ISPRS, Vol.34, Part 2W2, “Dynamic and Multi-Dimensional GIS", Bangkok, May 23-25, 2001 
61 
ocarbon- 
ar. 
gh widely 
irt Monte 
i analysis 
posed by 
j a fuzzy 
cated at 
atchewan, 
adsorption 
a tank, a 
i salt bath 
They had 
ever since 
;igned and 
scrubbers, 
ream prior 
n line. The 
l from gas 
f the site 
loved from 
>ed of in a 
) located 
ng (Roper 
ciates Ltd., 
the nature 
Site. “It is 
iommodate 
cently, the 
jve ground 
outh of the 
the site and 
ensate was 
aepage into 
994). Two 
ned in the 
a between 
vironmental 
iducted by 
n, Including 
92, Clifton 
' Consulting 
1998. These 
ce sampling 
nitoring well 
ance of free 
: in several 
1995, 1996; 
dual phase 
I samples in 
ìe, Toluene, 
;he AST and 
ational, Inc., 
ately model 
he saturated 
mensions in 
:tured media, 
ideling which 
could model 
sorption, and 
microbial processes based on oxygen-limited, anaerobic, first- 
order, or Monod-type biodegradation kinetics as well as 
anaerobic or first-order sequential degradation involving 
multiple daughter species. 
Here, four Light Non-Aqueous Phase Liquid (LNAPL) Benzene, 
Toluene, Ethylene, and Xylene (BTEX) are identified as main 
contaminants of concern that are considered to contribute most 
significantly to risks for the exposed populations due to their 
relatively high concentration, and their dangerous toxicity in 
nature. 
Depending on the inputs of required model parameters, sets of 
time series contaminants concentration results will be derived. 
They will be sequentially employed for the risk assessment 
process. 
2.2 Individual Based Probabilistic and Possibilistic Risk 
Assessment 
Uncertainties inherently exist in the environmental process due 
to sparse and imprecise natures of the available information. 
Variabilities existing in the individuals make it even more 
complicated to evaluate the relevant human health risk. Truly, 
exposures will vary as individual vary in terms of age, gender, 
health status, exposure duration, propensity to drink water, 
body size and so on. 
Generally, most of the previous risk analysts argued that risk 
should be measured through considering the probability of a 
damage that may occur following exposure of a target to 
contaminants. A number of authors have adapted statistical 
data related to variables such as drinking water ingestion, soil 
ingestion, and residential tenure into forms designed for use in 
probabilistic risk assessment. U.S. Environmental Protection 
Agency (USEPA) also accepted probabilistic risk analysis and 
simulation modeling as appropriate tools for risk analysis as 
CERCLA or “superfund” sites. 
2.2.1 Risk Characterization 
Contaminant concentrations for the site health risk assessment 
were gathered during a field effort which included sampling of 
surface and subsurface of the site and the time series 
simulation results from groundwater transport model. Exposure 
pathways and exposure populations were identified based on 
the observed spatial patterns of site contamination and on-site 
specific land use and behavior of potential receptors. Here, 
one potential exposure pathway, groundwater ingestion is 
carried through the quantitative risk analysis process. 
Standard exposure pathway intake models were used to 
estimate chronic contaminants ingestions. 
r nf . CWxIRxCFxFIxEFxED 
BWxAT 
where, CDI = chronic daily intake of contaminant (mg/kg/day); 
CW = contaminant concentrations in groundwater (pg/L); IR = 
groundwater ingestion rate (LVday); CF= conversion factor (10' 
mg/pg); FI = relative fraction absorbed from water (100%); EF 
= frequency of site contact (350 days/365 days); ED = 
exposure duration (30 years); BW = body weight (kg); AT = 
averaging time (years) 
The results from equation (1) can then be taken as inputs to 
the risk characterization process. USEPA method for health 
risk effect is employed and given in the following equation: 
Excess Lifetime Cancer Risk (ELCR)= Chronic daily intake 
(CDI)* Cancer Potential Factor (CSF) (2) 
The Slope factor is generally derived from the animal 
experiment data, and used with administered doses to 
estimate probability of increased cancer incidence over a life 
time. 
Monte Carlo Simulation 
Monte Carlo simulation addresses the weakness of the current 
risk assessment methods pose in the uncertainty analysis. In 
extending the regular methods for public health risk 
assessment, Monte Carlo simulation involves several valuable 
technologies to estimate both point values and full distributions 
for the exposure and risk (Burmaster et al., 1989) These 
extended techniques make the analyses more informative to 
risk managers and members of the public by giving some 
perspective of the uncertainty behind the point estimate 
(Finkel, 1990). 
The first step in the Monte Carlo simulation is to determine 
PDFs to describe sensitive variables in the uncertainty 
analysis, probability distributions for each of exposure 
parameter values were estimated using site-specific data, 
estimates from the literature review, and judgment from 
receptor behavior. For each uncertain parameter, a single 
value is randomly generated based on the probability 
distribution function. These single values are then used in the 
subsequent model calculations to produce a single answer. A 
large number of iterations are performed, repeating this 
procedure of generating random values and performing model 
calculation. Therefore, the results will finally lead to a 
distribution of possible values of daily intake rather than a 
single value (Burmaster et al., 1988). Chronic contaminant 
ingestion values were calculated for each pathway and 
exposed population using Monte Carlo simulation model as 
describe below. And Results are then used as inputs for the 
further risk analysis. 
Parameters distribution function and sampling monitoring well 
data are listed in following table: 
Table 2. Parameter Distribution Summary for various groups 
Distributi 
on 
Symbol 
PDF features 
Mean 
St.Dev 
Contaminants concentration in 
groundwater (ng /L) 
Concentr 
ation 
data 
from site 
cw 
Groundwater ingestion rate 1R 
(L/davl 
Age group 1: 0-2 years old 
Lognorm 
al 
IR, 
0.3 
0.I5 
Age group 2: 2-7 years old 
Lognorm 
al 
1R 2 
0.6 
0.3 
Age group 3: 7-12 years old 
Lognorm 
al 
IR 3 
0.7 
0.4 
Age group 4: 12-18 years old 
Lognorm 
al 
1R 4 
1.5 
0.8 
Age group 5: >18 
Lognorm 
al 
IR, 
2 
1.0 
Body Weiqht(kq) BW 
Age group 1: 0-2 years old 
Lognorm 
al 
BW, 
7.2 
2.12 
Age group 2: 2-7 years old 
Lognorm 
al 
BW> 
15.6 
3.55 
Age group 3: 7-12 years old 
Lognorm 
al 
BW, 
28.2 
4.49 
Age group 4: 12-18 years old 
Lognorm 
al 
BW 4 
49.2 
8.39 
Age group 5: >18 
Lognorm 
al 
BW 5 
70.1 
14.7 
Averaqed exposure time(years) AT 
Aqe qroup 1: 0-2 years old 
Normal 
AT, 
70 
14 
Aqe qroup 2: 2-7 years old 
Normal 
at 2 
70 
14 
Aqe group 3: 7-12 years old 
Normal 
AT 3 
60 
14 
Aqe qroup 4: 12-18 years old 
Normal 
at 4 
50 
10 
Aqe qroup 5: >18 
Normal 
at 5 
30 
6