Environ Sci Pollut Res (2016) 23:1234–1245
DOI 10.1007/s11356-015-5335-4
RESEARCH ARTICLE
A comparison of risk modeling tools and a case study for human
health risk assessment of volatile organic compounds
in contaminated groundwater
Lu Han 1 & Linbo Qian 1 & Jingchun Yan 1 & Rongqin Liu 1,2 & Yihua Du 3 & Mengfang Chen 1
Received: 5 June 2015 / Accepted: 27 August 2015 / Published online: 10 September 2015
# Springer-Verlag Berlin Heidelberg 2015
Abstract In order to promote the risk-based strategy in the
investigation, assessment, and remediation of Chinese brown-
field sites, the Health and Environmental Risk Assessment
(HERA) software was developed. It is vital to validate the
HERA model and compare the inter-model differences of HE-
RA model against other available risk assessment tools. This
paper discusses the similarities and differences between the
Risk-Based Corrective Action (RBCA) Tool Kit and the HE-
RA model by evaluating the health risk of organic contami-
nated groundwater sources for a chemical works in China for
the first time. Consequently, the HERA and RBCA models
yielded the identical results for Site-Specific Assessment
Criteria (SSAC) under the commercial redevelopment. How-
ever, the HERA estimated more conservative and stringent
SSACs under the residential scenario based on the different
exposure calculations. The inhalation of indoor vapors was the
most predominated exposure pathway for all the volatile or-
ganic compounds (VOCs) determined using the RBCA and
HERA models. According to the HERA model, inhalation of
Responsible editor: Marcus Schulz
Electronic supplementary material The online version of this article
(doi:10.1007/s11356-015-5335-4) contains supplementary material,
which is available to authorized users.
* Mengfang Chen
mfchen@issas.ac.cn
1
Key Laboratory of Soil Environment and Pollution Remediation,
Institute of Soil Science, Chinese Academy of Sciences,
Nanjing 210008, China
2
School of Environmental Studies, China University of Geosciences,
Wuhan 430074, China
3
Computer Network Information Center, Chinese Academy of
Sciences, Beijing 100190, China
chloroform may cause the highest unacceptable carcinogenic
risk at 2.31×10−3 under the residential scenario. Therefore, it
is recommended that a risk-based remedial strategy be
developed to ensure the safe and sustainable redevelop-
ment of the site.
Keywords Health risk assessment . Groundwater . Volatile
organic compounds . Principal component analysis . Model
validation
Introduction
Volatile organic compounds (VOCs) have been used ex-
tensively as raw materials or industrial degreaser in pet-
rochemical manufacturing and the synthesis of pesticide,
resins, and plastics (Fan et al. 2009; Ma et al. 2011).
However, they are ubiquitously present at many indus-
trial and disposal sites due to the improper storage, ap-
plication, and disposal during manufacturing processes.
They are considered to be the primary contaminants in
groundwater due to the better transportation ability and
water solubility (Kitwattanavong et al. 2013; Lopez
et al. 2008; Lapworth et al. 2012). The VOCs may po-
tentially destroy the groundwater environment and pose
a threat to the safety of public water supplies. In addi-
tion, these organic pollutants can migrate upward into
the atmosphere, causing public health threats from
chronic inhalation pathway (Cachada et al. 2012; John-
ston and Gibson 2013).
To restore groundwater to generic standards such as
drinking water standards is challenging in time and ef-
fort. To minimize expenditure and avoid excessive re-
mediation while protecting human health and the envi-
ronment, a risk-based and sustainable environmental
Environ Sci Pollut Res (2016) 23:1234–1245
management approach recommended by the United
States Environmental Protection Agency (USEPA) has
been widely accepted in many developed countries
(Baciocchi et al. 2010; Bardos et al. 2008; Cachada
et al. 2012; Nieuwenhuijsen et al. 2006; USEPA 1996,
2012; Faerber and Herzog 2010). Risk assessment deter-
mines the need for environmental remediation, quantify-
ing the probability and the deleterious implications of
contaminants to human health and the environment as
defined by the International Organization for Standardi-
zation (ISO 2012). Currently, many risk assessment
models and tools have been freely or commercially ap-
plied in developed countries, such as the Risk-Based
Corrective Action (RBCA) Tool Kit developed by GSI
Environmental Inc. (GSI 2014) and Contaminated Land
Exposure Assessment (CLEA) model (Jeffries and Mar-
tin 2009). The quantitative environmental models effi-
ciently support the risk-based assessment approach and
assist in the subsequent environmental decision-making
process (Carlon et al. 2008). Additionally, a comparison
of tools for contaminated sites and brownfield assess-
ment has been already performed (Pinedo et al. 2014;
Stezar et al. 2013).
The risk-based management of brownfield sites is
still in its infancy and is receiving increasing attention
in China (Chen 2010; Luo et al. 2013; Ren et al. 2015;
Wei et al. 2015; Zhou et al. 2012). In July 2014, the
Ministry of Environmental Protection of China pub-
lished the technical guideline for risk assessment of
contaminated sites (CRAG 2014). In order to develop
a predictive model of risk assessment supporting the
implementation of CRAG guideline, the Health and En-
vironmental Risk Assessment (HERA) model was de-
veloped. The HERA model has fully complied with
the CRAG guideline (2014) and incorporated the tech-
nical algorithms in extensive validation against RBCA
Tool Kit and CLEA model. Under the CRAG frame-
work, HERA can develop site-specific environmental
cleanup criteria from multimedia exposure for human
health and/or environmental receptors through a back-
ward calculation based on the acceptable target hazard
quotient (THQ) and/or target cancer risk (TCR) and
evaluate the potential cancer risks and/or hazard quo-
tients to humans through a forward calculation on the
basis of the actual concentration in the media, using a
tiered risk evaluation approach (CRAG 2014). The ex-
posure pathways mainly consist of oral ingestion and
dermal contact with soil, plant uptake from soil, inges-
tion of contaminated groundwater, inhalation of vapors
and particulates from the soil source, and inhalation of
vapors from the groundwater. The validation and the
inter-model differences of HERA model against other
established models are still not well known.
1235
Therefore, the purpose of this study is to quantify the
potential risks from organic contaminated groundwater
sources for a chemical works in China using the HERA
model (version 1.01), interpret the technical algorithms
recommended by the CRAG framework, and compare
the inter-model differences of HERA model with RBCA
Tool Kit. The specific studies undertaken include (1)
using statistical methods to analyze the sampling data
and to enhance data interpretation; (2) applying HERA
model and RBCA Tool Kit to derive the Site-Specific
Assessment Criteria (SSAC) through individual expo-
sure pathway, evaluating the risk modeling outcomes
and discussing the inter-model differences; and (3) un-
dertaking a detailed quantitative risk assessment for the
VOCs contaminated groundwater using HERA model. It
is envisaged that conclusions of this study will assist in
making informed decisions in sustainable management
of remedial activities at the site.
Methodology
Site description
The former chemical works is located in Jiangsu province in
China, which has been operating for more than 30 years. The
manufacturing activities were ceased following the national
policy of shutting down and relocating contaminated indus-
tries. The main products consisted of methamidophos pesti-
cide, chloral-alkali, ethers, acetals, and chlorobenzenes.
Monochlorobenzene and carbon tetrachloride were used as
raw materials in manufacturing 3-methyl-diphenyl ether and
3-phenoxy-benzaldehyde. The site investigation has revealed
a leakage around the wastewater treatment plant. Additional
spillage from the pipelines containing the raw materials was
also identified. This is as a result of improper excavation in the
historical reconstruction of the plants.
As shown in Fig. 1, the site with an area of 85,476 m2 is
surrounded by roads on western, eastern, and southern bound-
aries, and a river to the north. A village is situated approxi-
mately 85 m to the south west of the site. The geological
sequence of the site is characterized by backfill, silt clay, silt,
and clay (Fig. 2a) with the line of the geological section (1-1′)
being shown in Fig. 1. The silt clay layer acts as an aquitard
and some discontinuous perched water bodies exist between
backfill and silt clay layers. The silt layer, located approxi-
mately between 9.5 and 17 m below ground level, is relatively
permeable compared with the silt clay and clay layer encoun-
tered at the site, as the hydraulic conductivity was measured in
the range of 2.53×10−4 to 3.5×10−4 cm/s. The static ground-
water table is recorded in the geological section (Fig. 2a). The
groundwater contour is illustrated in Fig. 2b, suggesting
groundwater predominantly flowing from south west to north
1236
Fig. 1 Site layout showing groundwater sampling locations
east. Physical parameters were determined including the frac-
tions of organic carbon, porosity, moisture content, and bulk
density in unsaturated and saturated zones.
Sample collection and analysis
The sampling location map is presented in Fig. 1. Six-
teen groundwater monitoring wells were installed in the
study area and the screening depths of the wells ranged
from 9 to 15 m below the ground level. The sample
collection from groundwater each monitoring well was
analyzed for VOCs and semi-volatile organic com-
pounds (SVOCs) within 24 h according to the protocols
of the Ministry of Environmental Protection of China
(MEPPRC 2009). Ten VOCs including benzene, tolu-
ene, ethyl benzene, monochlorobenzene, 1,2-dichloro-
benzene, 1,3-dichlorobenzene, 1,4-dichlorobenzene, car-
bon tetrachloride, chloroform, and dichloromethane
were initially selected for this study based on their de-
tection during the monitoring process. Correlation anal-
ysis and multivariate principal component analysis
(PCA) were performed on the sampling data in order
to identify likely groupings between the contaminants.
In addition, the Bfactor score^ for each sampling
Environ Sci Pollut Res (2016) 23:1234–1245
location was calculated in PCA and was plotted to clus-
ter sampling location with similar contaminants (Critto
et al. 2003).
Conceptual site model
A conceptual site model will identify relations of con-
taminant sources–exposure pathways–potential recep-
tors at the specific site (USEPA 1996). The conceptual
site model for the chemical works is presented in
Fig. 3. The contaminant sources are likely to be locat-
ed within the manufacturing workshops, wastewater
treatment plant, and storage areas to the west of the
site. Leakage of toxic waste into the underlying
groundwater may have occurred. As the future land
use for the site has not been decided, the site will be
evaluated under both residential and commercial sce-
narios for the purpose of protecting human health.
Children up to 6 years old, adults under residential
scenario, and adults under commercial scenario are
considered on-site human receptors. Vapors are mainly
transported vertically through the diffusion in the pores
of aquiclude and bio-decay is not considered for the
purpose of conservatism. The on-site receptors may
Environ Sci Pollut Res (2016) 23:1234–1245
Fig. 2 The geologic profile of the
site (a); groundwater contours (b)
inhale indoor and outdoor vapor emitting from ground-
water pollution sources. The off-site villagers are un-
likely to inhale outdoor vapor through atmospheric dif-
fusion as the prevailing wind direction is in an east-
northeasterly direction. In addition, the villagers will
not be affected by inhalation of groundwater indoor
vapor since the groundwater flows from southwest to
northeast (Fig. 2b). The risk of drinking groundwater
is considered low as the shallow aquifer is not potable.
1237
Health risk assessment
The TCR level of 10−6 for the carcinogenic contaminants
and the THQ of 1 for the non-carcinogenic contaminants
as recommended by the CRAG (2014) were used for the
backward SSAC derivation. The Johnson-Ettinger model
(Johnson and Ettinger 1991) and ASTM guidelines
(ASTM 2004) were used to predict indoor or ambient
vapor concentrations through the migration from
1238
Fig. 3 Conceptual site model for the chemical works
groundwater sources into buildings and atmosphere
(Eqs. 1 and 2 in Table S1).
The HERA model and RBCA Tool Kit were selected to
calculate the SSAC through indoor vapor and outdoor vapor
inhalation pathways, respectively. Technical algorithm for the
derivation of SSAC and calculation of carcinogenic risks (CR)
and non-carcinogenic hazard quotients (HQ) is summarized in
Table S1 (Eqs. 3 to 12). The calculation of total cancer risks
(CRT) or overall hazard index (HI) for each chemical of con-
cern was the sum of CR or HQ of individual exposure path-
ways. It should be noted that the groundwater allocation factor
(WAF, 0.2) was taken into account for the calculation of Site-
Specific Assessment Criteria in case of the non-carcinogenic
contaminants (SSACnc) to take account of background expo-
sure in accordance with the CRAG guideline (2014), as shown
in Eq. 3 in Table S1. In order to facilitate the comparison of the
results, the effect of background exposure was also considered
in RBCA Tool Kit through multiplying the input parameter of
reference concentration parameters (RfC) by the same coeffi-
cient (0.2) for each chemical of concern (Eq. 4 in Table S1).
The physiochemical and toxicity parameters for each chem-
ical are reported in Table S2. In the study, the toxicity param-
eters of chemicals were mainly from the USEPA Integrated
Risk Information System (IRIS 2011), the Texas Risk Reduc-
tion Program (TRRP 2013), and the California Office of Envi-
ronmental Health Hazard Assessment (OEHHA 2013) in order
of decreasing preference. According to the International Agen-
cy for Research on Cancer (IARC 2014), benzene is classified
as a human carcinogen (Group 1); dichloromethane is regarded
as a probable carcinogen to humans (Group 2A); ethyl ben-
zene, 1,4-dichlorobenzene, carbon tetrachloride, and chloro-
form are possible carcinogens to humans (Group 2B); and the
remaining chemicals of concern are not classifiable as to carci-
nogenicity to humans (Group 3). Therefore, benzene, ethyl
benzene, 1,4-dichlorobenzene, carbon tetrachloride,
Environ Sci Pollut Res (2016) 23:1234–1245
chloroform, and dichloromethane have been assigned both
cancer unit risk (UR) and RfC. Toluene, monochlorobenzene,
1,2-dichlorobenzene, and 1,3-dichlorobenzene have been
assigned reference concentration parameters only. The detailed
input parameters necessary for the risk assessment are listed in
Table S3. The exposure parameters and indoor air and building
parameters were default values adopted from the CRAG
(2014). The outdoor air parameters, soil parameters, and
groundwater parameters were mostly site-specific value.
A parameter sensitivity analysis was undertaken in the
study to evaluate the sensitivity of an individual parameter
used in the health risk assessment (USEPA 2001). The sensi-
tivity ratio (SR) was calculated on the bases of the ratio of the
percentage change in output to the percentage change in input
for a specific input variable and Eq. 13 which is shown in
Table S1.
Results and discussion
Contaminant levels in groundwater
The concentration of contaminants is shown in Fig. 4a–c and
the statistical summaries including the maximum, minimum,
and average concentrations of the contaminants in groundwa-
ter and their detected frequencies of 16 sampling points are
summarized in Table 1. A significant variation in the contam-
inant levels in groundwater can indicate the presence of highly
heterogeneous aquifer. The average contaminant concentra-
tions were decreased in the order of chloroform >
monochlorobenzene>carbon tetrachloride>benzene>dichlo-
romethane>1,3-dichlorobenzene>1,4-dichlorobenzene>eth-
yl benzene > toluene > 1,2-dichlorobenzene. Chloroform,
monochlorobenzene, carbon tetrachloride, benzene, and di-
chloromethane were detected at the maximum concentrations
of 670.00, 130.36, 26.00, and 10.37, 3.19 mg/L with their
detection frequency being 81.25, 75.00, 25.00, 62.5, and
18.75 %, respectively. Concentrations of the remaining con-
taminants were lower than 0.1 mg/L and were occasionally
detected in the samples. The aqueous solubility in water for
chloroform, monochlorobenzene, benzene, and carbon tetra-
chloride are higher relative to other contaminants. Therefore,
they may be the major contribution to the contaminated
groundwater. Additionally, bivariate correlation analysis indi-
cated that the half lives derived from the first-order decay were
notably correlated with the maximum concentrations of the
contaminants (Spearman r=0.914, p<0.01).
Contaminant concentrations for the ten pollutants and chlo-
ride ion were statistically analyzed using PCA. The results of
the corresponding analysis are presented in Table S4 to
Table S6. The correlation matrix revealed strong correlation
between chloride ion and benzene, between 1,4-dichloroben-
zene and 1,2-dichlorobenzene and 1,3-dichlorobenzene, as
Environ Sci Pollut Res (2016) 23:1234–1245 1239

a 10
b 1000

100
Benzene Chloroform

Concentration (mg/L)
1 Toluene Carbon tetrachloride
Concentration (mg/L)

Dichloromethane
Ethyl benzene 10

0.1 1

0.1
0.01
0.01

1E-3 1E-3
GW1

GW2

GW3

GW4

GW5

GW6

GW7

GW8

GW9

GW10

GW13

GW14

GW15

GW1

GW2

GW3

GW4

GW5

GW6

GW7

GW8

GW9

GW10

GW13

GW14

GW15
Monitoring wells Monitoring wells

c d
100 GW14
Monochlorobenzene
1,2-dichlorobenzene
GW10
1,3-dichlorobenzene
10 1,4-dichlorobenzene
Concentration (mg/L)

GW15
GW6

PC2 (26.83 %)
1 0.5

GW7GW9
GW8 GW3
GW16 GW12
0.1 GW13 GW11
0.0 GW5
GW4 GW2
0.01
-1.0
-0.5 -0.5
0.0 )
4%
0.0 GW1
1E-3 PC 0.5
0.5
.6
14
GW1

GW2

GW3

GW4

GW5

GW6

GW7

GW8

GW9

GW10

GW13

GW14

GW15

1 (4 1.0
0.5 1.0 3(
9 %)
1.5
PC
Monitoring wells 1.5 2.0

Fig. 4 The individual contaminant concentrations in sampling location, chemicals at GW11, GW12, and GW16 being not detected (a–c). Principal
component analysis (PCA) scattergram of Bfactor scores^ for each sampling location (d)

Table 1 Summary of
contaminant concentrations in COCs Concentration (mg/L) Detected frequency (%)
groundwater
Max Min Average

Benzene 10.37 bdl 1.34 62.50

Toluene 0.026 bdl 0.0078 43.75
Ethyl benzene 0.043 bdl 0.016 25.00
Monochlorobenzene 130.36 bdl 30.53 81.25
1,2-Dichlorobenzene 0.0084 bdl 0.0039 25.00
1,3-Dichlorobenzene 0.11 bdl 0.041 43.75
1,4-Dichlorobenzene 0.067 bdl 0.025 37.50
Carbon tetrachloride 26.00 bdl 6.62 25.00
Chloroform 670.00 bdl 51.70 75.00
Dichloromethane 3.19 bdl 1.13 18.75
Cl− 4790.00 61.00 533.75 100.00
Depth (mbgl) 9.5 8.3 9
pH 7.6 5.9 7.2
Temperature (°C) 19.9 18.4 19.3

Groundwater samples were taken between November 7 and May 8, 2013

bdl below detection limit, mbgl meters below ground level
1240 Environ Sci Pollut Res (2016) 23:1234–1245

well as between dichloromethane and carbon tetrachloride and
chloroform, respectively (Table S4). The first three principal
components (PCs) were extracted, accounting for approxi-
mately 82 % (PC1, 40.58 %; PC2, 26.83 %; PC3, 14.64 %)
of the total variance of the outcome (Table S5). The first prin-
cipal component (PC1) was associated with toluene, carbon
tetrachloride, and chloroform. The second principal compo-
nent (PC2) was related to ethyl benzene, 1,2-dichlorobenzene,
1,3-dichlorobenzene, and 1,4-dichlorobenzene. The third
principal component (PC3) was linked to benzene,
monochlorobenzene, and chloride ion (Table S6). The PCA
scatter diagram of Bfactor score^ for each sampling location is
presented in Fig. 4d. The figure revealed that the sampling
locations could be clustered into two subgroups. Most
sampling locations in the first subgroup including GW14,
GW10, GW15, GW7, GW8, GW9, GW3, GW16, GW13,
GW5, GW11, and GW12 with relative lower concentration
were related to PC2 and PC3 mainly. The second subgroup
was associated with GW1, GW2, and GW4, which was clus-
tered in the plane of PC1 and PC3 attributable to higher con-
centrations of monochlorobenzene and benzene and trace
amounts of dichlorobenzene. The most highlighted sampling
location for GW6 in the plane of PC1 and PC2 was ascribed to
the highest level of chloroform and carbon tetrachloride and
relative higher concentration of monochlorobenzene. There-
fore, the characterization of contaminated groundwater by
principal component analysis may provide the potential risk
hotspot. GW1 is situated in the manufacturing workshop,

Table 2 The specific technical algorithms for SSAC calculation in RBCA and HERA

Commercial scenario

Non-
T HQ  Rf Di  W AF  BW a  AT nc  365 days=year
carcinogens SSAC HERA ¼ ¼
nc
V F  RV a  E F  EDa
T HQ  R f C  RV a  W A F  BW a  AT nc  365 days=year T HQ  R f C  W AF  AT nc  365 days=year
¼ (1)
BW a  V F  RV a  E F  EDa V F  E F  EDa
nc 365 days=year
SSAC RBCA
nc ¼ T HQR fVCAT
FE FEDa  W A F (2)
Carcinogens
T CR  BW a  AT ca  365 days=year T CR  BW a  AT ca  365 days=year
SSAC HERA ¼ ¼ U RBW a ¼
ca
S F i  V F  E F  EDa  RV a RV a  V F  E F  EDa  RV a
T CR  AT ca  365 days=year
(3)
U R  V F  E F  EDa
ca 365 days=year
SSAC RBCA
ca ¼ T CRAT
U RV FE FEDa (4)
Residential scenario
Non-
T HQ  Rf Di  W AF  BW c  AT nc  365 days=year
carcinogens SSAC HERA ¼ ¼
nc
V F  RV c  E F  EDc
T HQ  R f C  RV a  BW c  AT nc  365 days=year
 W AF (5)
BW a  V F  RV c  E F  EDc
nc 365 days=year
SSAC RBCA
nc ¼ T HQR fVCAT
FE FEDc  W A F (6)

SSAC RBCA a =RV a

nc
SSAC HERA
¼ BW
BW c =RV c (7)
nc

T CR  AT ca  365 days=year EDc RV c EDa RV a  T CR  AT ca  365 days=year EDc RV c EDa RV a 
Carcinogens
SSAC HERA ¼ þ BW a ¼ þ BW a
ca
S F i  V F  E F BW c U RBW a
RV a  V F  E F BW c

(8)

ca 365 days=year
SSAC RBCA
ca ¼ T CRAT
U RV FE F ðEDc þ EDa Þ (9)
 
SSAC RBCA RV c EDc þEDa
ca
SSAC HERA
¼ BW
RV a 
a
BW c ðEDc þ EDa Þ (10)
ca

SSAC (mg/L), site-specific assessment criteria; subscripts Bnc^ and Bca^ mean non-carcinogens and carcinogens, respectively; superscript BHERA ^ and
BRBCA ^ mean HERA model or from the RBCA Tool Kit; THQ (unitless), target hazard quotient; TCR (unitless), target cancer risk; RfDi (mg/kg/day),
reference dose; UR (1/mg/m3 ), unit risk; SFi (1/(mg/kg/day)), cancer slope factor; RfC (mg/m3 ), reference concentration; VF (mg m−3 /mg L−1 ),
volatilization factors; ATnc or ATca (year), averaging time for non-carcinogens or for carcinogens; EF (days/year), exposure frequency; ED (year),
exposure duration; RV (m3 /day), the daily inhalation rate; BW (kg), body weight; subscripts Bc ^ and Ba ^ mean child and adult, respectively
Environ Sci Pollut Res (2016) 23:1234–1245 1241

while GW2, GW4, and GW6 were within the raw material and
product warehouse. Therefore, these monitoring wells are
considered within the potential groundwater pollution
sources.
Model comparison between RBCA and HERA models
The technical algorithm to calculate SSAC is listed in
Table S1. As illustrated in Eqs. 1 and 2, both RBCA and
HERA models have implemented identical transport models
for the calculation of volatilization factors (VF) for individual
exposure route under commercial and residential redevelop-
ment scenarios. Therefore, the identical results for VF
were obtained using HERA model and RBCA Tool Kit
(Table 3). Additionally, the calculated VF indicates that
the human receptors in future may be more vulnerable to
the VOCs volatile into the closed space because the VFs
calculated for indoor vapor pathways were approximately
three orders of magnitude higher than those for outdoor
vapor exposure pathways (Table 3).
For the SSAC calculation, there is a fine distinction be-
tween HERA model and RBCA Tool Kit, although most
similarities exist between the two models (Table S1). In detail,
HERA model applied reference dose (RfDi, mg/kg/day) and/
or cancer slope factor (SFi, 1/(mg/kg/day)) as the acceptable
daily exposure reference to the human receptor and considered
body weight (BW) and daily inhalation rate (RV) for specific
receptors at the same time (Eqs. 3 and 7). Compared with
HERA model, RBCA Tool Kit used the RfC (mg/m3) and/or
UR (1/mg/m3) as the acceptable daily exposure reference
without respect to the parameters of receptors (Eqs. 4 and 8).
RfC and UR can be converted into RfDi and SFi through mul-
tiplying by the ratio of RV and BW for adult or their inverse
form, respectively (Eqs. 5 and 9). Therefore, the same result
will be obtained when the adult receptor is taken into account
but different outcomes when multiple receptors are considered
using the two models.
The specific technical algorithms for SSAC calcula-
tion using RBCA and HERA is listed in Table 2 and
the calculated results are listed in Table 3. As shown in
Table 3, identical SSACs under the commercial redevel-
opment scenario were calculated through the conversion
between RfDi and RfC or between SFi and UR for only
adult receptors (Eqs. 1 to 4 in Table 2). However,

Table 3 Volatilization factors and Site-Specific Assessment Criteria from RBCA and HERA tools

Chemicals RBCA Tool Kit HERA model

VFov VFiv SSACiv SSACov VFov VFiv SSACiv SSACov

mg m−3/mg L−1 mg/L mg m−3/mg L−1 mg/L

Commercial scenario
Benzene 9.52E−07 1.77E−04 14.36 >1770 9.52E−07 1.77E−04 14.36 >1770
Toluene 9.09E−07 1.70E−04 >530 >530 9.09E−07 1.70E−04 >530 >530
Ethyl benzene 8.57E−07 1.60E−04 14.00 >169 8.57E−07 1.60E−04 14.00 >169
Monochlorobenzene 7.93E−07 1.47E−04 132.01 >502 7.93E−07 1.47E−04 132.01 >502
1,2-Dichlorobenzene 6.51E−07 1.20E−04 97.71 >150 6.51E−07 1.20E−04 97.71 >150
1,3-Dichlorobenzene 7.51E−07 1.40E−04 22.28 >110 7.51E−07 1.40E−04 22.28 >110
1,4-Dichlorobenzene 6.74E−07 1.25E−04 4.09 >73.8 6.74E−07 1.25E−04 4.09 >73.8
Carbon tetrachloride 1.71E−06 3.22E−04 2.90 >805 1.71E−06 3.22E−04 2.90 >805
Chloroform 9.19E−07 1.71E−04 1.43 795.69 9.19E−07 1.71E−04 1.43 795.69
Dichloromethane 9.67E−07 1.78E−04 67.11 >15400 9.67E−07 1.78E−04 67.11 >15,400
Residential scenario
Benzene 9.52E−07 4.44E−04 3.42 >1770 9.52E−07 4.44E−04 2.92 >1770
Toluene 9.09E−07 4.25E−04 >530 >530 9.09E−07 4.25E−04 >530 >530
Ethyl benzene 8.57E−07 4.01E−04 3.33 >169 8.57E−07 4.01E−04 2.85 >169
Monochlorobenzene 7.93E−07 3.69E−04 37.72 >502 7.93E−07 3.69E−04 20.41 >502
1,2-Dichlorobenzene 6.51E−07 2.99E−04 27.92 >150 6.51E−07 2.99E−04 15.11 >150
1,3-Dichlorobenzene 7.51E−07 3.49E−04 6.37 >110 7.51E−07 3.49E−04 3.45 >110
1,4-Dichlorobenzene 6.74E−07 3.12E−04 0.97 >73.8 6.74E−07 3.12E−04 0.83 >73.8
Carbon tetrachloride 1.71E−06 8.05E−04 0.69 >805 1.71E−06 8.05E−04 0.59 >805
Chloroform 9.19E−07 4.27E−04 0.34 473.62 9.19E−07 4.27E−04 0.29 404.96
Dichloromethane 9.67E−07 4.44E−04 15.98 >15400 9.67E−07 4.44E−04 13.66 >15,400

The subscripts Biv^ and Bov^ are for indoor and outdoor vapor pathway; B>value^ indicates that the calculated SSAC exceeds the compound’s solubility
1242 Environ Sci Pollut Res (2016) 23:1234–1245

SSACs under the residential redevelopment scenario cal-
culated using HERA software were more conservative
than those calculated using RBCA Tool Kit, especially
for the non-carcinogenic contaminants. The differences
in SSACs from these two models are explained. The
critical receptors for carcinogenic contaminants under
the residential scenario are children and adults. As a
result, the individual RV and BW of the receptor were
taken into account in HERA model (Eqs. 8 in Table 2),
whereas RBCA Tool Kit considered the exposure dura-
tion (ED) of individual receptor only (Eqs. 9 in
Table 2). Although the toxicity parameter of SFi could
be converted into UR, an identical result could not be
obtained and the ratio of SSAC for carcinogen derived
from RBCA (SSAC ca R BC A ) and that from HERA
(SSACcaHERA) is shown in Eqs. 10 in Table 2. In the
case of non-carcinogens, children were selected as the
most sensitive receptors under the residential scenario.
SSAC was predicted in the same way and the specific
technical algorithms are listed in Eqs. 5 and 6. Like-
wise, the ratio of SSAC for non-carcinogen (SSAC nc)
from the two models is shown in Eqs. 7 in Table 2. Con-
sequently, the SSAC derived from RBCA Tool Kit for
non-carcinogenic chemicals under the residential scenar-
io are 1.85 times higher than those derived from HERA
model, whereas the SSAC derived from RBCA Tool Kit
for carcinogenic chemicals under the residential scenario
are slightly less with 1.17 times higher than those derived
from HERA model. Therefore, the main discrepancy in
exposure calculations was responsible for the different
SSACs calculated using RBCA and HERA tools.
Detailed quantitative risk assessment using HERA model
Integrated SSAC comparing with actual concentration
of chemicals
The technical algorithm for integrated SSAC is listed in
Eqs.11 and 12 in Table S1. The results for SSACs for single
exposure pathway and the integrated SSACs from multiple
exposures derived by HERA are summarized in Tables 3
and 4. Typically, SSACs for indoor vapor inhalation route
under both residential and commercial scenarios were almost
identical to their respective integrated SSACs indicating that
the inhalation of groundwater indoor vapor was a critical path-
way. This is because of the larger exposure frequency (EF)
setting (Table S3) and stronger volatilization factor (VF). The
results shown in Table 4 revealed a variety of target cleanup
criteria for different chemicals of concern. The SSAC values
under residential scenario were more conservative than those
under commercial scenario as children were considered to be
the critical receptors under the residential scenario. From the
toxicity point, SSAC values for carcinogenic compounds were
basically more stringent due to their serious toxicity, thus the

Table 4 Site-Specific
Assessment Criteria for COCs Residential Commercial
contaminants of concerns
SSACint (mg/L) SSAC exceedance SSACint (mg/L) SSAC exceedance

Benzenea 2.92 GW1 (10.37) 14.33 –

Tolueneb >530 – >530 –
Ethyl benzenea 2.85 – 13.97 –
Monochlorobenzeneb 20.40 GW1 (72.01) 131.77 –
GW2 (130.36)
GW4 (120.66) –
GW6 (63.90) –
1,2-Dichlorobenzeneb 15.10 – 97.53 –
1,3-Dichlorobenzeneb 3.44 – 22.24 –
1,4-Dichlorobenzenea 0.83 – 4.08 –
Carbon tetrachloridea 0.59 GW6 (26.00) 2.90 GW6 (26.00)
Chloroforma 0.29 GW1 (0.49) 1.42 GW6 (670.00)
GW6 (670.00)
GW13 (0.73)
GW15 (0.78)
Dichloromethanea 13.65 – 66.99 –

B>value^ indicates that the calculated SSAC exceeds the compound’s solubility; SSACint, integrated SSAC for
multiple exposure pathway. a carcinogenic chemical; b non-carcinogenic chemical.
a
Carcinogenic chemical
b
Non-carcinogenic chemical
Environ Sci Pollut Res (2016) 23:1234–1245 1243

target threshold level of the contaminants in the groundwater and GW6 were in the range of 3.13 to 6.39 times over the
had to be strictly restricted ranging from 0.29 to 13.65 mg/L. SSACs. The concentration of benzene (10.37 mg/L) also
In terms of non-carcinogenic substances, SSACs for 1,3-di- exceeded the SSAC of 2.92 mg/L at GW1. Under the com-
chlorobenzene, 1,2-dichlorobenzene, and mercial scenario, contaminant concentrations for chloroform
monochlorobenzene were relatively lenient in the range of and carbon tetrachloride at GW6 are 470.63 and 8.98 times
3.44–20.40 mg/L under residential scenario, whereas SSACs above their respective SSACs. Therefore, potential risks
for toluene was greater than their corresponding solubility associated with inhalation of indoor vapor have been
(530 mg/L), suggesting that toluene may pose a negligible identified for benzene, monochlorobenzene, carbon tetra-
health effect. chloride, and chloroform under the residential redevelop-
Contaminant concentrations at several groundwater moni- ment, and carbon tetrachloride and chloroform under
toring locations were in excess of SSACs for chloroform, commercial scenario.
carbon tetrachloride, monochlorobenzene, and benzene under
the residential scenario (Table 4). The level of the exceedance Carcinogenic risk and non-carcinogenic hazard assessment
was increased in the order of chloroform>carbon tetrachlo-
ride>monochlorobenzene>benzene. The concentrations in The results of CRT and HI via multi-exposure pathways for
groundwater for chloroform (670.00 mg/L) and carbon tetra- each chemical of concern under residential and commercial
chloride (26.00 mg/L) at GW6 were found to be 2309.29 and scenarios are explained in Fig. 5a–d. The collected data for
44.05 times over their respective SSACs. The levels of CRT and/or HI exhibited that the unacceptable risk or hazard
monochlorobenzene in groundwater at GW1, GW2, GW4, effects were associated with benzene, monochlorobenzene,

a 4
b 3
10 10
Residential 2 Commercial
10
2 10
1
0 10
10
-6
Cancer risk*10

0
-6

10
Cancer risk*10

-2
10
-1
10
-4
10
-2
10
-6
10 -3
10
-8
10 10
-4

-10 -5
10 10
GW1
GW2
GW3
GW4
GW5
GW6
GW7
GW8
GW9
GW10
GW11
GW12
GW13
GW14
GW15
GW16

GW1
GW2
GW3
GW4
GW5
GW6
GW7
GW8
GW9
GW10
GW11
GW12
GW13
GW14
GW15
GW16
c 2
Monitoring wells d 10 1
Monitoring wells
10

10
1 Residential 10
0
Commercial
0 -1
10 10
-1 -2
10 10
Hazard index

Hazard index

-2 -3
10 10
-3 -4
10 10
-4 -5
10 10
-5 -6
10 10
-6 -7
10 10
-7 -8
10 10
GW1
GW2
GW3
GW4
GW5
GW6
GW7
GW8
GW9
GW10
GW11
GW12
GW13
GW14
GW15
GW16

GW1
GW2
GW3
GW4
GW5
GW6
GW7
GW8
GW9
GW10
GW11
GW12
GW13
GW14
GW15
GW16

Monitoring wells Monitoring wells

Fig. 5 The total carcinogenic risk and hazard index under residential and 1,3-dichlorobenzene; open triangles, 1,4-dichlorobenzene; open down-
commercial exposure scenarios. Filled squares, benzene; filled circles, pointing triangles, carbon tetrachloride; plus symbols, chloroform;
toluene; filled triangles, ethyl benzene; filled down-pointing triangles, asterisks, dichloromethane
monochlorobenzene; open squares, 1,2-dichlorobenzene; open circles,
1244
chloroform, and carbon tetrachloride. The remaining
chemicals showed negligible harmful effects. The calculated
carcinogenic risk levels were in the range of 10−10 to 10−6 at
most of the groundwater monitoring wells, with the exception
of GW1, GW6, GW13, and GW15. Under the residential
scenario, the cancer risk at GW1 for benzene and at GW13
and GW15 for chloroform was 3.55×10−6, 2.46×10−6, and
2.67×10−6, respectively, which are considered as Bpossible
risk^ between 1×10−6 and 1×10−5 proposed by Sexton et al.
(2007). The calculated cancer risk of 4.40×10−5 for carbon
tetrachloride at GW6 under residential redevelopment falls
into Bprobable risk^ between 1×10−5 and 1×10−4. The cancer
risk of 2.31×10−3 for chloroform exceeds the regulatory limit
of 1×10−4 classified as a Bdefinite risk^. In comparison with
the residential scenario, the cancer risks for benzene and
carbon tetrachloride under commercial scenario were de-
creased one order of magnitude, but the risk of 4.71×
10−4 for chloroform and 8.98×10−6 for carbon tetrachlo-
ride at GW6 still posed a considerable risk for inhala-
tion of groundwater indoor vapors. In terms of the non-
carcinogenic hazard index under the residential scenario,
carbon tetrachloride, benzene, and monochlorobenzene
were found to be of concern as the HI in the range of
1.02 to 6.39 at GW1, GW2, GW4, and GW6 exceeded
the threshold value of 1. Furthermore, chloroform with
the highest HI of 19.62 for residential scenario and 3.04
for commercial scenario were considered to pose the
most serious toxic effect to human health.
Sensitivity analysis of parameters
The parameter sensitivity analysis was performed under resi-
dential scenario as an example. The SR of the parameters for
residential redevelopment is shown in Table S4. The outcomes
indicated that volume/area ratio (LB) and indoor air exchange
rate (ER) were the extraordinary sensitive parameters with an
SR value of −0.95, and they were negatively correlated with
both CRT and HI. While the SR for areal fraction of cracks (η),
the depth to the foundation (Zcrack), and the foundation thick-
ness (Lcrack) was approximately 0, indicating that the parame-
ters were relatively insensitive to both CRT and HI estimates.
With regard to the exposure parameters of receptors, data of
adults were more sensitive (SR>0.65) relative to those of
children (SR<0.32) including the BW, the indoor vapor ex-
posure frequency (EFiv-a), the exposure duration (EDa), and
daily inhalation rate (RVa) in terms of CRT assessment. Addi-
tionally, the BW was inversely related to both CRT and HI,
suggesting that increasing the BW of receptor would weaken
the deleterious effects of pollutants regardless of other factors.
In contrast to CRT, children were taken into account only for
the HI calculation. Therefore, the exposure parameters
discussed above for children were the most sensitive parame-
ters with SR values between 0.95 and 1.00. Consequently, the
Environ Sci Pollut Res (2016) 23:1234–1245
exposure parameters related to the indoor vapor inhalation
were more sensitive and profoundly affected the assessment
of CRT and HI.
Conclusions
In this study, modeling validations and comparisons including
the similarities and differences between HERA software that
implemented technical algorithm and exposure scenarios of
Chinese risk assessment guideline (CRAG 2014) and an inter-
nationally accepted RBCATool Kit have been undertaken. Fol-
lowing the validation, human health risk assessment of volatile
organic solvents in groundwater has been presented under the
commercial and residential scenarios for a former chemical
works in China. Detailed model comparisons indicated the
identical SSACs under the commercial redevelopment scenario
from both models, but more conservative and stringent SSAC
values under the residential scenario using HERA model than
those obtained from RBCATool Kit. These are attributed to the
identical algorithms on transport models implemented in both
models and the difference in exposure calculation under the
residential scenarios. The inhalation of indoor vapor was iden-
tified to be the most predominant exposure pathway for VOCs
determined using the RBCA and HERA models.
The outcome of the risk assessment using HERA model
identified unacceptable carcinogenic risks posed by benzene,
monochlorobenzene, carbon tetrachloride, and chloroform in
the range of 2.46×10−6 to 2.31×10−3 and the unacceptable
highest non-carcinogenic hazard quotient for chloroform of
19.62 under the residential redevelopment. Under the commer-
cial scenario, carcinogenic risk levels of 4.71×10−4 and 8.98×
10−6 have been calculated for carbon tetrachloride and chloro-
form, respectively. Finally, it is revealed from the parameter
sensitivity analysis that the most sensitive parameters are those
relating to the pathway of indoor vapor inhalation and pro-
foundly affect the assessment of CRT and HI.
It can be concluded from the above analyses that
potential health risks exist under future redevelopment
scenarios and development of a risk-based remedial
strategy is recommended to ensure the safe and sus-
tainable redevelopment of the site.
Acknowledgments The National High Technology Research and De-
velopment Program (863 Program, 2013AA06A208), the Science and
Technology Service Network Initiative (STS, KFJ-EW-STS-091), the
Public Health Research Funding from the Ministry of Environmental
Protection of the People’s Republic of China (201309005), and the Na-
tional Science Foundation of China (41471404 and 51309214) are
acknowledged.
Environ Sci Pollut Res (2016) 23:1234–1245
References
ASTM (American Society for Testing and Materials) (2004) Standard
guide for risk-based corrective action. ASTM E-2081-00,
Philadelphia, PA
Baciocchi R, Berardi S, Verginelli I (2010) Human health risk assess-
ment: models for predicting the effective exposure duration of on-
site receptors exposed to contaminated groundwater. J Hazard Mater
181(1–3):226–233
Bardos P, Andersson-Skold Y, Blom, S, Keuning S, Pachon C, Track T
(2008) Special Session 1 Brownfields, Bioenergy and Biofeedstocks
Special Session 2 Green Remediation. 10th International UFZ-
Deltares/TNO Conference on Soil-Water Systems. Milan, Italy
Cachada A, Pato P, Rocha-Santos T, Ferreira da Silva E, Duarte AC
(2012) Levels, sources and potential human health risks of organic
pollutants in urban soils. Sci Total Environ 430:184–192
Carlon C, Pizzol L, Critto A, Marcomini A (2008) A spatial risk assess-
ment methodology to support the remediation of contaminated land.
Environ Int 34:397–411. doi:10.1016/j.envint.2007.09.009
Chen M (2010) Analytical integration procedures for the derivation of
risk-based generic assessment criteria for soil. Hum Ecol Risk
Assess 16:1295–1317
CRAG (Technical guidelines for risk assessment of contaminated sites)
(2014) Beijing: Ministry of Environmental Protection of the
People's Republic of China [in Chinese]. http://kjs.mep.gov.cn/
hjbhbz/bzwb/trhj/trjcgfffbz/201402/t20140226_ 268358.htm.
Accessed 2 January 2014
Critto A, Carlon C, Marcomini A (2003) Characterization of contaminat-
ed soil and groundwater surrounding an illegal landfill (S. Giuliano,
Venice, Italy) by principal component analysis and kriging. Environ
Pollut 122:235–244
Faerber T, Herzog J (2010) Solid waste management and environmental
remediation: environmental remediation technologies, regulations
and safety. NOVA Science Publishers, New York, pp 335–362.
ISBN 9781612092591
Fan C, Wang GS, Chen YC, Ko CH (2009) Risk assessment of exposure
to volatile organic compounds in groundwater in Taiwan. Sci Total
Environ 407:2165–2174
GSI (GSI Environmental Inc) (2014) http://gsi-net.com/en/software.html.
Accessed 20 January 2014
IARC (International Agency for Research on Cancer) (2014) List of
classifications by alphabetical order. http://monographs.iarc.fr/
ENG/Classification/index.php. Accessed 10 June 2014
IRIS (Integrated Risk Information System) (2011) Full List of IRIS
Chemicals. http://cfpub.epa.gov/ncea/iris/index.cfm?fuseaction=
iris.showSubstanceList. Accessed 13 July 2014
ISO (International Organization for Standardization) (2012) Soil
Quality—Assessment of Impact from Soil Contaminated with
Petroleum Hydrocarbons. ISO 11504: 2012. Geneva, Switzerland
Jeffries J, Martin I (2009) Updated technical background to the CLEA
model; science report: SC050021/SR3. Environmental Agency
U.K., Bristol
Johnson PC, Ettinger RA (1991) Heuristic model for predicting the intru-
sion rate of contaminant vapours into buildings. Environ Sci
Technol 25:1445–1452
Johnston JE, Gibson JM (2013) Screening houses for vapour intrusion
risks: a multiple regression analysis approach. Environ Sci Technol
47:5595–5602
Kitwattanavong MK, Prueksasit T, Morknoy D, Tunsaringkarn T,
Siriwong W (2013) Health risk assessment of petrol station workers
in the inner city of Bangkok, Thailand, to the exposure to BTEX and
carbonyl compounds by inhalation. Hum Ecol Risk Assess 19:
1424–1439
1245
Lapworth DJ, Baran N, Stuart ME, Ward RS (2012) Emerging organic
contaminants in groundwater: a review of sources, fate and occur-
rence. Environ Pollut 163:287–303
Lopez E, Schuhmacher M, Domingo JL (2008) Human health risks of
petroleum-contaminated groundwater. Environ Sci Pollut Res 15(3):
278–288
Luo F, Song J, Chen M, Wei J, Pan YY, Yu HB (2013) Risk assessment of
manufacturing equipment surfaces contaminated with DDTs and
dicofol. Sci Total Environ 468:176–185
Ma XM, Anand D, Zhang XF, Talapatra S (2011) Adsorption and desorp-
tion of chlorinated compounds from pristine and thermally treated
multiwalled carbon nanotubes. J Phys Chem C 115(11):4552–4557
MEPPRC (Beijing: Ministry of Environmental Protection of the People’s
Republic of China) (2009) Monitoring and analysis methods of wa-
ter and wastewater, 4th edn. China Environmental Science Press,
Beijing [in Chinese]
Nieuwenhuijsen M, Paustenbach D, Duarte-Davidson R (2006) New de-
velopments in exposure assessment: the impact on the practice of
health risk assessment and epidemiological studies. Environ Int
32(8):996–1009
OEHHA (Office of Environmental Health Hazard Assessment) (2013)
OEHHA Toxicity Criteria Database. http://www.oehha.ca.gov/
tcdb/index.asp. Accessed 13 July 2014
Pinedo J, Ibanez R, Irabien A (2014) A comparison of models for
assessing human risks of petroleum hydrocarbons in polluted soils.
Environ Model Softw 55:61–69
Ren WX, Geng Y, Ma Z, Sun L, Xue B, Fujita T (2015) Reconsidering
Brownfield redevelopment strategy in China’s old industrial zone: a
health risk assessment of heavy metal contamination. Environ Sci
Pollut Res 22:2765–2775
Sexton K, Linder SH, Marko D, Bethel H, Lupo PJ (2007) Comparative
assessment of air pollution-related health risks in Houston. Environ
Health Perspect 115(10):1388–1393
Stezar IC, Pizzol L, Critto A, Ozunu A, Marcomini A (2013) Comparison
of risk-based decision-support systems for Brownfield site rehabili-
tation: DESYRE and SADA applied to a Romanian case study. J
Environ Manag 131:383–393. doi:10.1016/j.jenvman.2013.09.022
TRRP (Texas Risk Reduction Programme) (2013) Toxicity and physical
and chemical parameters summary table that used to derive the
Texas State Protective Concentration Levels (PCL). http://www.
tceq.state.tx.us/remediation/trrp/ trrppcls.html. Accessed 13
July 2014
USEPA (United States Environmental Protection Agency) (1996) Soil
screening guidance: technical background document. EPA/540/
R95/128. Office of Solid Waste and Emergency Response,
Washington, DC
USEPA (United States Environmental Protection Agency) (2001) Risk
assessment guidance for superfund vol. 3, part A, process for
conducting probabilistic risk assessment. Office of Emergency and
Remedial Response, Washington, DC
USEPA (United States Environmental Protection Agency) (2012) Human
health risk assessment. Strategic research action plan 2012–2016, 601/
R-12/007. Office of Research and Development, Washington, DC
Wei J, Chen M, Song J, Luo F, Han L, Li CP, Dong MG (2015)
Assessment of human health risk for an area impacted by a large-
scale metallurgical refinery complex in Hunan, China. Hum Ecol
Risk Assess 21:863–881
Zhou YY, Zhou HY, Tang YD, Yan ZG, Zhang CY, Jiang L, Li FS, Bai
LP, Liu L (2012) Development of soil screening levels for contam-
inated sites assessment in Beijing. International Conference on
Civil, Architectural and Hydraulic Engineering, Zhangjiajie, China