Goodchild, L. T. Steyaert, B. O. Parks,
C. Jhonston, D. Maidment, M. Crane, & S. Glendinning
(Eds.), GIS and environmental modelling: Progress and
research issues (pp. 357–362). Fort Collins: GIS World.
Fytianos, K., & Christophoridis, C. (2004). Nitrate, arsenic and
chloride pollution of drinking water in northern Greece,
elaboration by applying GIS. Environmental Monitoring
and Assessment, 93(1), 55–67.
Gartrell, M. J., Craun, J. C., Podrebarac, D. S., & Gunderson,
E. L. (1985). Pesticides, selected elements, and other
chemicals in infant and toddler total diet samples, October
1978–September 1979. Journal of Association of Official
Analytical Chemists, 66(5), 842–861.
Gay, J. R., & Korre, A. (2006). A spatially-evaluated method-
ology for assessing risk to a population from contaminated
land. Environmental Pollution, 142, 227–234.
Gochfeld, M. (1998). Principles of toxicology. In R. Wallace
(Ed.), Maxcy-Rosenau-Last. Public health & preventive
medicine (14th ed.). Stamford: Appleton-Lange.
Goodchild, M. F., Parks, B. O., & Steyaert, L. T. (Eds.).
(1993). Environmental modeling with GIS. New York:
Oxford University Press.
Goovaerts, P., AvRuskin, G., Meliker, J., Slotnick, M., Jacquez,
G., & Nriagu, J. (2005). Geostatistical modeling of the
spatial variability of arsenic in groundwater of southeast
Michigan. Water Resources Research, 41(7), 1–19.
Grandjean, P., & Bach, E. (1986). Indirect exposures: The
significance of bystanders at work and at home. AIHA
Journal, 47(12), 819–824.
Gunderson, E. L. (1995). FDA Total Diet Study, July 1986–
April 1991, dietary intakes of pesticides, selected ele-
ments, and other chemicals. Journal of AOAC
International, 78(6), 1353–1363.
Guo, H.-R., & Tseng, M.-P. (2000). Arsenic drinking water and
bladder cancer: Comparison between studies based on
cabcer registry and death certificates. Environmental
Geochemistry and Health, 22, 83–91.
Hadi, A., & Parveen, R. (2004). Arsenicosis in Bangladesh:
Prevalence and socio-economic correlates. Public Health,
118(8), 559–564.
Hargrove, W. W., Levine, D. A., Miller, M. R., Coleman, P. R.,
Pack, D. L., & Durfee, R. C. (2001). GIS and risk
assessment: A fruitful combination. In ESRI international
user conference. San Diego, USA.
Harris, J. S. (1997). Evaluating possible human exposure
pathways to populations relative to hazardous materials
sites. In ESRI international user conference. San Diego,
USA.
Harrison, P. T. C., & Holmes, P. (2006). Assessing risks to
human health. Issues in Environmental Science and
Technology, 22(22), 65–83.
Hasnat, M. A. (2005). Assessment of arsenic mitigation
options; adverse pregnancy outcomes due to chronic
arsenic exposure; and the impact of nutritional status on
development of arsenicosis in Bangladesh. PhD. Thesis,
The National Centre from Epidemiology and Population
Health. The Australian National University, Australia.
Hassan, M. M., Atkins, P. J., & Dunn, C. E. (2003). The spatial
pattern of risk from arsenic poisoning: A Bangladesh case
study. Journal of Environmental Science and Health,
A38(1), 1–24.
Hassan, M. M., Atkins, P. J., & Dunn, C. E. (2005). Social
implications of arsenic poisoning in Bangladesh. Social
Science and Medicine, 61, 2201–2211.
Heck, J. E., Gamble, M. V., Chen, Y., Graziano, J. H., Slav-
kovich, V., Parvez, F., et al. (2007). Consumption of
folate-related nutrients and metabolism of arsenic in
Bangladesh. American Journal of Clinical Nutrition,
85(5), 1367–1374.
Hellweger, F. L., Wilson, H. L., Naranjo, E. M., & Anid, P. J.
(2002). Adding human health risk analysis tools to geo-
graphic information systems. Transactions in GIS, 6(4),
471–484.
Hiscock, K. M., Lovett, A. A., & Parfitt, J. P. (1995).
Groundwater vulnerability assessment: two case studies
using GIS methodology. The Quarterly Journal of Engi-
neering Geology, 28, 179–194.
Hsu, L. T., & Cheng, C. C. (1999). Integrating landscape
models in forest landscape analyses using GIS: An
example in Taiwan. GIS Development, 1–5.
Islam, L. N., Nabi, A., Rahman, M. M., Khan, M. A., & Kazi,
A. I. (2004a). Association of clinical complications with
nutritional status and the prevalence of leukopenia among
arsenic patients in Bangladesh. International Journal of
Environmental Research and Public Health,1(2), 74–82.
Islam, M. N., & Nehaluddin, M. (2002). Hydrogeology and
arsenic contamination in Bangladesh. In M. A. Feroze &
M. A. Chowdhury (Eds.), Arsenic mitigation in Bangla-
desh. Government of the People’s Republic of
Bangladesh: Local Government Division, Ministry of
Rural Development & Cooperatives.
Islam, M. R., Paul, D. K., & Shaha, R. K. (2004b). Nutritional
importance of some leafy vegetables available in Ban-
gladesh. Pakistan Journal of Biological Sciences, 7(8),
1380–1384.
Kelly, C., & Lunn, R. J. (1999). Development of a contami-
nated land assessment system based on hazard to surface
water bodies. Water Resources Research, 33(6), 1377–
1386.
Korre, A., Durucan, S., & Koutroumani, A. (2002). Quantita-
tive-spatial assessment of the risks associated with high pb
loads in soils around Lavrio, Greece. Applied Geochem-
istry, 17, 1029–1045.
Korte, N. E., & Fernando, Q. (1991). A review of Arsenic (III)
in groundwater. CRC Critical Reviews in Environmental
Control CCECAU, 21(1), 1–39.
Laparra, J. M., Velez, D., Barbera, R., Farre, R., & Montoro,
R. (2005). Bioavailability of inorganic Arsenic in cooked
rice: Practical aspects for human risk assessment.
Journal of Agricultural and Food Chemistry, 53(22),
8829–8833.
Lin, C. W. (2002). Mapping soil lead and remediation needs in
contaminated soils. Enviornmental Geochemistry and
Health, 24(1), 23–33.
164 Environ Geochem Health (2009) 31:143–166
123
Liu, C. W., Liang, C. P., Huang, F. M., & Hsueh, Y. M. (2006).
Assessing the human health risks from exposure of inor-
ganic arsenic through oyster (Crassostrea gigas)
consumption in Taiwan. Science of the Total Environ-
ment, 361(1–3), 57–66.
Llobet, J. M., Falco, G., Casas, C., Teixido, A., & Domingo, J. L.
(2003). Concentration of Arsenic, Cadmium, Mercury, and
Lead in common foods and estimated daily intake by chil-
dren, adolescents, adults and seniors of Catalonia, Spain.
Journal of Agricultural and Food Chemistry, 51(3), 838–
842.
Lovertt, A. A., Parfitt, J. P., & Brainard, J. S. (1997). Using
GIS in risk analysis: A case study of hazardous waste
transport. Risk Analysis, 17(5), 625–633.
Luzi, L. (1995). Application of available modeling to zoning of
landslide hazard in Fabriano Area, Central Italy. In JEC-
GI ‘95 (pp. 398–403). The Hague.
Maidment, R. D. (1996). Environmental modelling within GIS.
In M. F. Goodchild, L. T. Steyaert, B. O. Parks,
C. Jhonston, D. Maidment, M. Crane, & S. Glendinning
(Eds.), GIS and environmental modelling: Progress and
research issues (pp. 315–324). New York: Oxford Uni-
versity Press.
Mauro, C. D., Hoogerwerf, M., & Sinke, A. J. C. (2000). A GIS
Based risk assessment model: Application on sites con-
taminated by chlorinated solvents and petroleum
hydrocarbons (BTEX). In Fourth International Confer-
ence on Integrating GIS and Environmental Modelling
(GIS/EM4): Problems, Prospects and Research Needs.
Alberta, Canada.
Mazumder, G. D. N., Haque, R., Ghosh, N., De, B. K., Santra,
A., Chakroborti, D., et al. (1998). Arsenic levels in
drinking water and the prevalence of skin lesions in West
Bengal, India. International Journal of Epidemiology,
27(5), 871–877.
McGrath, D., Zangh, C. S., & Carton, O. (2004). Geostatistical
analyses and hazard assessment on soil lead in silvermines
area, Ireland. Environmental Pollution, 127(2), 239–248.
Meharg, A. A. (2006). Arsenic contamination of Bangladesh
rice. In R. Naidu, E. Smith, G. Owens, P. Bhattacharya, &
P. Nadebaum (Eds.), Managing Arsenic in the environ-
ment: From soil to human health (pp. 273–282).
Australia: CSIRO.
Meharg, A. A., & Rahman, M. M. (2003). Arsenic contami-
nation of Bangladesh paddy field soils: Implications for
rice contribution to arsenic consumption. Environmental
Science Technology, 37(2), 229–234.
Ministry of Agriculture, Fisheries and Food (MAFF). (1999).
Total Diet Study: Aluminum, arsenic, cadmium, chro-
mium, cropper, lead, mercury, nickel, selenium, tin and
zinc. Food surveillance information sheet, No. 191. Lon-
don: HMSO.
Mitra, S. R., Mazumder, D. N. G., Basu, A., Block, G., Haque, R.,
Samanta, S., et al. (2004). Nutritional factors and suscep-
tibility to arsenic—caused skin lesions in West Bengal,
India. Environmental Health Perspectives, 112(10), 1104–
1109.
Mohri, T., Hisanaga, A., & Ishinishi, N. (1990). Arsenic intake
and excretion by Japanese adults: A 7-day duplicate diet
study. Food Chemical Toxicity, 28(7), 521–529.
Morra, P., Bagli, S., & Spadoni, G. (2006). The analysis of
human healthy risk with a detailed procedure operating in
a GIS environment. Environment International, 32(4),
444–454.
Naidu, R., & Skinner, H. (1999). Arsenic contamination of
rural ground water supplies in Bangladesh and India:
Implications for soil quality, animal and human health. In
C. Barber, B. Humphries, & J. Dixon (Eds.), International
conference on diffuse pollution (pp. 407–417). Perth,
Australia: CSIRO.
Ortigosa, G. R., Leo, G. A., & Gatto, M. (2000). VVF: Inte-
grating modelling and GIS in a software tool for habitat
suitability assessment. Environment Modelling Software,
15(1), 1–12.
Owens G., Rahman M. M., Heinrich T., & Naidu R. (2004).
Bangladesh–Australia Centre for Arsenic Mitigation
Program (BACAMP): Program 3: Safe Food, Sect. 1:
Arsenic Food Chain Assessment. University of South
Australia, Consultancy report for GHD Pty Ltd.
Parrish, D. A., Townsend, L., Saunders, J., Carney, G., &
Langston, C. (1993). US EPA region 6 comparative risk
project: Evaluating ecological risk. In M. F. Goodchild,
B. O. Parks, & L. T. Steyaert (Eds.), Environmental
modeling with GIS (pp. 318–331). New York: Oxford
University Press.
Patel, S. K., Shrivas, K., Brandt, R., Jakubowski, N., Corns,
W., & Hoffmann, P. (2005). Arsenic contamination in
water, soil, sediment and rice of central India. Environ-
mental Geochemistry and Health, 27(2), 131–145.
Pomroy, C., Charbonneau, S. M., McCullough, R. S., & Tam,
G. K. (1980). Human retention studies with 74As. Toxi-
cology and Applied Pharmacology, 53(3), 550–556.
Reis, A. P., Sousa, A. J., Da Silva, F., & Fonseca, C. (2005).
Application of geostatistical methods to arsenic data from
soil samples of the Cova dos Mouros mine (Vila Verde-
Portugal). Enviornmental Geochemistry and Health,
27(3), 259–270.
Rejeski, D. (1993). GIS and risk: A three-cultural problem. In
M. F. Goodchild, B. O. Parks, & L. T. Steyaert (Eds.),
Environmental modeling with GIS (pp. 318–331). New
York: Oxford University Press.
Roberts, S. M., Weimar, W. R., Vinson, J. R. T., Munson, J. W.,
& Bergeron, R. J. (2002). Measurement of Arsenic bio-
availability in soil using primate model. Toxicological
Sciences, 67, 303–310.
Robson, M. (2003). Methodologies for assessing exposures to
metals: Human host factor. Ecotoxicology and Environ-
mental Safety, 56(1), 104–109.
Romic, M., & Romic, D. (2003). Heavy metals distribution in
agricultural topsoils in urban area. Environmental Geol-
ogy, 43(7), 795–805.
Roychowdhury, T., Tokunaga, H., & Ando, M. (2003). Survey
of arsenic and other heavy metals in food composites and
drinking water and estimation of dietary intake by the
villagers from an arsenic-affected area of West Bengal,
India. The Science of the Total Environment, 308, 15–35.
Roychowdhury, T., Uchino, T., Tokunaga, H., & Ando, M.
(2002). Survey of arsenic in food composites from as
aresenic-affected area of West Bengal, India. Food
Chemical Toxicology, 40, 1611–1621.
Environ Geochem Health (2009) 31:143–166 165
123
Ryan, P. B., Huet, N., & MacIntosh, D. L. (2000). Longitudinal
investigation of exposure to arsenic, cadmium, and lead in
drinking water. Enviornmental Health Perspectives,
108(8), 731–735.
Sabel, C. E., Gatrell, A. C., Loytonen, M., Maasilta, P., &
Jokelainen, M. (2000). Modelling exposure opportunities:
estimating relative risk for motor neuron disease in Fin-
land. Social Science Medicine, 50(7–8), 1121–1137.
Sapunar-Postruznik, J., Bazulic, D., & Kubala, H. (1996).
Estimation of dietary intake of arsenic in the general
population of Republic of Croatia. The Science of the
Total Environment, 191, 119–123.
Schoof, R. A., Yost, L. J., Eickhoof, J., Crecelius, E. A.,
Cragin, D. W., Meacher, D. M., et al. (1999). A market
basket survey of inorganic arsenic in food. Food and
Chemical Toxicology, 37(8), 839–846.
Sekhar, K. C., Chary, N. S., Kamla, C. T., Rao, J. V., Balaram, V.,
& Anjaneyulu, Y. (2003). Risk assessment and pathway
study of arsenic in industrially contaminated sites of Hy-
derabad: a case study. Environment International, 29(5),
601–611.
Serre, M. L., Kolovos, A., Christakos, G., & Modis, K. (2003).
An application of the holistochastic human exposure
methodology to naturally occurring arsenic in Bangladesh
Drinking Water. Risk Analysis, 23(3), 515–528.
Smith, E., Naidu, R., & Alston, A. M. (1998). Arsenic in the
soil environment: A review. Advances in Agronomy, 64,
149–195.
Stein, A., Staritsky, I., & van Groenigen, J. W. (1995). Inter-
active GIS for environmental risk assessment.
International Journal Geographic Information System,
9(5), 509–525.
Sydelko, P. J., Dolph, J. E., Majerus, K. A. & Taxon, T. N.
(2000). An advance object-based software framework for
complex ecosystem modeling and simulation. In Fourth
International Conference on Integrating GIS and Envi-
ronmental Modeling (GIS/EM4): Problems, Prospects
and Research Needs (pp. 1–9). Alberta, Canada.
Tchounwou, P. B., Centeno, J. A., & Patlolla, A. K. (2004).
Arsenic toxicity, mutagenesis, and carcinogenesis—A
health risk assessment and management approach.
Molecular and Cellular Biochemistry, 255, 47–55.
Tseng, W. P. (1977). Effects and dose-response relationships of
skin cancer and Blackfoot diseases with arsenic. Envi-
ronmental Health Perspectives, 19, 109–119.
Tsuda, T., Inoue, T., Kojima, M., & Aoki, S. (1995). Market
basket and duplicate portion estimation of dietary intakes
of cadmium, mercury, arsenic, copper, manganese, and
zinc by Japanese adults. Journal of AOAC International,
78(6), 1363–1368.
Uchino, T., Roychowdhury, T., Ando, M., & Tokunaga, H.
(2006). Intake of arsenic from water, food composites and
excretion through urine, hair from a studied population in
West Bengal, India. Food and Chemical Toxicology, 44,
455–461.
Wadge, G., Wislocki, A. P., & Pearson, E. J. (1993). Spatial
analysis in GIS for natural hazards assessment. In M. F.
Goodchild, B. O. Parks, & L. T. Steyaert (Eds.), Envi-
ronmental modeling with GIS (pp. 332–338). USA:
Oxford University Press.
Watanabe, C., Kawata, A., Sudo, N., Sekiyama, M., Inaoka, T.,
Bae, M., et al. (2004). Water intake in an Asian population
living in arsenic contaminated area. Toxicology and
Applied Pharmacology, 198, 272–282.
WHO. (2001). Arsenic and arsenic compounds (2nd ed.,
Environmental health criteria, 224). International Pro-
gramme on Chemical Safety, 521 S. Geneva: World
Health Organization.
Williams, P. N., Price, A. H., Raab, A., Hossain, S. A., Feld-
mann, J., & Meharg, A. A. (2005). Variation in arsenic
speciation and concentration in paddy rice related to
dietary exposure. Enviornmental Science and Technology,
39, 5531–5540.
Wilson, H. L. (1997). Simple environmental exposure models
in a GIS Framework. In Term project report. University of
Texas, Austin.
Yost, L. J., Tao, S.-H., Egan, S. K., Barraj, L. M., Smith, K. M.,
Tsuji, J. S., et al. (2004). Estimation of dietary intake of
inorganic arsenic in U.S. children. Human and Ecological
Risk Assessment, 10(3), 473–483.
Yu, W. H., Harvey, C.m., & Harvey, C. F. (2003). Arsenic in
groundwater in Bangladesh: A geostatistical and epide-
miological framework for evaluating health effects and
potential remedies. Water Resources Research, 39(6),
1–17.
Zandbergen, P. A. (1998). Urban watershed ecological risk
assessment using GIS: A case study of the Brunette River
watershed in British Columbia, Canada. Journal of Haz-
ardous Materials, 61(1), 163–173.
Zhang, J., Chen, X., Parkpian, P., & Sriratana, M. (2001). GIS
application on arsenic contamination and its risk assess-
ment in Ronphibun, Nakhorn Si Thammarat, Thailand.
Geographical Information Sciences, 7(2), 69–78.
Zhang, J., Hodgson, J., & Erkut, E. (2000). Using GIS to assessthe
risks of hazardous materials transport in networks. European
Journal of Operational Research, 121(2), 316–329.
166 Environ Geochem Health (2009) 31:143–166
123
